Compare commits

...
Author SHA1 Message Date
Martin Kroeker 5f2a3c05cd Revert "Rewrite &= -> = and simplify the initial blocking phase." 2018-07-03 21:42:28 +02:00
Martin Kroeker a83f01e0ee Merge pull request #1662 from martin-frbg/cmake-avx512
Add -march=skylake-avx512 to AVX512 compile check and suppress its ou…
2018-07-03 17:40:09 +02:00
Martin Kroeker b74aef2816 Add -march=skylake-avx512 to AVX512 compile check and suppress its output 2018-07-03 14:41:44 +02:00
Martin Kroeker a9fa805007 Merge pull request #1660 from martin-frbg/issue1659
Fix typo that broke compilation with DYNAMIC_ARCH and NO_AVX2
2018-07-02 17:48:19 +02:00
Martin Kroeker 9d15a3bd16 Fix typo that broke compilation with DYNAMIC_ARCH and NO_AVX2
fixes 1659
2018-07-02 14:40:41 +02:00
Martin Kroeker bbf2124970 set version number to 0.3.2.dev 2018-07-01 12:01:51 +02:00
Martin Kroeker 1392eba488 set version number to 0.3.2.dev 2018-07-01 12:01:16 +02:00
Martin Kroeker 61659f8765 Merge pull request #1648 from martin-frbg/nofort
Handle NOFORTRAN=0
2018-07-01 11:56:40 +02:00
Martin Kroeker 3d3c19717c Merge pull request #1655 from martin-frbg/issue1641
Fix apparent off-by-one error in calculation of MAX_ALLOCATING_THREADS
2018-07-01 08:41:22 +02:00
Martin Kroeker 24e344038d Merge pull request #1654 from martin-frbg/avx512check
Add compiler option to avx512 test and hide test output
2018-07-01 01:17:03 +02:00
Martin Kroeker 4e9c34018e Fix apparent off-by-one error in calculation of MAX_ALLOCATING_THREADS
fixes #1641
2018-06-30 23:57:50 +02:00
Martin Kroeker f5243e8e1f Add compiler option to avx512 test and hide test output 2018-06-30 23:47:44 +02:00
Martin Kroeker ba8388cee0 Merge pull request #1651 from martin-frbg/avx512-nodgemm
Disable the 16x2 DTRMM kernel on SkylakeX as well
2018-06-30 17:48:03 +02:00
Martin Kroeker 6e54b0a027 Disable the 16x2 DTRMM kernel on SkylakeX as well 2018-06-30 17:31:06 +02:00
Martin Kroeker 40c8cbc3bf Merge pull request #1650 from martin-frbg/avx512-nodgemm
Disable the AVX512 DGEMM kernel for now
2018-06-30 13:05:46 +02:00
Martin Kroeker d3c9eb4c7d Merge pull request #1639 from martin-frbg/dyn_list
Add DYNAMIC_LIST option for user-defined list of dynamic targets
2018-06-30 13:05:30 +02:00
Martin Kroeker f0a8dc2eec Disable the AVX512 DGEMM kernel for now
due to #1643
2018-06-30 11:34:48 +02:00
Martin Kroeker cc92257ea6 Update Makefile 2018-06-27 00:09:21 +02:00
Martin Kroeker 2aba1b1658 Merge branch 'develop' into nofort 2018-06-27 00:07:32 +02:00
Martin Kroeker 8396e9e777 Handle NOFORTRAN=0 2018-06-27 00:00:27 +02:00
Martin Kroeker bfad307ed7 Merge pull request #1647 from martin-frbg/armv7-dot
Remove premature exits from ARMV7 xdot codes
2018-06-26 22:27:30 +02:00
Martin Kroeker b83e4c60c7 Remove premature exit for INC_X or INC_Y zero 2018-06-26 20:46:42 +02:00
Martin Kroeker e344db269b Remove premature exit for INC_X or INC_Y zero 2018-06-26 20:45:57 +02:00
Martin Kroeker 545b82efd3 Remove premature exit for INC_X or INC_Y zero 2018-06-26 20:45:00 +02:00
Martin Kroeker e322a951fe Remove premature exit for INC_X or INC_Y zero 2018-06-26 20:44:13 +02:00
Martin Kroeker ff2f171036 Merge pull request #1644 from martin-frbg/revert-filterout
Revert changes to NOFORTRAN handling in Makefile
2018-06-26 10:15:15 +02:00
Martin Kroeker 092175cfec Revert changes to NOFORTRAN handling from 952541e 2018-06-26 08:09:52 +02:00
Martin Kroeker 750162a05f Try gradual fallback for cores not in the dynamic core list 2018-06-25 21:02:31 +02:00
Martin Kroeker e6d93f20f1 Merge pull request #2 from martin-frbg/develop
merge develop
2018-06-25 20:48:10 +02:00
Martin Kroeker c38c65eb65 Merge pull request #1 from xianyi/develop
Merge xianyi:develop into develop
2018-06-25 20:45:56 +02:00
Martin Kroeker ce3651516f Merge pull request #1642 from oon3m0oo/develop
Rewrite &= -> = and simplify the initial blocking phase.
2018-06-25 19:23:40 +02:00
Craig Donner 0144068537 Rewrite &= -> = and simplify the initial blocking phase. 2018-06-25 15:08:55 +01:00
Martin Kroeker 1833a67071 Add support for a user-defined list of dynamic targets 2018-06-23 19:42:15 +02:00
Martin Kroeker 0b2b83d9ed Add support for a user-defined list of dynamic targets 2018-06-23 19:41:32 +02:00
Martin Kroeker 62cf769aa6 Merge pull request #1638 from martin-frbg/issue1637
Expose the CBLAS interface to the IxAMIN functions and have make build it
2018-06-23 15:01:02 +02:00
Martin Kroeker eb71d61c7c Expose CBLAS interface to BLAS extensions iXamin 2018-06-23 13:31:09 +02:00
Martin Kroeker 9cf22b7d91 Build cblas_iXamin interfaces 2018-06-23 13:27:30 +02:00
Martin Kroeker cc66743b66 Merge pull request #1634 from oon3m0oo/develop
Fix data races reported by TSAN.
2018-06-21 21:01:03 +02:00
oon3m0oo 2aa0a5804e Use BLAS rather than CBLAS in test_fork.c (#1626)
This is handy for people not using lapack.
2018-06-21 18:47:45 +02:00
Craig Donner 28c28ed275 Fix data races reported by TSAN. 2018-06-21 16:41:02 +01:00
oon3m0oo a399d00425 Further improvements to memory.c. (#1625)
- Compiler TLS is now used only used when the compiler supports it
- If compiler TLS is unsupported, we use platform-specific TLS
- Only one variable (an index) is now in TLS
- We only access TLS once per alloc, and never when freeing
- Allocation / release info is now stored within the allocation itself, by
  over-allocating; this saves having external structures do the bookkeeping, and
  reduces some of the redundant data that was being stored (such as addresses)
- We never hit the alloc lock when not using SMP or when using OpenMP (that was
  my fault)
- Now that there are fewer tracking structures I think this is a bit easier to
  read than before
2018-06-20 22:04:03 +02:00
Martin Kroeker f66b9c8826 Merge pull request #1630 from martin-frbg/x86-march
Add -march=skylake-avx512 to flags if target is skylake x
2018-06-20 21:51:57 +02:00
Martin Kroeker 2946c46024 Merge pull request #1631 from oon3m0oo/stack
Avoid declaring arrays of size 0 when making large stack allocations.
2018-06-20 21:51:38 +02:00
Craig Donner 05978528c3 Avoid declaring arrays of size 0 when making large stack allocations. 2018-06-20 17:03:18 +01:00
Martin Kroeker ef6f0b645e Merge pull request #1629 from martin-frbg/issue1628
Make gfortran link libomp for clang in the tests; avoid two typical gotchas with NOFORTRAN
2018-06-20 16:41:13 +02:00
Martin Kroeker 0c5b7b400b Add -march=skylake-avx512 to flags if target is skylake x 2018-06-20 15:16:19 +02:00
Martin Kroeker 952541e840 Need to use filter-out to handle NOFORTRAN not set 2018-06-20 13:20:30 +02:00
Martin Kroeker 9369d3e6e5 Modify NOFORTRAN tests to always check the value; fix rewriting of NO_FORTRAN 2018-06-19 23:28:06 +02:00
Martin Kroeker 10b70c904d Handle erroneous user settings NOFORTRAN=0 and NO_FORTRAN 2018-06-19 20:53:19 +02:00
Martin Kroeker 6a5ab083b7 Handle special case of gfortran+clang+OpenMP 2018-06-19 20:47:33 +02:00
Martin Kroeker 1f9e4f3193 Handle special case of gfortran+clang+OpenMP 2018-06-19 20:46:36 +02:00
Martin Kroeker 5a6a2bed9a Merge pull request #1623 from fenrus75/fast-thread
Initialize only the required subset of the jobs array, fix barriers and improve switch ratio on SkylakeX and Haswell. For issue #1622
2018-06-18 09:02:40 +02:00
Martin Kroeker 2d8cc7193a Support upcoming Intel Cannon Lake CPUs as Skylake X (#1621)
* Support  upcoming Cannon Lake as Skylake X
2018-06-17 23:38:14 +02:00
Arjan van de Ven 2ddc96c9e5 make WMB / MB safer on x86-64
make it so that

if (foo)
	RMB;
else
	MB;

is always done correctly and without syntax surprises
2018-06-17 18:06:24 +00:00
Arjan van de Ven 7e39ffe113 On x86-64, make MB/WMB compiler barriers
Whie on x86(64) one does not normally need full memory barriers, it's
good practice to at least use compiler barriers for places where on other
architectures memory barriers are used; this prevents the compiler
from over-optimizing.
2018-06-17 17:53:15 +00:00
Arjan van de Ven 73de17664d Add missing barriers in gemm scheduler
a few places in the gemm scheduler code were missing barriers;
the code likely worked OK due to heavy use of volatile / _Atomic
but there's no reason to get this incorrect
2018-06-17 17:50:43 +00:00
Arjan van de Ven 6eb4b9ae7c Tune HASWELL SWITCH_RATIO as well
Similar to the SKYLAKEX patch, 32 seems to work best
(much better than 4 or 16)

Before (4)

   Matrix          SGEMM cycles    MPC                                   DGEMM cycles      MPC
  48 x 48               15554.3    7.2       0.2%                             30353.8      3.7       0.3%
  64 x 64               30346.8    8.7       1.6%                             63495.0      4.1      -0.1%
  65 x 65               81668.1    3.4    -123.3%                             82705.2      3.3     -21.2%
  80 x 80              105045.9    4.9     -95.5%                            115226.0      4.5      -2.2%
  96 x 96              152461.2    5.8     -74.3%                            148156.3      6.0      16.4%
 112 x 112             188505.2    7.5     -42.2%                            171187.3      8.2      36.4%
 128 x 128             257884.0    8.1     -39.5%                            224764.8      9.3      46.0%

Intermediate (16)

   Matrix          SGEMM cycles    MPC                                   DGEMM cycles      MPC
  48 x 48               15565.7    7.2       0.2%                             30378.9      3.7       0.2%
  64 x 64               30430.2    8.7       1.3%                             63046.4      4.2       0.6%
  65 x 65               27306.0   10.1      25.3%                             38879.2      7.1      43.0%
  80 x 80               51008.7   10.1       5.1%                             61007.6      8.4      45.9%
  96 x 96               70856.7   12.5      19.0%                             83403.1     10.6      53.0%
 112 x 112              84769.9   16.6      36.0%                             99920.1     14.1      62.9%
 128 x 128              84213.2   25.0      54.5%                            113024.2     18.6      72.8%

After (32)

   Matrix          SGEMM cycles    MPC                                   DGEMM cycles      MPC
  48 x 48               15537.3    7.2       0.3%                             30537.0      3.6      -0.3%
  64 x 64               30352.7    8.7       1.6%                             62597.8      4.2       1.3%
  65 x 65               36857.0    7.5      -0.8%                             56167.6      4.9      17.7%
  80 x 80               42552.6   12.1      20.8%                             69536.7      7.4      38.3%
  96 x 96               52101.5   17.1      40.5%                             91016.1      9.7      48.7%
 112 x 112              63853.7   22.1      51.8%                            110507.4     12.7      58.9%
 128 x 128              73966.1   28.4      60.0%                            163146.4     12.9      60.8%
2018-06-17 17:08:36 +00:00
Arjan van de Ven 5c6f008365 Tune param.h for SkylakeX
param.h defines a per-platform SWITCH_RATIO, which is used as a measure for how fine
grained the blocks for gemm need to be split up. Many platforms define this to 4.

The reality is that the gemm low level implementation for SkylakeX likes bigger blocks
due to the nature of SIMD... by tuning the SWITCH_RATIO to 32 the threading performance
improves significantly:

Before
   Matrix          SGEMM cycles    MPC                                   DGEMM cycles      MPC
  48 x 48               10756.0   10.5      -0.5%                             18296.7      6.1      -1.7%
  64 x 64               20490.0   12.9       1.4%                             40615.0      6.5       0.0%
  65 x 65               83528.3    3.3    -210.9%                             96319.0      2.9     -83.3%
  80 x 80              101453.5    5.1    -166.3%                            128021.7      4.0     -76.6%
  96 x 96              149795.1    5.9    -143.1%                            168059.4      5.3     -47.4%
 112 x 112             191481.2    7.3    -105.8%                            204165.0      6.9     -14.6%
 128 x 128             265019.2    7.9     -99.0%                            272006.4      7.7      -5.3%

After
   Matrix          SGEMM cycles    MPC                                   DGEMM cycles      MPC
  48 x 48               10666.3   10.6       0.4%                             18236.9      6.2      -1.4%
  64 x 64               20410.1   13.0       1.8%                             39925.8      6.6       1.7%
  65 x 65               34983.0    7.9     -30.2%                             51494.6      5.4       2.0%
  80 x 80               39769.1   13.0      -4.4%                             63805.2      8.1      12.0%
  96 x 96               45169.6   19.7      26.7%                             80065.8     11.1      29.8%
 112 x 112              57026.1   24.7      38.7%                             99535.5     14.2      44.1%
 128 x 128              64789.8   32.5      51.3%                            117407.2     17.9      54.6%

With this change, threading starts to be a win already at 96x96
2018-06-17 15:47:50 +00:00
Arjan van de Ven d148ec4ea1 Don't use _Atomic for jobs sometimes...
The use of _Atomic leads to really bad code generation in the compiler
(on x86, you get 2 "mfence" memory barriers around each access with gcc8, despite
x86 being ordered and cache coherent). But there's a fallback in the code that
just uses volatile which is more than plenty in practice.

If we're nervous about cross thread synchronization for these variables, we should
make the YIELD function be a compiler/memory barrier instead.

performance before (after last commit)

   Matrix          SGEMM cycles    MPC                                   DGEMM cycles      MPC
  48 x 48               10630.0   10.6       0.7%                             18112.8      6.2      -0.7%
  64 x 64               20374.8   13.0       1.9%                             40487.0      6.5       0.4%
  65 x 65              141955.2    1.9    -428.3%                            146708.8      1.9    -179.2%
  80 x 80              178921.1    2.9    -369.6%                            186032.7      2.8    -156.6%
  96 x 96              205436.2    4.3    -233.4%                            224513.1      3.9     -97.0%
 112 x 112             244408.2    5.8    -162.7%                            262158.7      5.4     -47.1%
 128 x 128             321334.5    6.5    -141.3%                            333829.0      6.3     -29.2%

Performance with this patch (roughly a 2x improvement):

   Matrix          SGEMM cycles    MPC                                   DGEMM cycles      MPC
  48 x 48               10756.0   10.5      -0.5%                             18296.7      6.1      -1.7%
  64 x 64               20490.0   12.9       1.4%                             40615.0      6.5       0.0%
  65 x 65               83528.3    3.3    -210.9%                             96319.0      2.9     -83.3%
  80 x 80              101453.5    5.1    -166.3%                            128021.7      4.0     -76.6%
  96 x 96              149795.1    5.9    -143.1%                            168059.4      5.3     -47.4%
 112 x 112             191481.2    7.3    -105.8%                            204165.0      6.9     -14.6%
 128 x 128             265019.2    7.9     -99.0%                            272006.4      7.7      -5.3%
2018-06-17 15:39:15 +00:00
Arjan van de Ven 9e162146a9 Only initialize the part of the jobs array that will get used
The jobs array is getting initialized in O(compiled cpus^2) complexity.
Distros and people with bigger systems will use pretty high values
(128 or 256 or more) for this value, leading to interesting bubbles
in performance.

Baseline (single threaded performance) gets roughly 13 - 15 multiplications per cycle
in the interesting range (threading kicks in at 65x65 mult by 65x65).
The hardware is capable of 32 multiplications per cycle theoretically.

   Matrix          SGEMM cycles    MPC                                   DGEMM cycles      MPC
  48 x 48               10703.9   10.6       0.0%                             17990.6      6.3       0.0%
  64 x 64               20778.4   12.8       0.0%                             40629.2      6.5       0.0%
  65 x 65               26869.9   10.3       0.0%                             52545.7      5.3       0.0%
  80 x 80               38104.5   13.5       0.0%                             72492.7      7.1       0.0%
  96 x 96               61626.4   14.4       0.0%                            113983.8      7.8       0.0%
 112 x 112              91803.8   15.3       0.0%                            180987.3      7.8       0.0%
 128 x 128             133161.4   15.8       0.0%                            258374.3      8.1       0.0%

When threading is turned on
TARGET=SKYLAKEX F_COMPILER=GFORTRAN  SHARED=1 DYNAMIC_THREADS=1 USE_OPENMP=0  NUM_THREADS=128

  Matrix          SGEMM cycles    MPC                                   DGEMM cycles      MPC
  48 x 48               10725.9   10.5      -0.2%                             18134.9      6.2      -0.8%
  64 x 64               20500.6   12.9       1.3%                             40929.1      6.5      -0.7%
  65 x 65             2040832.1    0.1   -7495.2%                           2097633.6      0.1   -3892.0%
  80 x 80             2063129.1    0.2   -5314.4%                           2119925.2      0.2   -2824.3%
  96 x 96             2070374.5    0.4   -3259.6%                           2173604.4      0.4   -1806.9%
 112 x 112            2111721.5    0.7   -2169.6%                           2263330.8      0.6   -1170.0%
 128 x 128            2276181.5    0.9   -1609.3%                           2377228.9      0.9    -820.1%

There is a deep deep cliff once you hit 65x65

With this patch

   Matrix          SGEMM cycles    MPC                                   DGEMM cycles      MPC
  48 x 48               10630.0   10.6       0.7%                             18112.8      6.2      -0.7%
  64 x 64               20374.8   13.0       1.9%                             40487.0      6.5       0.4%
  65 x 65              141955.2    1.9    -428.3%                            146708.8      1.9    -179.2%
  80 x 80              178921.1    2.9    -369.6%                            186032.7      2.8    -156.6%
  96 x 96              205436.2    4.3    -233.4%                            224513.1      3.9     -97.0%
 112 x 112             244408.2    5.8    -162.7%                            262158.7      5.4     -47.1%
 128 x 128             321334.5    6.5    -141.3%                            333829.0      6.3     -29.2%

The cliff is very significantly reduced.
(more to follow)
2018-06-17 15:32:03 +00:00
Martin Kroeker 47bf0dba8f Add build-time option for OMP scheduler; document MULTITHREAD_THRESHOLD range (#1620)
* Allow choosing the OpenMP scheduler and add range hint for GEMM_MULTITHREAD_THRESHOLD
* Amended description of GEMM_MULTITHREAD_THRESHOLD
to reflect #742 making it track floating point operations rather than matrix size
2018-06-15 11:25:05 +02:00
Martin Kroeker 12603b7dbb Merge pull request #1618 from oon3m0oo/less_locking
Remove the need for most locking in memory.c.
2018-06-15 00:10:29 +02:00
Craig Donner bf40f806ef Remove the need for most locking in memory.c.
Using thread local storage for tracking memory allocations means that threads
no longer have to lock at all when doing memory allocations / frees. This
particularly helps the gemm driver since it does an allocation per invocation.
Even without threading at all, this helps, since even calling a lock with
no contention has a cost:

Before this change, no threading:
```
----------------------------------------------------
Benchmark             Time           CPU Iterations
----------------------------------------------------
BM_SGEMM/4          102 ns        102 ns   13504412
BM_SGEMM/6          175 ns        175 ns    7997580
BM_SGEMM/8          205 ns        205 ns    6842073
BM_SGEMM/10         266 ns        266 ns    5294919
BM_SGEMM/16         478 ns        478 ns    2963441
BM_SGEMM/20         690 ns        690 ns    2144755
BM_SGEMM/32        1906 ns       1906 ns     716981
BM_SGEMM/40        2983 ns       2983 ns     473218
BM_SGEMM/64        9421 ns       9422 ns     148450
BM_SGEMM/72       12630 ns      12631 ns     112105
BM_SGEMM/80       15845 ns      15846 ns      89118
BM_SGEMM/90       25675 ns      25676 ns      54332
BM_SGEMM/100      29864 ns      29865 ns      47120
BM_SGEMM/112      37841 ns      37842 ns      36717
BM_SGEMM/128      56531 ns      56532 ns      25361
BM_SGEMM/140      75886 ns      75888 ns      18143
BM_SGEMM/150      98493 ns      98496 ns      14299
BM_SGEMM/160     102620 ns     102622 ns      13381
BM_SGEMM/170     135169 ns     135173 ns      10231
BM_SGEMM/180     146170 ns     146172 ns       9535
BM_SGEMM/189     190226 ns     190231 ns       7397
BM_SGEMM/200     194513 ns     194519 ns       7210
BM_SGEMM/256     396561 ns     396573 ns       3531
```
with this change:
```
----------------------------------------------------
Benchmark             Time           CPU Iterations
----------------------------------------------------
BM_SGEMM/4           95 ns         95 ns   14500387
BM_SGEMM/6          166 ns        166 ns    8381763
BM_SGEMM/8          196 ns        196 ns    7277044
BM_SGEMM/10         256 ns        256 ns    5515721
BM_SGEMM/16         463 ns        463 ns    3025197
BM_SGEMM/20         636 ns        636 ns    2070213
BM_SGEMM/32        1885 ns       1885 ns     739444
BM_SGEMM/40        2969 ns       2969 ns     472152
BM_SGEMM/64        9371 ns       9372 ns     148932
BM_SGEMM/72       12431 ns      12431 ns     112919
BM_SGEMM/80       15615 ns      15616 ns      89978
BM_SGEMM/90       25397 ns      25398 ns      55041
BM_SGEMM/100      29445 ns      29446 ns      47540
BM_SGEMM/112      37530 ns      37531 ns      37286
BM_SGEMM/128      55373 ns      55375 ns      25277
BM_SGEMM/140      76241 ns      76241 ns      18259
BM_SGEMM/150     102196 ns     102200 ns      13736
BM_SGEMM/160     101521 ns     101525 ns      13556
BM_SGEMM/170     136182 ns     136184 ns      10567
BM_SGEMM/180     146861 ns     146864 ns       9035
BM_SGEMM/189     192632 ns     192632 ns       7231
BM_SGEMM/200     198547 ns     198555 ns       6995
BM_SGEMM/256     392316 ns     392330 ns       3539
```

Before, when built with USE_THREAD=1, GEMM_MULTITHREAD_THRESHOLD = 4, the cost
of small matrix operations was overshadowed by thread locking (look smaller than
32) even when not explicitly spawning threads:
```
----------------------------------------------------
Benchmark             Time           CPU Iterations
----------------------------------------------------
BM_SGEMM/4          328 ns        328 ns    4170562
BM_SGEMM/6          396 ns        396 ns    3536400
BM_SGEMM/8          418 ns        418 ns    3330102
BM_SGEMM/10         491 ns        491 ns    2863047
BM_SGEMM/16         710 ns        710 ns    2028314
BM_SGEMM/20         871 ns        871 ns    1581546
BM_SGEMM/32        2132 ns       2132 ns     657089
BM_SGEMM/40        3197 ns       3196 ns     437969
BM_SGEMM/64        9645 ns       9645 ns     144987
BM_SGEMM/72       35064 ns      32881 ns      50264
BM_SGEMM/80       37661 ns      35787 ns      42080
BM_SGEMM/90       36507 ns      36077 ns      40091
BM_SGEMM/100      32513 ns      31850 ns      48607
BM_SGEMM/112      41742 ns      41207 ns      37273
BM_SGEMM/128      67211 ns      65095 ns      21933
BM_SGEMM/140      68263 ns      67943 ns      19245
BM_SGEMM/150     121854 ns     115439 ns      10660
BM_SGEMM/160     116826 ns     115539 ns      10000
BM_SGEMM/170     126566 ns     122798 ns      11960
BM_SGEMM/180     130088 ns     127292 ns      11503
BM_SGEMM/189     120309 ns     116634 ns      13162
BM_SGEMM/200     114559 ns     110993 ns      10000
BM_SGEMM/256     217063 ns     207806 ns       6417
```
and after, it's gone (note this includes my other change which reduces calls
to num_cpu_avail):
```
----------------------------------------------------
Benchmark             Time           CPU Iterations
----------------------------------------------------
BM_SGEMM/4           95 ns         95 ns   12347650
BM_SGEMM/6          166 ns        166 ns    8259683
BM_SGEMM/8          193 ns        193 ns    7162210
BM_SGEMM/10         258 ns        258 ns    5415657
BM_SGEMM/16         471 ns        471 ns    2981009
BM_SGEMM/20         666 ns        666 ns    2148002
BM_SGEMM/32        1903 ns       1903 ns     738245
BM_SGEMM/40        2969 ns       2969 ns     473239
BM_SGEMM/64        9440 ns       9440 ns     148442
BM_SGEMM/72       37239 ns      33330 ns      46813
BM_SGEMM/80       57350 ns      55949 ns      32251
BM_SGEMM/90       36275 ns      36249 ns      42259
BM_SGEMM/100      31111 ns      31008 ns      45270
BM_SGEMM/112      43782 ns      40912 ns      34749
BM_SGEMM/128      67375 ns      64406 ns      22443
BM_SGEMM/140      76389 ns      67003 ns      21430
BM_SGEMM/150      72952 ns      71830 ns      19793
BM_SGEMM/160      97039 ns      96858 ns      11498
BM_SGEMM/170     123272 ns     122007 ns      11855
BM_SGEMM/180     126828 ns     126505 ns      11567
BM_SGEMM/189     115179 ns     114665 ns      11044
BM_SGEMM/200      89289 ns      87259 ns      16147
BM_SGEMM/256     226252 ns     222677 ns       7375
```

I've also tested this with ThreadSanitizer and found no data races during
execution.  I'm not sure why 200 is always faster than it's neighbors, we must
be hitting some optimal cache size or something.
2018-06-14 16:54:58 +01:00
Martin Kroeker ed682a4a0c Merge pull request #1619 from martin-frbg/issue1580
Update OSX deployment target to 10.8
2018-06-14 17:48:51 +02:00
Martin Kroeker fcb77ab129 Update OSX deployment target to 10.8
fixes #1580
2018-06-14 16:57:58 +02:00
Martin Kroeker 26e1cfb653 Merge pull request #1607 from martin-frbg/dynarch
Move some x86_64 DYNAMIC_ARCH targets to new DYNAMIC_OLDER option
2018-06-14 16:52:55 +02:00
Martin Kroeker c628c6fa59 Merge pull request #1612 from oon3m0oo/cpus
Fixed a few more unnecessary calls to num_cpu_avail.
2018-06-14 16:51:31 +02:00
Martin Kroeker 67d81ab49d Merge pull request #1609 from martin-frbg/issue1529
Create OpenBLASConfig.cmake in cmake builds as well
2018-06-12 23:00:24 +02:00
Martin Kroeker 2f957947a6 Merge pull request #1613 from xianyi/revert-1600-noyield
Revert "Use usleep instead of sched_yield by default"
2018-06-11 17:14:49 +02:00
Martin Kroeker de8fff671d Revert "Use usleep instead of sched_yield by default" 2018-06-11 17:05:27 +02:00
Martin Kroeker 6f71c0fce4 Return a somewhat sane default value for L2 cache size if cpuid retur… (#1611)
* Return a somewhat sane default value for L2 cache size if cpuid returned something unexpected

Fixes #1610, the KVM hypervisor on Google Chromebooks returning zero for CPUID  0x80000006, causing DYNAMIC_ARCH
builds of OpenBLAS to hang
2018-06-11 13:26:19 +02:00
Craig Donner c2545b0fd6 Fixed a few more unnecessary calls to num_cpu_avail.
I don't have as many benchmarks for these as for gemm, but it should still
make a difference for small matrices.
2018-06-11 10:17:16 +01:00
Martin Kroeker e65f451409 include CMakePackageConfigHelpers 2018-06-10 15:09:43 +02:00
Martin Kroeker 02634b549b Add template for OpenBLASConfig.cmake 2018-06-10 09:25:46 +02:00
Martin Kroeker 0bea6bb9e7 Create OpenBLASConfig.cmake from cmake as well 2018-06-10 09:24:37 +02:00
Martin Kroeker 3313e4b946 Merge pull request #1608 from martin-frbg/issue874
Enable parallel make on MS Windows by default
2018-06-09 19:57:33 +02:00
Martin Kroeker e9cd11768c Enable parallel make on MS Windows by default
fixes #874
2018-06-09 17:54:36 +02:00
Martin Kroeker 63f7395fb4 Move some DYNAMIC_ARCH targets to new DYNAMIC_OLDER option 2018-06-09 16:31:38 +02:00
Martin Kroeker 1cbd8f3ae4 Move some DYNAMIC_ARCH targets to new DYNAMIC_OLDER option 2018-06-09 16:30:46 +02:00
Martin Kroeker 6c2d90ba77 Move some DYNAMIC_ARCH targets to new DYNAMIC_OLDER option 2018-06-09 16:29:17 +02:00
Martin Kroeker 0297b3211a Merge pull request #1605 from oon3m0oo/develop
Improve performance of GEMM for small matrices when SMP is defined.
2018-06-09 12:42:34 +02:00
Craig Donner 66316b9f4c Improve performance of GEMM for small matrices when SMP is defined.
Always checking num_cpu_avail() regardless of whether threading will actually
be used adds noticeable overhead for small matrices.  Most other uses of
num_cpu_avail() do so only if threading will be used, so do the same here.
2018-06-07 15:29:13 +01:00
Martin Kroeker 6adc4b7b36 Merge pull request #1601 from martin-frbg/zaxpy
Use a single thread for small input size in zaxpy
2018-06-07 14:09:58 +02:00
Martin Kroeker 2ade0ef085 Merge pull request #1600 from martin-frbg/noyield
Use usleep instead of sched_yield by default
2018-06-07 12:42:00 +02:00
Martin Kroeker e8880c1699 Use a single thread for small input size
copies daxpy improvement from #27, see #1560
2018-06-07 10:26:55 +02:00
Martin Kroeker ed7c4a043b Use usleep instead of sched_yield by default
sched_yield only burns cpu cycles, fixes #900,  see also #923, #1560
2018-06-07 10:18:26 +02:00
Martin Kroeker cf234a0561 Merge pull request #1589 from fenrus75/skylakex
Initial support for SkylakeX / AVX512
2018-06-06 22:07:09 +02:00
Martin Kroeker ae2a33128b Merge pull request #1599 from martin-frbg/c_check_avx512
Improved AVX512 test case for c_check
2018-06-06 18:42:42 +02:00
Martin Kroeker e4718b1fee Better AVX512 test case 2018-06-06 16:51:30 +02:00
Martin Kroeker 9b87b64262 Improve AVX512 testcase
clang 3.4 managed to accept the original test code, only to fail on the actual Skylake asm later
2018-06-06 16:49:00 +02:00
Martin Kroeker 0218b884c1 Merge pull request #1598 from martin-frbg/issue1593-2
Restore _Atomic define before stdatomic.h for old gcc
2018-06-06 12:48:26 +02:00
Martin Kroeker 83da278093 Update common.h 2018-06-06 09:27:49 +02:00
Martin Kroeker 358d4df2bd Merge branch 'develop' into issue1593-2 2018-06-06 09:21:41 +02:00
Martin Kroeker 06d43760e4 Restore _Atomic define before stdatomic.h for old gcc
see #1593
2018-06-06 09:18:10 +02:00
Martin Kroeker a4af8861ff Merge pull request #1597 from martin-frbg/cmake-avx512
Check build system support for AVX512 instructions
2018-06-06 07:22:20 +02:00
Martin Kroeker 7fb62aed7e Check build system support for AVX512 instructions 2018-06-05 23:29:33 +02:00
Martin Kroeker f6021c798d Re-enable QUIET_MAKE 2018-06-05 19:09:38 +02:00
Martin Kroeker e8002536ec disable quiet_make for the moment 2018-06-05 18:23:01 +02:00
Martin Kroeker ce6317f6c0 Merge pull request #1594 from martin-frbg/issue1593
Fix inverted condition in _Atomic declaration
2018-06-05 16:02:51 +02:00
Martin Kroeker 15a78d6b66 export NO_AVX512 setting 2018-06-05 15:58:34 +02:00
Martin Kroeker 354a976a59 Fix inverted condition in _Atomic declaration
fixes #1593
2018-06-05 10:31:34 +02:00
Martin Kroeker 38ad05bd04 Extend loop range to find SkylakeX in force_coretype 2018-06-05 10:26:49 +02:00
Martin Kroeker b7feded85a Propagate NO_AVX512 via CCOMMON_OPT 2018-06-05 10:24:05 +02:00
Martin Kroeker dc9fe05ab5 Update cpuid_x86.c 2018-06-04 17:10:19 +02:00
Martin Kroeker 8be027e4c6 Update dynamic.c 2018-06-04 14:36:39 +02:00
Martin Kroeker ac7b6e3e9a Fix misplaced endif 2018-06-04 08:23:40 +02:00
Martin Kroeker fc66a0ec0b Merge pull request #1590 from martin-frbg/avx512_check
Disable AVX512 (Skylake X) support if the build system is too old
2018-06-04 08:18:38 +02:00
Arjan van de Ven 89372e0993 Use AVX512 also for DGEMM
this required switching to the generic gemm_beta code (which is faster anyway on SKX)
for both DGEMM and SGEMM

Performance for the not-retuned version is in the 30% range
2018-06-03 22:17:27 +00:00
Martin Kroeker ef626c6824 typo fix 2018-06-04 00:13:19 +02:00
Martin Kroeker 83fec56a3f Disable AVX512 (Skylake X) support if the build system is too old 2018-06-04 00:01:11 +02:00
Martin Kroeker 5a51cf4576 Separate Skylake X from Skylake 2018-06-03 23:41:33 +02:00
Martin Kroeker 5a92b311e0 Separate Skylake X from Skylake 2018-06-03 23:29:07 +02:00
Martin Kroeker a7d0f49cec Add SKYLAKEX to DYNAMIC_CORE list only if AVX512 is available 2018-06-03 23:13:25 +02:00
Martin Kroeker f1fb9a4745 Propagate NO_AVX512 if needed 2018-06-03 13:48:27 +02:00
Martin Kroeker 0023515733 Typo fix (misplaced parenthesis) 2018-06-03 13:22:59 +02:00
Arjan van de Ven 99c7bba8e4 Initial support for SkylakeX / AVX512
This patch adds the basic infrastructure for adding the SkylakeX (Intel Skylake server)
target. The SkylakeX target will use the AVX512 (AVX512VL level) instruction set,
which brings 2 basic things:
1) 512 bit wide SIMD (2x width of AVX2)
2) 32 SIMD registers (2x the number on AVX2)

This initial patch only contains a trivial transofrmation of the Haswell SGEMM kernel
to AVX512VL; more will follow later but this patch aims to get the infrastructure
in place for this "later".

Full performance tuning has not been done yet; with more registers and wider SIMD
it's in theory possible to retune the kernels but even without that there's an
interesting enough performance increase (30-40% range) with just this change.
2018-06-03 07:58:52 +00:00
Martin Kroeker 36c4523d85 Merge pull request #1587 from matthew-brett/fix-compile-error-early-glibc
Revert "take out unused variables"
2018-06-02 10:02:38 +02:00
Matthew Brett a8002e283a Revert "take out unused variables"
This reverts commit e5752ff9b3.

The variables i and n are used in the `#if !__GLIBC_PREREQ(2, 7)`
branch.

Closes gh-1586.
2018-06-01 23:20:00 +01:00
Martin Kroeker 401adddb2b Merge pull request #1585 from martin-frbg/lapack-253
Fixes from Lapack-Reference PR 253
2018-06-01 18:59:33 +02:00
Martin Kroeker c5b13d4e10 Fixes from netlib PR 253 2018-06-01 15:14:45 +02:00
Martin Kroeker 677e42d7b0 Fixes from netlib PR 253
When minimal workspace is given in ?hesv_aa, ?sysv_aa, ?hesv_aa_2stage, ?sysv_aa_2stage, now no error is given
Quick return for ?laqr1
2018-06-01 15:12:59 +02:00
Martin Kroeker e2a8c35e5a Fixes from netlib PR253
LAPACKE interfaces for Aasen's functions now call ?sytrf_aa and ?hetrf_aa instead of ?sytrf and ?hetrf
2018-06-01 15:08:14 +02:00
Martin Kroeker 1a49fb1c05 Merge pull request #1584 from martin-frbg/issue1503
Work around name clash with Windows10's winnt.h
2018-05-31 21:56:04 +02:00
Martin Kroeker 8562d5787a Merge pull request #1583 from martin-frbg/issue1575
Handle INCX=0,INCY=0 case
2018-05-31 21:55:26 +02:00
Martin Kroeker 93f1eb09c3 Merge pull request #1582 from martin-frbg/develop-031
Update version number on the develop branch to 0.3.1.dev
2018-05-31 21:55:07 +02:00
Martin Kroeker c90bbda3df Merge pull request #1581 from martin-frbg/issue1574-2
Fix paths to LIN and EIG tests
2018-05-31 21:54:45 +02:00
Martin Kroeker 7df8c4f76f typo fix 2018-05-31 17:23:08 +02:00
Martin Kroeker 2fc748bf72 Restore optimized swap kernel now that we have a proper fix 2018-05-31 13:41:12 +02:00
Martin Kroeker a91f1587b9 Work around name clash with Windows10's winnt.h
fixes #1503
2018-05-31 13:26:00 +02:00
Martin Kroeker d1b7be14aa Handle INCX=0,INCY=0 case
Fixes #1575 (sswap/dswap failing the swap utest on x86) as suggested by atsampson.
2018-05-31 12:52:04 +02:00
Martin Kroeker b491b10057 Update version to 0.3.1.dev 2018-05-31 12:44:36 +02:00
Martin Kroeker 5fae96fb70 Update version to 0.3.1.dev 2018-05-31 12:43:45 +02:00
Martin Kroeker a7dbd4c57d Fix paths to LIN and EIG tests
should fix 1574
2018-05-31 11:19:33 +02:00
Martin Kroeker 2cae104b5e Merge pull request #1579 from martin-frbg/issue1574
Adapt lapack-test and blas-test to changes in netlib directory layout
2018-05-29 22:02:06 +02:00
Martin Kroeker 908d40be71 Adapt lapack-test and blas-test to changes in netlib directory layout
partial fix for #1574 - the problem with lapack_testing.py looks like an upstream bug
2018-05-29 14:27:46 +02:00
Zhang Xianyi 43e592ceb3 Add -lm for Android.
Conflicts:
	exports/Makefile
2018-05-24 21:02:42 +08:00
Martin Kroeker f0f27868d8 Merge pull request #1572 from martin-frbg/issue1571
Use the new zrot.c on POWER8 for crot as well
2018-05-23 22:55:37 +02:00
Martin Kroeker 961d25e9c7 Use the new zrot.c on POWER8 for crot as well
fixes #1571 (the old zrot.S assembly does not handle incx=0 correctly)
2018-05-23 22:54:39 +02:00
Martin Kroeker f5959f2543 Merge pull request #1567 from martin-frbg/mipstrmm
Revert " Switch mips32 target to USE_TRMM to fix complex TRMM"
2018-05-17 20:50:23 +02:00
Martin Kroeker 82012b960b Revert " Switch mips32 target to USE_TRMM to fix complex TRMM"
... as it was just a silly workaround for the issue seen in #1563, caused by #1419
2018-05-17 20:30:03 +02:00
Martin Kroeker 8dd3515fa2 Merge pull request #1565 from martin-frbg/mipstypo
Remove extraneous brace from previous commit of mips dsdot fix
2018-05-17 20:22:58 +02:00
Martin Kroeker 95f7f0229c Remove extraneous brace from previous commit 2018-05-17 18:43:59 +02:00
Martin Kroeker 5082fe4306 Merge pull request #1564 from martin-frbg/issue1563
Revert changes from PR#1419
2018-05-17 14:04:13 +02:00
Martin Kroeker 7a7619af6d Revert changes from PR#1419
at least one of these changes apparently is an oversimplification, leading to TRMM breakage on some platforms as observed in #1563
2018-05-17 11:40:08 +02:00
Martin Kroeker 9a400b7014 Merge pull request #1562 from martin-frbg/issue1561
Use correct data type for initializers of v2f64, v4f32
2018-05-15 17:46:09 +02:00
Martin Kroeker 893b535540 Use correct data type for initializers of v2f64, v4f32
Fixes #1561
2018-05-15 14:42:12 +02:00
Martin Kroeker 6791294312 Merge pull request #1559 from martin-frbg/buildconf
Add build-time configuration options to pkgconfig file
2018-05-14 18:49:53 +02:00
Martin Kroeker ddb8b124de Merge pull request #1558 from martin-frbg/instpc
Overwrite any pre-existing openblas.pc rather than append to it
2018-05-14 17:38:12 +02:00
Martin Kroeker 191746c493 Merge pull request #1557 from martin-frbg/getconfig
Add threading and OpenMP information to output
2018-05-14 17:37:55 +02:00
Martin Kroeker eb9b021d38 Add build-time configuration options to pkgconfig file 2018-05-14 00:10:15 +02:00
Martin Kroeker 7d7564568c Add build-time configuration options to pkgconfig file 2018-05-14 00:09:35 +02:00
Martin Kroeker a07843bc93 Overwrite any pre-existing openblas.pc rather than append to it 2018-05-12 22:11:27 +02:00
Martin Kroeker 41ae8e8d67 Add threading and OpenMP information to output
For #1416 and #1529, more information about the options OpenBLAS was built with is needed. Additionally we may want to add this data to the openblas.pc file (but not all projects use pkgconfig, and as far as I am aware the cmake module for accessing it does not make such "private" declarations available)
2018-05-12 12:11:38 +02:00
Zhang Xianyi 9c1aa0b0fe Merge pull request #1556 from WestAlgo/develop
move _Atomic define to common.h
2018-05-11 17:02:47 +08:00
zhiyong.dang 53457f222f move _Atomic define to common.h 2018-05-11 00:13:16 -07:00
Zhang Xianyi 458e3af5b1 Merge pull request #1555 from WestAlgo/develop
Change _STDC_VERSION__ to __STDC_VERSION__
2018-05-11 12:25:24 +08:00
Zhiyong Dang 3716267124 Change _STDC_VERSION__ to __STDC_VERSION__
Change-Id: Id3fa4e8d9eedd4ef7230df69b611e7f397301a42
2018-05-11 12:15:08 +08:00
Zhang Xianyi 50acc40613 Merge pull request #1536 from WestAlgo/develop
Fix race condition in blas_server_omp.c
2018-05-11 10:09:14 +08:00
Martin Kroeker c720f1f019 Merge pull request #1554 from martin-frbg/lapack-249
LAPACKE fixes from lapack PR249
2018-05-10 15:32:08 +02:00
Martin Kroeker d7d950fcf2 LAPACKE fixes from lapack PR249
Copied from Reference-LAPACK/lapack#249, this fixes out-of-bounds memory accesses
in the nancheck calls of the LAPACKE lacgv, lassq,larfg,larfb,larfx and mtr functions
2018-05-10 13:15:42 +02:00
Martin Kroeker 12398e53ce Merge pull request #1553 from martin-frbg/ifort-openmpflag
Change -openmp to -fopenmp for ifort entry as well
2018-05-09 14:39:52 +02:00
Martin Kroeker 193f835662 Change -openmp to -fopenmp for ifort entry as well 2018-05-09 12:34:09 +02:00
Martin Kroeker 7e3151ead7 Merge pull request #1551 from martin-frbg/f_check_fix
Fixes for ifort 2018
2018-05-09 09:02:52 +02:00
Martin Kroeker e3a069f108 Merge pull request #1550 from martin-frbg/ifort-openmpflag
Update compiler flag for openmp use with ICC
2018-05-09 09:02:38 +02:00
Martin Kroeker 6fff8c626a Merge pull request #1549 from martin-frbg/fix_ompcheck
Drop C-style "L" suffx from OPENMP version number tests in the LAPACK source
2018-05-08 23:52:55 +02:00
Martin Kroeker d2b9389f1b Fixes for ifort 2018
1. the already deprecated -openmp option was removed in 2018, switch to -fopenmp
2. add leading blank in search for "zho_ge__" symbol to work around misleading tags in the 2018 assembly
Expected to fix #1548
2018-05-08 21:55:37 +02:00
Martin Kroeker 65b8a5c5d8 Update compiler flag for openmp use with ICC
The deprecated -openmp option was finally removed in favor of -qopenmp or -fopenmp, picking the latter to stay compatible with Intel compiler versions before 2015 (when -q options were introduced). Fixes #1546
2018-05-08 21:47:10 +02:00
Martin Kroeker 9795adc7ef Drop C-style "L" suffix from OPENMP version number in check 2018-05-08 21:39:42 +02:00
Martin Kroeker 1a8e487c4a Drop C-style "L" suffix from OPENMP version number in check 2018-05-08 21:38:25 +02:00
Martin Kroeker 5966fd52a2 Drop C-style "L" suffix from OPENMP version number in check 2018-05-08 21:36:56 +02:00
Martin Kroeker dbafe6357b Merge pull request #1543 from martin-frbg/mips32
Fix MIPS32 build and add MIPS 1004K cpu (MT7621 SOC)
2018-05-02 22:47:45 +02:00
Martin Kroeker 71051259e0 Restore compiler options for mips P5600 target 2018-05-02 20:37:06 +02:00
Martin Kroeker 73cc321190 Add MIPS 1004K target 2018-05-02 20:27:56 +02:00
Martin Kroeker 018f2dad27 Switch mips32 target to USE_TRMM to fix complex TRMM 2018-05-02 20:25:32 +02:00
Martin Kroeker 9d5098dbc9 Add MIPS 1004K target (Mediatek MT7621 SOC) 2018-05-02 20:20:44 +02:00
Martin Kroeker d94d7baf7e Add mips32r2 api target 2018-05-02 20:17:26 +02:00
Martin Kroeker 3af1b5c805 Make cpuid_mips compile again and add 1004K cpu 2018-05-02 20:12:25 +02:00
Martin Kroeker 88e224f4c0 Merge pull request #1542 from martin-frbg/quickdiv64
Avoid out-of-bounds accesses in blas_quickdivide on big X86 systems
2018-05-02 18:11:50 +02:00
Martin Kroeker d0c0506588 Omit the divide table overflow check on small systems 2018-05-02 14:44:50 +02:00
Martin Kroeker e93355e5e1 Omit the table overflow check when building for small systems 2018-05-02 14:43:08 +02:00
Martin Kroeker c1eb06e102 Update common_x86_64.h 2018-04-29 14:40:12 +02:00
Martin Kroeker 8145ecd70b Avoid out-of-bounds reads from blas_quick_divide_table on big systems 2018-04-29 14:38:55 +02:00
Martin Kroeker 26ce518d46 Avoid out of bounds reads from blas_quick_divide_table on big systems
Should fix #1541
2018-04-29 14:34:33 +02:00
Martin Kroeker 1d27fa8507 Merge pull request #1539 from martin-frbg/ztrmv-1332
Disable multithreading in ztrmv
2018-04-27 23:10:21 +02:00
Martin Kroeker 802cf6b22d Merge pull request #1486 from martin-frbg/atomic
Use _Atomic instead of volatile for thread safety where C11 is supported
2018-04-27 23:09:57 +02:00
Martin Kroeker 894433a7c7 Update Makefile.rule 2018-04-27 12:08:06 +02:00
Zhiyong Dang 1b83341d19 Fix race condition in blas_server_omp.c
Change-Id: Ic896276cd073d6b41930c7c5a29d66348cd1725d
2018-04-27 17:00:42 +08:00
Martin Kroeker 954f1832de Merge pull request #1540 from martin-frbg/mips32-zasum
Fix typo in MIPS P5600 complex ASUM code selection
2018-04-25 23:23:00 +02:00
Martin Kroeker 941ad280a8 Fix typo in MIPS P5600 complex ASUM code selection 2018-04-25 22:50:10 +02:00
Martin Kroeker a8ed428bab Disable multithreading in ztrmv
BLAS-Tester shows that the same problem exists as with DTRMV (issue #1332)
2018-04-25 22:35:46 +02:00
Martin Kroeker 1da365312a Merge pull request #1538 from martin-frbg/arm7utest
Fix handling of zero INCX, INCY in ArmV7 AXPY and ROT
2018-04-25 08:38:58 +02:00
Martin Kroeker 2d0929fa7c Move the test for zero incx,incy in ARMV7 ROT
to pass the related utest (see #1469)
2018-04-24 22:43:00 +02:00
Martin Kroeker 125343cc88 Drop test for zero incx,incy in armv7 AXPY
...to pass the related utest (see #1469)
2018-04-24 22:39:50 +02:00
Martin Kroeker 8a3b6fa108 Use generic zrot.c on ppc64/POWER6 to work around utest failure from … (#1535)
* Use generic C implementation of zrot on ppc64/POWER6 to work around utest failure from #1469
2018-04-23 19:05:49 +02:00
Martin Kroeker 78694f1b7e Merge pull request #1534 from xianyi/revert-1333-haswell32
Revert "Fix 32bit HASWELL builds"
2018-04-22 23:34:17 +02:00
Martin Kroeker 9c5518319a Revert "Fix 32bit HASWELL builds" 2018-04-22 20:20:04 +02:00
Martin Kroeker 86f49c529d Merge pull request #1532 from martin-frbg/utest-cblas
Do not try to build the fork utest when NO_CBLAS=1
2018-04-20 23:44:15 +02:00
Martin Kroeker 625c74a38f fork utest depends on CBLAS 2018-04-20 15:43:59 +02:00
Martin Kroeker 5fcaca6438 fork utest depends on CBLAS 2018-04-20 15:42:13 +02:00
Martin Kroeker 4fcdd24459 Merge pull request #1530 from ashwinyes/develop_20180419_Tx2AutoDetect
ARM64: Enable Auto Detection of ThunderX2T99
2018-04-19 14:10:57 +02:00
Ashwin Sekhar T K 68a3c4fca6 ARM64: Enable Auto Detection of ThunderX2T99 2018-04-19 09:05:25 +00:00
Martin Kroeker 0c4718c57a Merge pull request #1523 from martin-frbg/utest_waith
Include sys/types.h for proper typedefs related to wait()
2018-04-15 13:09:30 +02:00
Martin Kroeker f29389c7ac Merge pull request #1520 from martin-frbg/cpucounts
Catch invalid cpu count returned by CPU_COUNT_S
2018-04-14 22:24:34 +02:00
Martin Kroeker 734d7c6a93 Include sys/types.h for proper typedefs related to wait()
Should fix #1519
2018-04-14 18:59:46 +02:00
Martin Kroeker 7c861605b2 Catch invalid cpu count returned by CPU_COUNT_S
mips32 was seen to return zero here, driving nthreads to zero with subsequent fpe in blas_quickdivide
2018-04-14 18:29:10 +02:00
Martin Kroeker 2ca0faf495 Merge pull request #1515 from martin-frbg/mipsdot
Correct precision of mips dsdot
2018-04-11 08:21:25 +02:00
Martin Kroeker 0fe434598b Fix precision of mips dsdot 2018-04-10 23:30:59 +02:00
Martin Kroeker 15c437e092 Merge pull request #1512 from ararslan/aa/travis-macos-2
Add macOS to the Travis testing matrix: Take 2!
2018-04-07 23:31:26 +02:00
Alex Arslan b966bd79d5 Add a BINARY=32 build to macOS 2018-04-07 12:29:57 -07:00
Alex Arslan 2e988dbf35 Add macOS to the Travis testing matrix 2018-04-07 10:56:34 -07:00
Martin Kroeker be6090d396 Merge pull request #1511 from xianyi/revert-1510-aa/travis-macos
Revert "Add macOS to the Travis testing matrix"
2018-04-07 13:29:31 +02:00
Martin Kroeker daae8fd197 Revert "Add macOS to the Travis testing matrix" 2018-04-07 13:27:24 +02:00
Martin Kroeker 20c6c38e51 Merge branch 'develop' into atomic 2018-04-07 12:09:39 +02:00
Martin Kroeker a1fb7670f7 Merge pull request #1510 from ararslan/aa/travis-macos
Add macOS to the Travis testing matrix
2018-04-07 12:07:12 +02:00
Martin Kroeker 6c99c97489 Merge pull request #1509 from ararslan/aa/dragonfly
Add DragonFly to exports/Makefile
2018-04-07 12:06:57 +02:00
Alex Arslan 6a0930560e Add macOS to the Travis testing matrix 2018-04-06 17:53:58 -07:00
Alex Arslan 24f8d5b624 Add DragonFly to exports/Makefile
Its exclusion was an oversight on my part.
2018-04-06 17:30:10 -07:00
Martin Kroeker 77b4dbd53b Merge pull request #1506 from martin-frbg/issue1497
Fix thread races and infinite looping on systems with many cpus
2018-04-05 23:46:36 +02:00
Martin Kroeker bc4c3bca01 Merge pull request #1507 from martin-frbg/threads_usage
Underline importance of NUM_THREADS setting for BUFFER allocation
2018-04-05 08:54:07 +02:00
Martin Kroeker 6b0a9d135c Merge pull request #1508 from ararslan/aa/wording
Minor changes to wording and formatting in the README
2018-04-05 08:53:38 +02:00
Alex Arslan 137ccd9dd9 Minor changes to wording and formatting in the README
The wording in some places is not grammatically correct. This change
also provides minor adjustments to the Markdown formatting which provide
modest improvements to readability.
2018-04-04 14:30:32 -07:00
Martin Kroeker 84923dedb7 Merge pull request #1505 from ararslan/aa/compiler
Compile with cc rather than gcc whenever possible
2018-04-04 22:45:33 +02:00
Martin Kroeker 8ec28ff461 Remove unguarded use of _Atomic and fix tabbing 2018-04-04 22:40:30 +02:00
Martin Kroeker ca8ca796d3 Underline importance of NUM_THREADS setting for BUFFER allocation
following augray's suggestion from #1451, and incorporating ashwinyes' comments from #1141 on the importance of NUM_THREADS even for single-threaded builds.
2018-04-04 22:26:51 +02:00
Alex Arslan 8f811a9312 Reinstate macOS logic 2018-04-04 11:41:45 -07:00
Alex Arslan 36a17536ca Compile with cc rather than gcc whenever possible 2018-04-04 11:26:54 -07:00
Martin Kroeker bb9876db33 Fix thread races and infinite looping on systems with many cpus
On systems with more than 64 cpus, blas_quickdivide will sometimes return zero which creates bogus workloads when used for the stride calculation. This then leads to threads spinning incessantly waiting for a status change that never happens, as seen in #1497.
This patch also fixes several data races that were found by helgrind and/or tsan while debugging the issue.
2018-04-04 18:16:52 +02:00
Martin Kroeker d636b418af Merge pull request #1504 from ararslan/aa/openbsd
Allow building on OpenBSD
2018-04-04 15:26:46 +02:00
Martin Kroeker a460c92577 Merge pull request #1501 from martin-frbg/issue875
Add workaround for old gcc and clang versions
2018-04-04 15:26:21 +02:00
Alex Arslan 33f838393c Add OpenBSD and DragonFly to community supported platforms 2018-04-03 16:42:01 -07:00
Alex Arslan a41d241a0e Add support for DragonFly BSD 2018-04-03 16:39:29 -07:00
Alex Arslan 8da6b6ae52 Allow building on OpenBSD
With this change, OpenBLAS builds and all tests pass on OpenBSD 6.2
using Clang. Tested on x86-64 only, with and without DYNAMIC_ARCH=1.
2018-04-02 10:48:22 -07:00
Martin Kroeker 01c4b82f04 Update memory.c 2018-03-31 22:32:06 +02:00
Martin Kroeker 93db123f7e Update memory.c 2018-03-29 13:13:49 +02:00
Martin Kroeker 752fdb5dd8 Add workaround for old gcc and clang versions
Old gcc and clang do not handle constructor arguments, finally fix #875 as discussed there, using the fedora patch
2018-03-29 11:56:56 +02:00
Martin Kroeker 07ed01e97f Merge pull request #1500 from martin-frbg/issue1474
Correct index variables used in MFlops calculation
2018-03-28 09:15:34 +02:00
Martin Kroeker 35c5a32309 Correct index variables used in MFlops calculation
Fixes #1474
2018-03-27 21:52:29 +02:00
Martin Kroeker c7b55b6082 Merge pull request #1499 from quickwritereader/develop
Implemented missing vsx simd  kernels for power8 blas1/2 double. z13 modifications
2018-03-27 21:43:23 +02:00
Martin Kroeker 840e01061f Merge pull request #1491 from martin-frbg/ddot_mt
Add multithreading support for Haswell DDOT
2018-03-27 21:43:05 +02:00
QWR QWR 28ca97015d power8:Added initial zgemv_(t|n) ,i(d|z)amax,i(d|z)amin,dgemv_t(transposed),zrot
z13: improved zgemv_(t|n)_4,zscal,zaxpy
2018-03-27 14:54:41 +00:00
Martin Kroeker 73c5ca74fa Merge pull request #1495 from martin-frbg/aff
Disable CPU affinity by default again
2018-03-19 18:03:25 +01:00
Martin Kroeker e453555d97 Disable CPU affinity by default again
This setting must have been changed unintentionally by my PR #1214 (probably leftover from unrelated tests)
2018-03-19 18:02:23 +01:00
Martin Kroeker 6a6ffaff1e Merge pull request #1494 from martin-frbg/x86_dsdot
Use generic/dot.c instead of the inferior arm/dot.c for x86 DSDOT
2018-03-17 15:26:47 +01:00
Martin Kroeker 28ac9ea5a6 Use generic/dot.c instead of the inferior arm/dot.c for x86 DSDOT
to resolve dsdot utest failure seen in #1492
2018-03-17 13:49:15 +01:00
Martin Kroeker a55694dd5b Declare dot_compute static to avoid conflicts in multiarch builds 2018-03-16 22:23:36 +01:00
Martin Kroeker 85a41e9cdb Add multithreading support for Haswell DDOT
copied from ashwinyes' implementation in dot_thunderx2t99.c
2018-03-16 16:58:47 +01:00
Martin Kroeker 40160ff3c1 Use _Atomic instead of volatile for thread safety where C11 is supported 2018-03-10 00:15:44 +01:00
Martin Kroeker 6a99fcce94 Use _Atomic instead of volatile for thread safety where C11 is supported
Suggested by dodomorandi in #660
2018-03-10 00:03:49 +01:00
Martin Kroeker 2c7392f07b Merge pull request #1482 from martin-frbg/haswell_axpy
Re-enable DAXPY AVX microkernels  for x86_64
2018-03-04 22:21:18 +01:00
Martin Kroeker 81215711a2 Re-enable DAXPY microkernels for x86_64
as the inaccuracies seen in the original testcase for #1332 appear to be due to an artefact that amplifies the very small rounding differences between FMA and discrete multiply+add
2018-03-04 19:37:03 +01:00
Martin Kroeker 809fd0d451 Rewrite ROTMG to address cases not covered by the netlib algorithm (#1480)
* Rewrite ROTMG based on the new implementation in GONUM based on the algorithm proposed by Tim Hopkins, see issue 1452 for the reference
* Correct ROTMG utest for issue1452 and add another from gonum, also correct transposition of expected and observed values in error messages
2018-03-04 17:39:56 +01:00
Martin Kroeker 72e65157df Merge pull request #1481 from martin-frbg/utest-fixup
Fix transposition of expected and computed values in error message
2018-03-03 22:43:56 +01:00
Martin Kroeker 69a8aa6de2 Fix transposition of expected and computed values in error message 2018-03-03 18:01:51 +01:00
Martin Kroeker 0ab5bf1746 Merge pull request #1476 from xsacha/patch-1
Fix CMake cross-compiling
2018-02-28 18:47:57 +01:00
Martin Kroeker 22167170b3 Merge pull request #1477 from quickwritereader/develop
Power8 blas3 copy-pack routines
2018-02-28 18:46:54 +01:00
Martin Kroeker 69d9f36ff4 Merge pull request #1468 from martin-frbg/martin-frbg-patch-1
Limit the additional locking from PRs 1052,1299 to non-OpenMP cases
2018-02-28 18:40:31 +01:00
Sacha f81815e48a Fix CMake cross-compiling
Without specifying thread count, NUM_THREADS would not be defined and CMake would fail.
This is because core count cannot be determined when cross-compiling.
2018-02-28 10:25:25 +10:00
Martin Kroeker 5f855d965d Merge pull request #1475 from ashwinyes/develop_20180227_utest_dsdot_fixes
ARM64: Fix utest dsdot errors
2018-02-27 14:04:16 +01:00
Ashwin Sekhar T K fa9ca65c0e ARM64: Fix utest dsdot errors 2018-02-27 10:47:55 +00:00
Martin Kroeker 719b68f077 Merge pull request #1473 from martin-frbg/p2align
Replace .align with .p2aligns in dscal.c and the Nehalem microkernels as well
2018-02-27 08:28:20 +01:00
Martin Kroeker fe9f15f2d8 Merge pull request #1472 from martin-frbg/utest-fixes
Fix limited DSDOT precision on arm,aarch64 and zarch
2018-02-26 22:48:07 +01:00
Martin Kroeker 497f0c3d8a Replace .align with .p2align in the Nehalem microkernels 2018-02-26 20:58:33 +01:00
Martin Kroeker ea37db828e Convert .align to .p2align for OSX compatibility 2018-02-26 20:48:03 +01:00
Martin Kroeker e6a0a3de73 Merge pull request #1471 from martin-frbg/p2align
Use .p2align instead of .align for portability on Haswell and Sandybridge
2018-02-26 12:28:01 +01:00
Martin Kroeker 6e70287776 Use generic/dot.c for DSDOT on ARMV5 and above
The default arm/dot.c is less precise when used for DSDOT, as shown by utest
2018-02-25 19:57:23 +01:00
Martin Kroeker 58f236ad73 Use generic/dot.c for DSDOT on zarch 2018-02-25 19:52:14 +01:00
Martin Kroeker e207107150 Use generic/dot.c for DSDOT on z13
The implementation in arm/dot.c has lower precision, as shown by the utest for dsdot.
2018-02-25 19:51:25 +01:00
Martin Kroeker c9d408064a Use dot.S also for DSDOT on CORTEXA57 2018-02-25 19:48:09 +01:00
Martin Kroeker 288d1a3f6e Use dot.S also for DSDOT on ARMV8 2018-02-25 19:45:16 +01:00
Martin Kroeker 7c1925acec Use .p2align instead of .align for compatibility on Sandybridge as well 2018-02-24 19:43:15 +01:00
Martin Kroeker 2359c7c1a9 Use .p2align instead of .align for portability
The OSX assembler apparently mishandles the argument to decimal .align, leading to a significant loss of performance 
as observed in #730, #901 and most recently #1470
2018-02-24 17:50:13 +01:00
Martin Kroeker 7646974227 Limit the additional locking from PRs 1052,1299 to non-OpenMP multithreading 2018-02-21 11:45:33 +01:00
Martin Kroeker e3a80e6aa8 Merge pull request #1466 from xianyi/revert-1464-issue1461
Revert "Add locks only for non-OPENMP multithreading"
2018-02-20 17:17:38 +01:00
Martin Kroeker 8866e393a2 Revert "Add locks only for non-OPENMP multithreading" 2018-02-20 17:17:12 +01:00
Martin Kroeker 9e87e6f3d8 Merge pull request #1464 from martin-frbg/issue1461
Add locks only for non-OPENMP multithreading
2018-02-20 14:08:24 +01:00
Martin Kroeker 3119b2ab4c Add locks only for non-OPENMP multithreading
to migitate performance problems caused by #1052 and #1299 as seen in #1461
2018-02-20 12:17:18 +01:00
Martin Kroeker e7366a4161 Restore the remaining utests (#1462)
* Restore the remaining utests

* Try fork test on Cygwin and Linux only, it hangs on at least ARMv8/Android as well

* Use generic sswap/dswap kernels for NEHALEM 32bit to fix fault found by the restored swap utest

* Disable zdotu test for MS cl to work around runtime error -1073741819 on AppVeyor for now
(probably coding error in the initialization of the complex numbers or wrong choice of zdotu API)
2018-02-20 10:07:17 +01:00
Martin Kroeker 6c0f79b787 Merge pull request #1463 from martin-frbg/rotmg2
Fix wrong conditionals in scaling loops of rotmg and update BLAS1 tests from netlib
2018-02-18 15:12:42 +01:00
Martin Kroeker 7cd7acf71e Merge pull request #1460 from martin-frbg/issue1425
Revert insidious suppression of the -fopenmp flag in the LAPACK subtree
2018-02-18 15:12:20 +01:00
Martin Kroeker 72f14a0363 Fix conditionals in the rescaling against GAMSQ 2018-02-18 12:54:52 +01:00
Martin Kroeker 53026dc63a Update single and double precision BLAS1 tests from LAPACK 3.8.0
adding tests for SROTMG, SROTM, SDSDOT, DROTMG, DROTM, DSDOT
2018-02-18 12:44:14 +01:00
Martin Kroeker 798f1595d5 Fix condition in both second scaling loops 2018-02-18 12:37:09 +01:00
the mslm 2c0a008281 dgemm_ncopy_4_ save/restore 2018-02-18 01:30:17 +00:00
the mslm c5425daa6b power8 ?gemm_tcopy save/restore 2018-02-16 23:36:46 +00:00
Martin Kroeker eaab622f03 Make "OMP task depend" sections conditional on OpenMP4, not just OpenMP
To allow compiling with gcc versions older than 4.9
2018-02-14 22:58:14 +01:00
Martin Kroeker 3cda1ce50a Revert insiduous suppression of the -fopenmp flag in the LAPACK subtree
This was added in #1046 citing a problem with mingw, but in effect it quietly reduces thread safety on all non-Windows platforms (while -fopenmp is already disabled for Windows builds through the toplevel Makefile.system). Removing the filter fixes #1425
2018-02-13 22:44:45 +01:00
Martin Kroeker 0391c07b17 Merge pull request #1457 from martin-frbg/issue1456
test_fork is not meant to be run (nor expected to work) with OpenMP
2018-02-11 23:13:18 +01:00
Martin Kroeker f4b095b1bb test_fork is not meant (nor expected) to be run with OpenMP
Fixes 1456
2018-02-11 20:58:27 +01:00
Martin Kroeker 6940c59a88 Merge pull request #1454 from martin-frbg/issue1452
Keep the flag handling separate from the scaling loops in rotmg
2018-02-11 20:48:04 +01:00
Martin Kroeker 42514fa24c Merge pull request #1449 from martin-frbg/armv8
Enable assembly kernels for the generic ARMV8 target and treat CortexA53,A72 as A57
2018-02-11 20:47:49 +01:00
Martin Kroeker 650077074a Add tests for rotmg 2018-02-10 14:18:21 +01:00
Martin Kroeker fe16a94fc2 Add rotmg tests for CMAKE MSVC+CLANG build 2018-02-10 14:17:41 +01:00
Martin Kroeker 632b8e0f05 Merge current Makefile from develop 2018-02-10 14:10:55 +01:00
Martin Kroeker a1bc0fcf07 Resurrect utest for rotmg and add testcase for issue 1452 2018-02-10 12:48:03 +01:00
Martin Kroeker 0464aa6784 Remove debug printfs 2018-02-09 23:06:50 +01:00
Martin Kroeker 55840f0bc9 Keep the flag handling separate from the scaling loops
Fixes #1452 and is more in line with how ATLAS does it. The earlier fix from #356 only moved the bug elsewhere, but we will never want the iterative rescaling to change the dflag setting and variable associations with each cycle.
2018-02-09 23:00:03 +01:00
Martin Kroeker a7485f3222 Merge pull request #1453 from martin-frbg/netlib228
Remove spurious EXTERNAL reference
2018-02-08 22:39:24 +01:00
Martin Kroeker d31f62cf02 Merge pull request #1450 from embray/cygwin/forking
Fix issues with forking on Cygwin
2018-02-08 22:39:11 +01:00
Martin Kroeker 150c7294a6 Remove spurious EXTERNAL reference
From Reference-LAPACK issue 228, remove spurious EXTERNAL reference to unused and nonexistent function xLACGV that could cause linking problems.
2018-02-08 14:57:13 +01:00
Erik M. Bray 8f5f614615 On Cygwin use mmap instead of Windows native allocation functions, which are not fork-safe. 2018-02-06 12:23:27 +01:00
Erik M. Bray f5fc109fbd Perform blas_thread_shutdown with pthread_atfork() on Cygwin
Even if we're directly using the win32 threading driver and not pthreads,
pthread_atfork still works fine to register a pre-fork handler, and is
necessary to restore the threading server to a pre-initialized state.
2018-02-06 12:23:27 +01:00
Erik M. Bray ce2028b425 Rewrite this test to work with ctest and re-enable it on the appropriate platforms (including Cygwin, which has fork()) 2018-02-06 12:23:27 +01:00
Martin Kroeker b47e6822aa Enable most assembly kernels in the generic ARMV8 target
ref #1439
2018-02-06 11:42:58 +01:00
Martin Kroeker 0ae5e14923 Detect CORTEX A53 and A72 as CORTEXA57 2018-02-06 11:38:18 +01:00
Martin Kroeker 95f2cea45b Merge pull request #1447 from martin-frbg/sparcfix
Generate CHAR_CORENAME for SPARC
2018-02-02 09:02:49 +01:00
Martin Kroeker e3c50643bb Fix my copypaste blunder with get_corename 2018-02-01 22:06:04 +01:00
Martin Kroeker efa84afd00 Use get_corename for SPARC as well 2018-02-01 18:20:38 +01:00
Martin Kroeker d1b512a01a Return a corename for SPARC 2018-02-01 18:15:15 +01:00
Martin Kroeker e0b02789ff Merge pull request #1445 from quickwritereader/develop
small fix inside ifdef z13mvc . (z13mvc code is not used in production)
2018-02-01 08:28:53 +01:00
Abdelrauf 60596a1abc Merge branch 'develop' into develop 2018-01-31 16:17:04 -08:00
Abdelrauf afd514c25d small fix inside ifdef z13mvc . (z13mvc code is not used in production) 2018-01-31 18:30:59 -05:00
Martin Kroeker 25c7e3992f Merge pull request #1443 from martin-frbg/sparcfix
Also #define SPARC in config.h when autodetecting
2018-01-31 23:49:01 +01:00
Martin Kroeker f45776ec1f Merge pull request #1440 from quickwritereader/develop
small corrections
2018-01-31 23:48:47 +01:00
Martin Kroeker e388459a27 Merge pull request #1419 from brada4/develop
Initialize unitialized values for repeated calls
2018-01-31 23:48:34 +01:00
Martin Kroeker 0ac824f6a5 Also #define SPARC in config.h when autodetecting
Fixes #1442
2018-01-31 22:02:00 +01:00
Abdelrauf f653e7a18d small fix
small fix inside ifdef z13mvc . (z13mvc code is not used in production)
2018-01-31 07:49:38 -08:00
Martin Kroeker 09e397e4f1 Merge pull request #1434 from xoviat/flang-wall
CMake: Remove unused wall option when FC=flang
2018-01-27 14:26:49 +01:00
Martin Kroeker 913e9546c0 Merge pull request #1436 from martin-frbg/cmaketrmm
Make USE_TRMM depend on TARGET_CORE not TARGET
2018-01-27 10:04:39 +01:00
the mslm f946a89432 zscal (case: real alpha=0 ) mikrokernel shift&mem fix , da_i as input reg. small typo fixes 2018-01-26 19:25:27 -08:00
Martin Kroeker 485df77612 Make USE_TRMM depend on TARGET_CORE not TARGET
Fixes #1432 (and possibly other DTRMM-related failures on Haswell and related architectures when built with cmake)
2018-01-26 23:20:00 +01:00
xoviat 038bfbb86c CMake: Remove unused wall option when FC=flang 2018-01-26 14:09:48 -06:00
Martin Kroeker 114fc0bae3 Merge pull request #1429 from martin-frbg/override_omp
When forcing USE_THREAD to zero, override USE_OPENMP as well
2018-01-24 09:02:55 +01:00
Martin Kroeker 4a4f6658de When forcing USE_THREAD=0, override USE_OPENMP as well
This avoids an error exit a few lines down as USE_THREAD=0 conflicts with USE_OPENMP=1
2018-01-23 21:33:21 +01:00
Martin Kroeker 0cf8046324 Merge pull request #1417 from xoviat/openblas-library-name
CMake: Use the correct library name on windows
2018-01-20 18:20:42 +01:00
Martin Kroeker e4c71a799a Merge pull request #1426 from quickwritereader/develop
(Z13 ) Blas1 mikrokernels can be inlined by gcc. Refactoring,fixes,tunings
2018-01-20 17:34:54 +01:00
Andrew e5752ff9b3 take out unused variables 2018-01-20 11:42:31 +01:00
the mslm 2619ad7ea5 Blas1 mikrokernels can be inlined by gcc. Refactoring ( symbolic operan
names). Some fixes and tunings
2018-01-19 19:24:35 -08:00
Andrew e5cc3d72c0 core.IdenticalExpr clang501 checker 2018-01-19 23:17:43 +01:00
Andrew 4938faa822 core.IdenticalExpr clang501 checker 2018-01-19 23:15:58 +01:00
Andrew 9fa986337d add missing brackets to silence indentation warnings gcc721 2018-01-19 23:11:12 +01:00
Andrew 8a0b086b28 add missing bracket for old glibc (cppcheck) 2018-01-12 22:35:48 +01:00
Andrew 3eed97f6b9 Initialize values to silence cppcheck 2018-01-12 22:35:00 +01:00
Andrew 13e137fbc9 Initialize uninitialized variables (cppcheck) 2018-01-12 22:33:41 +01:00
xoviat b3f24c9664 CMake: Use the correct library name on windows
FindBLAS and FindLAPACK don't correctly detect the OpenBLAS library
name on windows. I am unsure why this was originally set this way, but
it has caused me some trouble.
2018-01-11 11:34:53 -06:00
Martin Kroeker 3d23f45107 Merge pull request #1415 from quickwritereader/develop
(Z systems Z13) small fixes, some (i(dz)amin,i(dz)amax,(dz)dot,(dz)asum) mikrokernels…
2018-01-11 08:35:02 +01:00
Abdelrauf 87669d1c0a small fixes, some (i(dz)amin,i(dz)amax,(dz)dot,(dz)asum) mikrokernels can be inlined 2018-01-10 20:36:53 -05:00
Martin Kroeker 42285d8e70 Merge pull request #1410 from brada4/develop
Address warnings #1357
2018-01-06 20:02:46 +01:00
Martin Kroeker a76d1bb9e6 Merge pull request #1411 from christoph-conrads/fix-pkgconfig-path-escapes
Make: escape paths to pkg-config file
2018-01-06 20:01:35 +01:00
Christoph Conrads 7c7e2d9dc6 Make: escape paths to pkg-config file
Add double quotes around the path to the pkg-config file so that a path
containing whitespace does not break the build.
2018-01-05 17:08:55 -05:00
Andrew d602b99386 LAPACK helpers in C that need care too 2018-01-02 14:38:50 +01:00
Andrew 8aafa0473c address last warnings as seen by gcc7 2018-01-01 20:57:12 +01:00
Andrew 11a627c54e remove surplus parentheses to silence clang5 2018-01-01 20:56:26 +01:00
Andrew 4d0b005e5b Eliminate remaining unused results in kernels (clang5 analyzer) 2018-01-01 20:54:39 +01:00
Martin Kroeker b81656936f Merge pull request #1409 from martin-frbg/issue1292-2
Tag %1 and %2 as both input and output operands
2017-12-31 20:18:48 +01:00
Martin Kroeker b973990df2 Tag %1 and %2 as both input and output operands
fix from #1292 extended to the other gemv microkernels
2017-12-31 18:03:36 +01:00
Martin Kroeker 8fef2414b5 Merge pull request #1408 from xoviat/flang-ninja
Appveyor: speed up fortran builds
2017-12-30 14:52:21 +01:00
Martin Kroeker 1e31124eb0 Merge pull request #1406 from martin-frbg/issue1292
Tag %1 and %2 as both input and output
2017-12-30 14:52:03 +01:00
Martin Kroeker cc9500db41 Merge pull request #1403 from brada4/develop
Address few more warnings
2017-12-30 14:51:34 +01:00
xoviat b0652184ae Appveyor: enable building fortran with ninja 2017-12-29 19:58:35 -06:00
Martin Kroeker 723f396a20 Tag %1 and %2 as both input and output
The inline assembly modifies its input operands, so mark them as output to avoid surprises with optimization. Fixes #1292
2017-12-29 23:56:41 +01:00
Andrew 03e5ff0687 initialize potentially unitialized variables (clang5) 2017-12-26 09:24:24 +01:00
Andrew 47deec2c1a fix couple of dead assignment warnings 2017-12-22 00:56:35 +01:00
Andrew bfc2a88594 remove unused buffer 2017-12-22 00:55:40 +01:00
Martin Kroeker d741fc13d8 Merge pull request #1399 from martin-frbg/issue1398
Fix LAPACKE build problems with both cmake and make
2017-12-21 23:36:52 +01:00
Martin Kroeker 374260027d Add conditionals around ar calls for optional modules
The macOS ar aborts when it gets called with no input, see #1398
2017-12-21 20:42:30 +01:00
Martin Kroeker 599de9e598 Restore LAPACKE files for Xgeqpf, Xggsvd and Xggsvp
These were inadvertently dropped from the list in my PR #1095
2017-12-21 19:43:09 +01:00
Martin Kroeker 893bd14e92 Merge pull request #1393 from martin-frbg/daxpybug
Retire Piledriver/Steamroller/Excavator daxpy microkernels as well
2017-12-13 20:27:14 +01:00
Martin Kroeker 43c0622e7b Retire Piledriver/Steamroller/Excavator daxpy microkernels as well
related to issue #1332
2017-12-13 18:40:39 +01:00
Martin Kroeker 6aba7b66ce Merge pull request #1390 from martin-frbg/daxpybug
Use Sandybridge daxpy kernel on Haswell and Zen for now
2017-12-10 21:46:36 +01:00
Martin Kroeker 0623636c98 Use Sandybridge daxpy kernel on Haswell and Zen for now
The testcase from #1332 exposes a problem in daxpy_microk_haswell-2.c that is not seen with
any of the other Intel x86_64 microkernels.
2017-12-10 19:24:31 +01:00
Martin Kroeker 177b78c8b4 Issue1388 (#1389)
* Calculation of chunk range limits was ignoring num_cpu

bug introduced by me in #1262 - should fix #1388

* Calculation of range limits was ignoring num_cpu

bug introduced by me in #1262

* Calculation of chunk range limits was ignoring num_cpu

bug introduced by me in #1262

* Calculation of chunk range limits was ignoring num_cpu

bug introduced by me in #1262

* Calculation of chunk range limits was ignoring num_cpu

bug introduced by me in #1262

* Calculation of chunk range limits was ignoring num_cpu

bug introduced by me in #1262
2017-12-09 22:29:03 +01:00
Andrew 281a2b952f warning cleanup (#1380)
* dead increments in driver/level2

* dead increments in kernel/generic

* part dead increments in kernel/x86_64
2017-12-05 19:54:10 +01:00
Martin Kroeker c49c6b237d Merge pull request #1382 from martin-frbg/dtrmv-1332
Work around errors in multithreaded dtrmv
2017-12-05 19:53:23 +01:00
Martin Kroeker e2469a9ebc Merge pull request #1386 from martin-frbg/bignuma
Limit MAX_CPU to 1024 for now
2017-12-05 19:52:52 +01:00
Martin Kroeker 5b71f3a8e4 Merge pull request #1387 from martin-frbg/cmakeandroid
Explicitly link against libm on Android with cmake as well
2017-12-05 19:52:03 +01:00
Martin Kroeker 9381ac2748 Explicitly link against libm on Android with cmake as well
Patch from #1384
2017-12-05 13:02:48 +01:00
Martin Kroeker 28ae3ca76f Limit MAX_CPU to 1024 for now
Some Linux distributions (notably SuSE) have raised CPU_SETSIZE to 4096, apparently disregarding API limitations.
From #1348, the highest value to survive array initialization (on a desktop system) is 3232, and 1024 - which is the 
more usual CPU_SETSIZE limit, was demonstrated to work fine on an actual bignuma system.
2017-12-05 12:54:15 +01:00
Martin Kroeker b414283f48 Disable gemv unrolling
as a (hopefully temporary) workaround for #1332
2017-12-03 22:41:54 +01:00
Martin Kroeker 38763ec4f3 Disable multithreading for trmv
as a (hopefully temporary) workaround for #1332
2017-12-03 22:40:54 +01:00
Martin Kroeker 452fbef0bf Merge pull request #1381 from martin-frbg/ctest-warnings
Fix compiler warnings in ctest
2017-12-03 21:35:20 +01:00
Martin Kroeker 8c8313983b Fix compiler warnings in ctest
Various fixes for const correctness, stray tab characters and unused labels
2017-12-03 18:19:30 +01:00
Martin Kroeker 881a50c093 Merge pull request #1379 from martin-frbg/warnfix
Work around compiler warnings for unused variables
2017-12-03 13:04:02 +01:00
Martin Kroeker 8213385ab8 Work around compiler warnings for unused variables in the generic zgemm3m_Xcopy kernels 2017-12-02 22:51:58 +01:00
Martin Kroeker bede1c4fb4 Merge pull request #1372 from martin-frbg/param
Correct zgeadd_k prototype
2017-12-02 16:49:47 +01:00
Martin Kroeker 1d2da67841 Prefix make jobs with travis_wait (#1378)
* Prefix make with travis_wait to prevent it getting killed for producing no output

* Extend travis_wait to 30mins for the windows build

* Trying 45 mins wait time

* Increase travis_wait time to 45 minutes for linux builds as well
2017-12-02 12:59:27 +01:00
Martin Kroeker 0dc291d3fa Merge pull request #1377 from isuruf/threads
Allow overriding NUM_THREADS in cmake
2017-12-01 16:22:35 +01:00
Isuru Fernando e0ddd7d124 Allow overriding NUM_THREADS 2017-12-01 01:42:45 -06:00
Martin Kroeker adf4316f0e Merge pull request #1376 from xoviat/patch-2
[appveyor] fix test directory
2017-12-01 08:11:12 +01:00
xoviat 7fce11a5b8 [appveyor] fix test directory 2017-11-30 16:31:09 -06:00
Martin Kroeker c40f01ccea Merge pull request #1375 from xoviat/patch-1
[appveyor] Use out-of-tree build and cache
2017-11-30 22:43:54 +01:00
xoviat c567e34e6b [appveyor] fix syntax 2017-11-30 15:33:32 -06:00
xoviat c917278d23 [appveyor] Use out-of-tree build and cache 2017-11-30 15:30:10 -06:00
Martin Kroeker 0639ed1258 Merge pull request #1373 from mc10/patch-1
README: Use the SVG Travis badge
2017-11-30 12:54:52 +01:00
Kevin Ji f017e169dc README: Use the SVG Travis badge 2017-11-29 15:21:12 -08:00
Martin Kroeker 7e860acd38 Correct zgeadd_k prototype 2017-11-29 19:57:35 +01:00
Martin Kroeker db00a51e6b Merge pull request #1371 from martin-frbg/develop
Add trivially optimized DSDOT for POWER8
2017-11-29 19:55:21 +01:00
martin 7a4b3cfbf8 Add trivially optimized DSDOT for POWER8 2017-11-28 18:38:07 +01:00
Martin Kroeker 6c77b5f267 Merge pull request #1369 from martin-frbg/dsdot
Add optimized dsdot to all other x86_64 kernels that use sdot.c
2017-11-28 18:15:31 +01:00
Martin Kroeker d8b3c3c7db Merge pull request #1368 from brada4/develop
Eliminate warnings
2017-11-28 18:15:04 +01:00
Martin Kroeker beb18492fd Merge pull request #1366 from martin-frbg/develop
Update LAPACK to 3.8.0
2017-11-26 19:12:00 +01:00
Andrew 441a9c8385 more dead increments clang4 scan-build deadcode.deadstores 2017-11-26 17:24:08 +01:00
Andrew 1236dbe5a6 Eliminate 2-8 dead increments code 2017-11-26 13:26:11 +01:00
Andrew ef95cd471f elminate unread variable, after reiteration 3 of them (clang4) 2017-11-25 02:54:37 +01:00
Martin Kroeker c92cd6d162 Add trivially optimized dsdot based on sdot 2017-11-24 20:05:27 +01:00
Martin Kroeker cae5d9a20b Add trivially optimized dsdot based on sdot 2017-11-24 20:04:29 +01:00
Martin Kroeker 3d891c3106 Add trivially optimized dsdot based on sdot 2017-11-24 20:03:40 +01:00
Martin Kroeker 4fbdcfa823 Add trivially optimized dsdot based on sdot 2017-11-24 20:02:28 +01:00
Martin Kroeker 1bb6a96ebc Add trivially optimized dsdot based on sdot 2017-11-24 20:01:42 +01:00
Martin Kroeker 6bd163f37a Add trivially optimized dsdot based on sdot 2017-11-24 20:00:23 +01:00
Martin Kroeker f0333333d1 Add trivially optimized dsdot based on sdot 2017-11-24 19:59:28 +01:00
Andrew e14d50d86e eliminate Wunused-const gcc7 warning 2017-11-24 19:13:24 +01:00
Andrew e89b979b2c fix spurious compiler warning fix (no code change) 2017-11-24 18:39:04 +01:00
Andrew 7e9b29b9b8 fix spurious compiler warning (no code change) 2017-11-24 18:36:37 +01:00
martin 5056a044b2 fix location of lapacke_nancheck 2017-11-24 09:15:20 +01:00
martin 4054d32def update cmake files 2017-11-24 08:15:40 +01:00
martin 2d52f0f4c3 update cmakefiles for lapack 3.8.0 2017-11-23 21:22:01 +01:00
martin 3be5c3d343 Update LAPACK to 3.8.0 2017-11-23 18:13:35 +01:00
Martin Kroeker b18730f9e1 Merge pull request #1365 from xoviat/patch-1
[appveyor] use cmake from conda forge
2017-11-22 21:13:41 +01:00
xoviat be9e63c021 [appveyor] use cmake from conda forge 2017-11-21 18:44:02 -06:00
Martin Kroeker 21844238ca Merge pull request #1364 from martin-frbg/shmem-init
Handle shmem init failures in cpu affinity setup code
2017-11-19 12:50:16 +01:00
Martin Kroeker 07e7c36dac Handle shmem init failures in cpu affinity setup code
Failures to obtain or attach shared memory segments would lead to an exit without explanation of the exact cause.
This change introduces a more verbose error message and tries to make the code continue without setting cpu affinity.
Fixes #1351
2017-11-18 23:57:44 +01:00
Martin Kroeker 9251a2efde Merge pull request #1359 from brada4/develop
Eliminate mode variable where not needed in syrk interface
2017-11-18 23:47:17 +01:00
Martin Kroeker 06adbf2e41 Merge pull request #1347 from martin-frbg/issue1322
Change CBLAS complex functions to take void pointers
2017-11-18 23:46:57 +01:00
Martin Kroeker b46e2b57cc Make return parameter of cblas_Xdotc_sub, cblas_Xdotu_sub a void pointer as well 2017-11-18 20:28:02 +01:00
Martin Kroeker 3ce401f51b Make last parameter of cblas_Xdotc_sub/cblas_Xdotu_sub a void pointer as well 2017-11-18 18:58:40 +01:00
Martin Kroeker 1c9f4422b5 Fix declaration of cblas_Xdotc_sub and cblas_Xdotu_sub
last parameter of cblas_(c,z)dotc_sub and cblas_(c,z)dotu_sub should be void* as well
2017-11-18 18:56:30 +01:00
Andrew 27575d200a Eliminate mode variable where not needed 2017-11-15 15:32:38 +01:00
Martin Kroeker 6157d0902a Merge pull request #1358 from martin-frbg/unused_vars
Clean up spurious unused variables in the kernels
2017-11-15 11:31:43 +01:00
Martin Kroeker 3fea849bbf Remove unused variables from Haswell dtrmm and Bulldozer dtrsm 2017-11-14 23:35:10 +01:00
Martin Kroeker 8f177621bc Remove unused variables at0...at3 from ?symv_U 2017-11-14 23:32:25 +01:00
Martin Kroeker 5f402b7759 Remove unused (loop?) variable j from the gemv_n_4 implementations 2017-11-14 23:29:42 +01:00
Martin Kroeker 65bf0a343c Remove unused variable btpr 2017-11-14 23:25:50 +01:00
Martin Kroeker acf3d34bc5 Silence an unused variable warning with a cast
l2 cache size is not universally needed to assign default unrolling limits, but neither putting its declaration inside an ifdef nor cloning it into all ifdef sections that need it really makes sense here.
2017-11-14 23:23:44 +01:00
Martin Kroeker 8e75f7dcb4 Merge pull request #1353 from xoviat/patch-1
[appveyor] use flang from conda-forge
2017-11-10 22:16:31 +01:00
Martin Kroeker bd3546704c Merge pull request #1356 from martin-frbg/lapack-issue196
Break out of potentially infinite rescaling loop after 1000 iterations
2017-11-10 22:15:27 +01:00
Martin Kroeker 2df1e3372d Break out of potentially infinite rescaling loop after 1000 iterations
Inf values in the input vector will survive rescaling, causing an infinite loop. The value of 1000 is arbitrarily chosen as a large but finite value with the intention to never interfere with regular calculations.
2017-11-10 20:02:21 +01:00
Martin Kroeker 4271b2b158 Merge pull request #1354 from martin-frbg/shmem
Try to handle shmget or shmat failing
2017-11-10 09:11:03 +01:00
Martin Kroeker 148493df89 Merge branch 'develop' into shmem 2017-11-09 23:25:15 +01:00
Martin Kroeker 415555a9c1 Merge branch 'develop' into shmem 2017-11-09 23:20:54 +01:00
Martin Kroeker 2a6fef9a55 Try to handle shmget or shmat failing
also replaces one verbatim sched_yield with the YIELDING macro for consistency as suggested in #1351
2017-11-09 23:16:13 +01:00
xoviat 307305aeb5 [appeyor] use flang from conda-forge
This flang will be updated in the future. We leave cmake because it's
not yet released with fortran support
2017-11-09 15:10:02 -06:00
Martin Kroeker cc26cdce0c Merge pull request #1352 from martin-frbg/issue1351
Output an error message when shmat() fails
2017-11-09 21:08:16 +01:00
Martin Kroeker d8576826c4 Output an error message when shmat() fails
Observed in #1351 with SELinux as the likely culprit. Without the message, the user saw a segfault with no apparent reason
2017-11-09 17:31:44 +01:00
Martin Kroeker c6968edec4 Merge pull request #1350 from insertinterestingnamehere/flang
WIP: Support for Flang on Windows
2017-11-08 11:40:13 +01:00
Isuru Fernando 9268314290 Fix gensymbol script 2017-11-06 21:12:38 -06:00
Ian Henriksen 3ace0fda3f Merge pull request #1 from xoviat/patch-1
[appveyor] fixes
2017-11-06 15:17:24 -06:00
xoviat 3cfc64404a [appveyor] fixes 2017-11-06 15:05:20 -06:00
Ian Henriksen 72956e8950 Build MATGEN LAPACK routines by default when building with CMake. 2017-11-06 14:47:27 -06:00
Ian Henriksen 505dc08635 Update lapacke.cmake with routines added in LAPACK 3.7.0. 2017-11-06 14:43:33 -06:00
Ian Henriksen 61587b0670 Update lapack.cmake with additional routines from LAPACK version 3.7.0. 2017-11-06 14:41:02 -06:00
Ian Henriksen 632fc75d77 Allow using compilers other than gfortran in conjunction with
MSVC or clang-cl.
2017-11-06 14:39:12 -06:00
Martin Kroeker 2c222f1faa Modify complex CBLAS functions to take void pointers
Modify complex CBLAS functions to take void pointers instead of float or double arguments (to bring the prototypes in line with netlib and other implementations' cblas.h)
2017-11-05 15:53:14 +01:00
Martin Kroeker 66ac898f64 Change prototypes of all complex functions to use void*
Change prototypes of complex functions to use void pointers like the other implementations of CBLAS
2017-11-05 15:42:33 +01:00
Martin Kroeker ab87ee6b48 Merge pull request #1329 from martin-frbg/dsdot
(Trivial) optimized dsdot implementation for HASWELL
2017-10-25 19:13:38 +02:00
Martin Kroeker a07807caac Eliminate loop code when called as/from dsdot 2017-10-25 16:45:41 +02:00
Martin Kroeker b71f4fe681 Merge pull request #1334 from ashwinyes/develop_aarch64_20171024_addlocallabels
ARM64: Convert all labels to local labels
2017-10-24 19:50:03 +02:00
Ashwin Sekhar T K a0128aa489 ARM64: Convert all labels to local labels
While debugging/profiling applications using perf or other tools, the
kernels appear scattered in the profile reports. This is because the labels
within the kernels are not local and each label is shown as a separate
function.

To avoid this, all the labels within the kernels are changed to local
labels.
2017-10-24 11:40:05 +00:00
Martin Kroeker 627133f9ad Merge pull request #1333 from martin-frbg/haswell32
Fix 32bit HASWELL builds
2017-10-24 11:25:03 +02:00
Martin Kroeker 0e2cf102e1 Fix 32bit HASWELL 2017-10-24 10:07:44 +02:00
Martin Kroeker 5e3e91d0fc Split the microkernel workload into chunks of 32 floats for dsdot mode to limit loss of precision 2017-10-22 18:18:51 +02:00
Martin Kroeker 28c3fa8950 Add dsdot 2017-10-16 23:29:03 +02:00
Martin Kroeker 8ac87c1cb6 Implement DSDOT with unchanged sdot microkernels 2017-10-16 23:27:51 +02:00
Martin Kroeker b7cee00455 Merge pull request #1327 from martin-frbg/cmake-relapack
Make ReLAPACK available in cmake builds
2017-10-12 20:26:35 +02:00
Martin Kroeker 962b20a9bb Optionally add ReLAPACK to LIB_COMPONENTS 2017-10-12 17:02:01 +02:00
Martin Kroeker fbf83f4833 Add cmake build list file for ReLAPACK 2017-10-12 17:00:00 +02:00
Martin Kroeker 78cec6209c Add ReLAPACK option 2017-10-12 16:58:37 +02:00
Martin Kroeker c460027dbe Merge pull request #1325 from grisuthedragon/patch-1
Update README.md to include POWER8
2017-10-10 12:14:34 +02:00
Martin Köhler bfa9b9f6b2 Update README.md
Add POWER 8 to the list of additional architectures.
2017-10-10 10:12:04 +02:00
Martin Kroeker c7a8512d12 Cmake fixes for DYNAMIC_ARCH builds and whitespace in path names (#1323)
* prebuild.cmake: Put quotes around path names that may contain whitespace
(Copied from alexkaratakis' PR #1295)
* kernel/CMakeLists.txt: Fix common_lapack header inclusion and DYNAMIC_ARCH generation of ?neg_tcopy and ?laswp_ncopy files
* lapack/CMakeLists.txt: Use correct template for ?laswp_(plus,minus) functions
2017-10-09 23:34:18 +02:00
Martin Kroeker db72ad8f6a Merge pull request #1320 from timmoon10/develop
2D thread distribution for multi-threaded GEMMs
2017-10-08 23:31:33 +02:00
Martin Kroeker 97ecd4996a Merge pull request #1319 from martin-frbg/issue601
Fix out-of-bounds memory accesses exposed by xccblat3 testcase
2017-10-08 23:31:06 +02:00
Martin Kroeker 1eb43cccad Merge pull request #1317 from martin-frbg/power8-asm
Save and restore VSX registers
2017-10-08 23:30:46 +02:00
Martin Kroeker 9d92f526dd Comment out a code block that performs out-of-bounds memory accesses
...and does not appear to be needed even when it stays within the bounds of the array
2017-10-06 23:51:32 +02:00
Martin Kroeker 514d237257 Merge pull request #1279 from xsacha/develop
CMake improvements
2017-10-06 21:13:45 +02:00
Tim Moon 30486a356c Reduce number of data partitions in n. 2017-10-04 12:37:49 -07:00
Martin Kroeker e1b2502840 Merge pull request #1316 from timmoon10/develop
Variable thread count for multi-threaded GEMMs
2017-10-04 20:35:00 +02:00
Tim Moon 9de52b489a Cleaning up and documenting multi-threaded GEMM code. 2017-10-03 16:32:08 -07:00
Tim Moon 860dcfc703 Use 2D thread distribution for small GEMMs.
Allows maximum use of available cores if one of M and N is small and the other is large.
2017-10-03 13:43:39 -07:00
Martin Kroeker f96afd94b0 Fix out-of-bounds accesses where the data should be zero anyway 2017-10-01 01:06:39 +02:00
Martin Kroeker ebe84215e4 Merge pull request #1318 from pv/potrf-smoketest
Add trivial smoketest for xpotrf
2017-09-30 21:31:28 +02:00
Pauli Virtanen 845e6d750f Add trivial smoketest for xpotrf 2017-09-30 19:07:54 +02:00
Tim Moon a89d6711c6 Increasing flexibility of GEMM benchmark.
m, n, and k can be set to arbitrary constants. A and B matrices can be transposed independently.
2017-09-28 12:56:29 -07:00
Martin Kroeker 9c017a2218 Save and restore VSX registers 2017-09-28 12:17:09 +02:00
Tim Moon 0e6b11b708 Merge https://github.com/timmoon10/OpenBLAS into develop 2017-09-27 19:26:38 -07:00
Tim Moon 6aaa107865 Reducing threads for multi-threaded GEMMs on small matrices. 2017-09-27 19:25:33 -07:00
Martin Kroeker 00c42dc815 Merge pull request #1314 from martin-frbg/nofortran-fix-2
Rewrite NOFORTRAN conditionals
2017-09-26 10:34:18 +02:00
Martin Kroeker 79e754e548 Rewrite NOFORTRAN conditionals
... so that they do not trigger accidentally when NOFORTRAN is empty/unset
2017-09-25 23:45:14 +02:00
Martin Kroeker 2ccd7f6e0c Merge pull request #1310 from sva-img/develop
Added mips I6500 core
2017-09-22 09:34:54 +02:00
Shivraj Patil e3d844b062 Added mips I6500 core
Signed-off-by: Shivraj Patil <shivraj.patil@imgtec.com>
2017-09-22 11:57:43 +05:30
Martin Kroeker def146efed Merge pull request #1308 from sebastien-villemot/develop
Add support for TARGET=ZARCH_GENERIC and TARGET=Z13
2017-09-19 14:04:37 +02:00
Sébastien Villemot 7543e578a4 Add support for TARGET=ZARCH_GENERIC and TARGET=Z13 2017-09-19 12:16:42 +02:00
Martin Kroeker 601c71fe54 Merge pull request #1304 from martin-frbg/aix-build-fixes
(Plain make) build system fixes for AIX
2017-09-18 10:16:40 +02:00
Martin Kroeker 3810a6fd99 (Plain make) build system fixes for AIX
- retry fortran compiler test with aix-specific option if generic -m32/-m64 fails
- pass any custom ARFLAGS to lapack
- no addition of -m32/-m64 to the CFLAGS and FFLAGS on AIX
2017-09-18 01:29:21 +02:00
Martin Kroeker 742f54c235 Merge pull request #1303 from martin-frbg/imatcopy-rowscols
Fix cols/rows mixup in omatcopy 2nd step for BlasTrans cases
2017-09-14 21:46:26 +02:00
Martin Kroeker d674fbb4c7 Fix cols/rows mixup in omatcopy 2nd step for BlasTrans cases
Equivalent of #1244 (issue #899) for the non-complex cases. Fixes #1289
2017-09-14 19:59:05 +02:00
Martin Kroeker 2922c15f36 Merge pull request #1302 from martin-frbg/nofortran-fix
Remove default FEXTRALIBS in NOFORTRAN case
2017-09-14 11:54:20 +02:00
Martin Kroeker 3a245a376f Remove default FEXTRALIBS in NOFORTRAN case 2017-09-14 09:21:04 +02:00
Martin Kroeker 46c9357c72 Merge pull request #1288 from quickwritereader/develop
Optimized standard Blas Level-1,2 (excluding nrm2 functions) for z13 (double precision). Issue 884
2017-09-09 23:47:17 +02:00
Martin Kroeker 1c3e2d3dd5 Merge pull request #1293 from embray/cygwin/install
More canonical installation on Cygwin
2017-09-09 23:46:27 +02:00
Martin Kroeker f66d908282 Merge pull request #1299 from martin-frbg/race_fixes
Fix thread data races uncovered by gcc thread sanitizer
2017-09-09 23:41:53 +02:00
Martin Kroeker ba1f91f17b Convert another caller of "allocation" to LOCK_COMMAND
... as the "allocation" code jumped to now does UNLOCK_COMMAND instead of blas_unlock
2017-09-09 20:30:33 +02:00
Martin Kroeker f460776f0f Fix thread data races 2017-09-09 19:07:06 +02:00
Martin Kroeker e882f3d6f3 Fix thread data race in memory.c 2017-09-09 18:58:38 +02:00
Erik M. Bray dddedbab5d More canonical installation on Cygwin:
* The DLL is named cygopenblas.dll, not libopenblas.dll
* The import lib (still called libopenblas.dll.a) is installed
2017-09-07 14:18:56 +02:00
Abdurrauf 1cfdb2295d Optimized standard Blas Level-1,2 (excluding nrm2 functions) for z13 (double precision) 2017-09-06 16:41:08 +04:00
Martin Kroeker 00740c0e34 Merge pull request #1290 from martin-frbg/imatcopy
Use in-place transform shortcut only if matrix is square
2017-09-03 13:02:10 +02:00
Martin Kroeker 254db9bd7c Use in-place transform shortcut only if matrix is square 2017-09-03 09:52:55 +02:00
Martin Kroeker f2074f9ac1 Merge pull request #1286 from martin-frbg/baytrail
Fix coretype detection for Bay Trail Atom
2017-08-27 13:23:57 +02:00
Martin Kroeker aece65ea29 Fix coretype detection for Bay Trail Atom
My earlier PR #982 appears to have been incomplete in this regard - fixes #1285
2017-08-27 13:06:54 +02:00
Sacha ef64991506 Clean up config file writing. 2017-08-23 12:47:38 +10:00
Sacha 7a867082d8 Fix open_blas.config which was never working out-of-source. Remove need for gen_config_h.exe. If OpenMP is requested, do not silently ignore when it isn't available. 2017-08-23 11:16:24 +10:00
Sacha Refshauge a1b87eac6b Do not require Perl for MSVC if CMake >= 3.4 2017-08-23 07:19:02 +10:00
Sacha Refshauge 47ebce4d1a Clean up, fix old typos. Simplify arch usages. Move system arch check to earlier position. 2017-08-21 00:37:29 +10:00
Sacha Refshauge 69b560751c Improvements to previous commit (cross-compile).
Fix typos and bad if statements discovered in 0.2.20.
2017-08-20 22:50:31 +10:00
Sacha Refshauge 0a7a527a92 Add support for cross compiling.
Add support for not having host compiler as CMake cannot detect such a compiler.
Add support for not using getarch.
Successfully builds Android ARMV8. Any target can be added by supplying the TARGET_CORE config in prebuild.cmake.
2017-08-20 20:08:53 +10:00
Martin Kroeker 50715e8945 Merge pull request #1281 from sharkcz/armv8
fix detection of generic ARMv8 CPUs
2017-08-19 20:37:19 +02:00
Sacha Refshauge 11911fd941 Add kernel/Makefile.LA to CMake 2017-08-20 00:59:14 +10:00
Sacha Refshauge 408b4fe83f Add a CMake GCC and Clang target to Travis CI 2017-08-20 00:59:00 +10:00
Sacha Refshauge 4474465438 Remove _static usages for tests 2017-08-20 00:13:46 +10:00
Sacha Refshauge b9ec72546c Only run utest without NOFORTRAN, same as Makefile. Linux now compiles. 2017-08-20 00:13:24 +10:00
Sacha Refshauge 37858d1146 Fix threading usage in CMake: s/SMP/USE_THREAD/ 2017-08-19 15:07:42 +10:00
Dan Horák 1763e01567 fix detection of generic ARMv8 CPUs 2017-08-18 14:53:29 +02:00
Sacha Refshauge 6aac06587d Fix typos and use CMake OpenMP support. 2017-08-17 17:27:01 +10:00
7c1acc07f0 Fix bug that required fortran. Fix bug that needed CXX var. Remove redundant set vars. Fix threading detection. Do not attempt to run code if cross compiling. 2017-08-17 03:32:04 +10:00
38d273ea03 Drop some redundant vars and improve arch detection in CMake. 2017-08-17 02:04:36 +10:00
7242cdc4ec Allow CMake to determine if it is building static or shared. 2017-08-17 00:51:04 +10:00
90a4dab501 Let CMake deal with build type. 2017-08-17 00:35:54 +10:00
Martin Kroeker a8a342ccc4 Merge pull request #1277 from cconrads-scicomp/fix-installation-instructions
Make: fix installation instructions
2017-08-10 23:42:23 +02:00
Martin Kroeker 9e9a9553db Merge pull request #1276 from cconrads-scicomp/android_-lm_fix
ARM: do not add linker flag `-lm` unconditionally
2017-08-10 21:35:32 +02:00
Martin Kroeker be7c1b6324 Merge pull request #1275 from cconrads-scicomp/recognize-gfortran-on-arm
ARM: recognize gfortran pre-releases
2017-08-10 21:32:09 +02:00
Christoph Conrads be8e8228ae Make: show installation instructions after build 2017-08-10 14:22:26 -04:00
Christoph Conrads 77b197bc93 Make: fix installation instructions
The installation instructions as shown after successfully compiling
OpenBLAS are wrong because this arguments used during compilation have
to be provided to Make again.
2017-08-10 12:47:18 -04:00
Christoph Conrads 9cc6599056 ARM: do not add linker flag -lm unconditionally
On ARM the required math library depends on whether the soft floating
point ABI is used or not but this is already handled in
`Makefile.system`, lines 499-505.
2017-08-10 11:58:25 -04:00
Christoph Conrads da611dd81a ARM: recognize gfortran pre-releases
Without proper recognition of gfortran versions such as
> GNU Fortran (GCC) 4.9.x 20150123 (prerelease)
OpenBLAS assumes the presence of the G77 compiler. Consequently,
`-lgfortran` is missing from the pkg-config file.

The aforementioned compiler is the gfortran compiler in the Android repo
in a commit tagged as `ndk-r14`, cf. Paul Mustière's gfortran build
instructions for Android at https://github.com/buffer51/android-gfortran
2017-08-10 11:48:29 -04:00
Martin Kroeker e6b9b660c3 Merge pull request #1264 from isuruf/dyn
Support DYNAMIC_ARCH with CMake
2017-08-08 23:47:47 +02:00
Martin Kroeker 62635b1fb5 Merge pull request #1268 from jirutka/travis-2
Travis: Add jobs building with clang and disable job `LINUX64_MUSL USE_OPENMP=1`
2017-08-08 21:54:35 +02:00
Martin Kroeker 75ddfbd687 Change travis back to sudo true
,,,to see if this has any influence on the recent ld SIGKILLS
2017-08-08 16:39:13 +02:00
Isuru Fernando 2f12ea017b No strncasecmp with MSVC 2017-08-08 00:07:25 +05:30
Isuru Fernando d3b677fe87 Add commonobjs 2017-08-07 23:12:40 +05:30
Isuru Fernando 6f83004252 Test DYNAMIC_ARCH on appveyor 2017-08-06 19:17:31 +05:30
Isuru Fernando 505b218829 Merge remote-tracking branch 'upstream/develop' into dyn 2017-08-06 19:07:00 +05:30
Martin Kroeker 719fcc56b0 Merge pull request #1262 from martin-frbg/xmv_thread-splitting
Make sure that range limit of last thread never exceeds data size
2017-08-06 14:11:44 +02:00
Martin Kroeker 1296c89688 Merge pull request #1256 from isuruf/develop
Support compiling with clang on windows
2017-08-06 14:10:18 +02:00
Jakub Jirutka dfb91f2a31 Travis: Add jobs building with clang 2017-08-06 12:40:59 +02:00
Jakub Jirutka 2ec76011f9 Travis: Disable job "LINUX64_MUSL USE_OPENMP=1"
https://github.com/xianyi/OpenBLAS/pull/1255#issuecomment-320494610
2017-08-06 11:06:03 +02:00
Isuru Fernando 7abbe40980 Build all branches so that appveyor works in forks 2017-08-04 08:04:16 +05:30
Isuru Fernando ca32b66a1c New utest for clang 2017-08-04 07:58:04 +05:30
Isuru Fernando d9346930dd Merge remote-tracking branch 'upstream/develop' into develop 2017-08-04 07:57:55 +05:30
Martin Kroeker 0bfe5c6822 Merge pull request #1266 from ashwinyes/develop_thunderx2t99_fix_clang_compilation
THUDNERX2T99: Fix clang compilation
2017-08-03 15:33:28 +02:00
Ashwin Sekhar T K 4899d67f7d THUDNERX2T99: Fix clang compilation 2017-08-02 11:28:45 -07:00
Isuru Fernando 1d1854032b Add missing EXCAVATOR 2017-08-02 19:03:04 +05:30
Martin Kroeker ebb04e3265 Merge pull request #1259 from isuruf/cmake
CMake Improvements
2017-08-02 15:31:05 +02:00
Isuru Fernando 2c51a990ac Fix extra whitespaces. CMake parser macro fails with it
TODO: Fix the parser macro to strip trailing whitespaces
2017-08-02 18:26:57 +05:30
Isuru Fernando 7892434572 Add hemm3m and symm3m objects 2017-08-02 18:24:54 +05:30
Isuru Fernando d798487213 Fixes for dynamic_arch. almost there 2017-08-02 17:25:49 +05:30
Martin Kroeker 0ba64cee60 Update trmv_thread.c 2017-08-02 12:03:54 +02:00
Martin Kroeker 53aee36076 Merge pull request #1255 from jirutka/travis
Travis: Rewrite config, build and test also on Alpine Linux (musl libc)
2017-08-02 11:59:17 +02:00
Martin Kroeker c4e5ba1bfe Make sure that range_n of last thread never exceeds the actual data size when splitting the workload 2017-08-02 00:37:58 +02:00
Jakub Jirutka 486a485bb7 Travis: Allow job LINUX64_MUSL USE_OPENMP=1 to fail
See: https://github.com/xianyi/OpenBLAS/pull/1255#issuecomment-318692183
2017-08-01 21:58:36 +02:00
Jakub Jirutka 08c7d1ddf8 Travis: Disable some gcc warnings to avoid exceeding Travis limit
See: https://github.com/xianyi/OpenBLAS/pull/1255#issuecomment-318628666
2017-08-01 21:58:15 +02:00
Jakub Jirutka e0bd5b5c0e Travis: Build and test also on Alpine Linux (musl libc)
Alpine jobs needs sudo (for chroot), so they run on VMs infrastructure.
That's why they are much slower than other jobs.
2017-08-01 21:58:03 +02:00
Jakub Jirutka ae65f75572 Travis: Simplify configuration using Build Stages and APT addon
Using APT addon has nice side-effect - you don't need sudo anymore, so
it can run on Travis containers-based infrastructure that is much faster
than their VMs infrastructure (used when sudo is needed).

You've been still running on Ubuntu Presty builders, but new default is
Trusty. Thus I've explicitly set `dist: presty` to let it stay on
Presty, to not change build environment by this commit.
2017-08-01 21:52:03 +02:00
Martin Kroeker 1e9247c276 Merge pull request #1260 from xianyi/revert-1254-xbmv_range
Revert "Fix calculated range limit exceeding actual data size for last thread"
2017-08-01 20:07:32 +02:00
Isuru Fernando 251715d9ef configure kernel_core.h 2017-08-01 23:23:55 +05:30
Martin Kroeker a6f533b248 Revert "Fix calculated range limit exceeding actual data size for last thread" 2017-08-01 19:28:08 +02:00
Isuru Fernando 50deeb49b7 configure setparam 2017-08-01 22:32:47 +05:30
Isuru Fernando 4260215adf Support DYNAMIC_ARCH with cmake 2017-08-01 22:25:52 +05:30
Isuru Fernando 89ec2be110 Fix lapacke copying 2017-08-01 15:47:14 +05:30
Isuru Fernando 8381591f56 No need of a temp file for f77blas.h 2017-08-01 15:27:19 +05:30
Isuru Fernando d245caa49a Support out-of-source build 2017-08-01 15:16:14 +05:30
Isuru Fernando c9c13a8a0c Fix installing cblas.h and fix tabs 2017-08-01 14:58:49 +05:30
Martin Kroeker e70a6b92bf Merge pull request #1257 from martin-frbg/cgroups-prereq
Rework __GLIBC_PREREQ checks to avoid breaking non-glibc builds
2017-08-01 11:23:03 +02:00
Isuru Fernando 7a96499b29 Don't change timestamps 2017-08-01 13:43:59 +05:30
Martin Kroeker 63cfa32691 Rework __GLIBC_PREREQ checks to avoid breaking non-glibc builds 2017-07-31 21:02:43 +02:00
Martin Kroeker d537e0de8c Merge pull request #1254 from martin-frbg/xbmv_range
Fix calculated range limit exceeding actual data size for last thread
2017-07-31 17:46:40 +02:00
Isuru Fernando f00bbb9dbf Remove unnecessary line in appveyor 2017-07-30 00:00:37 +05:30
Isuru Fernando 1a02a087a1 Fix vcvarsall call in appveyor 2017-07-29 23:42:56 +05:30
Isuru Fernando 02c1f86055 Fix copying libopenblas.dll 2017-07-29 23:42:38 +05:30
Isuru Fernando 5e0f67c666 Make ARCH variable a CACHE variable 2017-07-29 23:30:15 +05:30
Isuru Fernando 7345795e64 Try adding RC to path 2017-07-29 22:16:53 +05:30
Isuru Fernando a36e976491 vsvarsall in appveyor 2017-07-29 21:58:53 +05:30
Isuru Fernando 1169f489a4 Fix CMAKE_C_COMPILER in appveyor 2017-07-29 21:54:32 +05:30
Isuru Fernando 7a4ebf825b add --yes to conda in appveyor.yml 2017-07-29 21:48:49 +05:30
Isuru Fernando ff17e3eb9f build clang-cl first 2017-07-29 21:47:15 +05:30
Isuru Fernando c56d4881f9 Fix appveyor.yml 2017-07-29 21:42:59 +05:30
Isuru Fernando b03d50b794 Test clang in appveyor.yml 2017-07-29 21:16:00 +05:30
Isuru Fernando ea1095135e Ninja complains that file openblas.def does not exist 2017-07-29 21:00:32 +05:30
Isuru Fernando 4c5df489db clang on windows needs FU='' 2017-07-29 20:59:28 +05:30
Isuru Fernando eb98fdddfc typedefs only for c 2017-07-29 20:38:16 +05:30
Isuru Fernando ca17b4b75c Fix complex support for MSVC headers 2017-07-28 11:50:29 +05:30
Isuru Fernando dc24914415 check compiler is msvc instead of msvc 2017-07-28 11:49:39 +05:30
Martin Kroeker 585c0010a5 Fix range limit exceeding actual data size in last step 2017-07-28 00:27:02 +02:00
Martin Kroeker 857f61bc5d Fix range limit exceeding data size in last step 2017-07-28 00:21:53 +02:00
Martin Kroeker 9332042d5f Fix range exceeding actual data size in quick_divide 2017-07-28 00:13:24 +02:00
Martin Kroeker ae93532fd3 Merge pull request #1249 from martin-frbg/cgroup
Honor cgroup/cpuset limits when enumerating cpus
2017-07-25 23:31:57 +02:00
Martin Kroeker c4af196a2d Honor cgroup/cpuset limits when enumerating cpus 2017-07-25 22:47:34 +02:00
Martin Kroeker 480e697681 Revert "Honor cgroup/cpuset limits when enumerating cpus" (#1246) 2017-07-24 16:17:50 +02:00
Zhang Xianyi 3c4c472584 Merge pull request #1236 from martin-frbg/l1cache
Use cpuid 4 with subleafs to query L1 cache size on Intel processors
2017-07-24 12:07:00 +08:00
Zhang Xianyi a797666fbe Bump develop version for 0.3.0. 2017-07-24 12:06:29 +08:00
Martin Kroeker 00774b1105 Add dummy implementation of cpuid_count for the CPUIDEMU case 2017-07-12 21:56:23 +02:00
Martin Kroeker 6497aae57c Use cpuid 4 with subleafs to query L1 cache size on Intel processors 2017-07-12 20:43:09 +02:00
2591 changed files with 83551 additions and 32606 deletions
+1
View File
@@ -5,6 +5,7 @@
*.def
*.o
*.out
*.tmp
lapack-3.1.1
lapack-3.1.1.tgz
lapack-3.4.1
+174 -29
View File
@@ -1,5 +1,179 @@
# XXX: Precise is already deprecated, new default is Trusty.
# https://blog.travis-ci.com/2017-07-11-trusty-as-default-linux-is-coming
dist: precise
sudo: true
language: c
jobs:
include:
- &test-ubuntu
os: linux
stage: test
compiler: gcc
addons:
apt:
packages:
- gfortran
before_script: &common-before
- COMMON_FLAGS="DYNAMIC_ARCH=1 TARGET=NEHALEM NUM_THREADS=32"
script:
- set -e
- travis_wait 45 make QUIET_MAKE=1 $COMMON_FLAGS $BTYPE
- make -C test $COMMON_FLAGS $BTYPE
- make -C ctest $COMMON_FLAGS $BTYPE
- make -C utest $COMMON_FLAGS $BTYPE
env:
- TARGET_BOX=LINUX64
- BTYPE="BINARY=64"
- <<: *test-ubuntu
env:
- TARGET_BOX=LINUX64
- BTYPE="BINARY=64 USE_OPENMP=1"
- <<: *test-ubuntu
env:
- TARGET_BOX=LINUX64
- BTYPE="BINARY=64 INTERFACE64=1"
- <<: *test-ubuntu
compiler: clang
env:
- TARGET_BOX=LINUX64
- BTYPE="BINARY=64 CC=clang"
- <<: *test-ubuntu
compiler: clang
env:
- TARGET_BOX=LINUX64
- BTYPE="BINARY=64 INTERFACE64=1 CC=clang"
- <<: *test-ubuntu
addons:
apt:
packages:
- gcc-multilib
- gfortran-multilib
env:
- TARGET_BOX=LINUX32
- BTYPE="BINARY=32"
- os: linux
stage: test
compiler: gcc
addons:
apt:
packages:
- binutils-mingw-w64-x86-64
- gcc-mingw-w64-x86-64
- gfortran-mingw-w64-x86-64
before_script: *common-before
script:
- travis_wait 45 make QUIET_MAKE=1 $COMMON_FLAGS $BTYPE
env:
- TARGET_BOX=WIN64
- BTYPE="BINARY=64 HOSTCC=gcc CC=x86_64-w64-mingw32-gcc FC=x86_64-w64-mingw32-gfortran"
# Build & test on Alpine Linux inside chroot, i.e. on system with musl libc.
# These jobs needs sudo, so Travis runs them on VM-based infrastructure
# which is slower than container-based infrastructure used for jobs
# that don't require sudo.
- &test-alpine
os: linux
stage: test
dist: trusty
sudo: true
language: minimal
before_install:
- "wget 'https://raw.githubusercontent.com/alpinelinux/alpine-chroot-install/v0.6.0/alpine-chroot-install' \
&& echo 'a827a4ba3d0817e7c88bae17fe34e50204983d1e alpine-chroot-install' | sha1sum -c || exit 1"
- alpine() { /alpine/enter-chroot -u "$USER" "$@"; }
install:
- sudo sh alpine-chroot-install -p 'build-base gfortran perl linux-headers'
before_script: *common-before
script:
- set -e
# XXX: Disable some warnings for now to avoid exceeding Travis limit for log size.
- alpine make QUIET_MAKE=1 $COMMON_FLAGS $BTYPE
CFLAGS="-Wno-misleading-indentation -Wno-sign-conversion -Wno-incompatible-pointer-types"
- alpine make -C test $COMMON_FLAGS $BTYPE
- alpine make -C ctest $COMMON_FLAGS $BTYPE
- alpine make -C utest $COMMON_FLAGS $BTYPE
env:
- TARGET_BOX=LINUX64_MUSL
- BTYPE="BINARY=64"
# XXX: This job segfaults in TESTS OF THE COMPLEX LEVEL 3 BLAS,
# but only on Travis CI, cannot reproduce it elsewhere.
#- &test-alpine-openmp
# <<: *test-alpine
# env:
# - TARGET_BOX=LINUX64_MUSL
# - BTYPE="BINARY=64 USE_OPENMP=1"
- <<: *test-alpine
env:
- TARGET_BOX=LINUX64_MUSL
- BTYPE="BINARY=64 INTERFACE64=1"
# Build with the same flags as Alpine do in OpenBLAS package.
- <<: *test-alpine
env:
- TARGET_BOX=LINUX64_MUSL
- BTYPE="BINARY=64 NO_AFFINITY=1 USE_OPENMP=0 NO_LAPACK=0 TARGET=core2"
- &test-cmake
os: linux
stage: test
compiler: clang
addons:
apt:
packages:
- gfortran
- cmake
dist: trusty
sudo: true
before_script:
- COMMON_ARGS="-DTARGET=NEHALEM -DNUM_THREADS=32"
script:
- set -e
- mkdir build
- CONFIG=Release
- cmake -Bbuild -H. $CMAKE_ARGS $COMMON_ARGS -DCMAKE_BUILD_TYPE=$CONFIG
- cmake --build build --config $CONFIG -- -j2
env:
- CMAKE=1
- <<: *test-cmake
env:
- CMAKE=1 CMAKE_ARGS="-DNOFORTRAN=1"
- <<: *test-cmake
compiler: gcc
env:
- CMAKE=1
- &test-macos
os: osx
stage: test
osx_image: xcode8
before_script:
- COMMON_FLAGS="DYNAMIC_ARCH=1 TARGET=NEHALEM NUM_THREADS=32"
- brew update
- brew install gcc # for gfortran
script:
- travis_wait 45 make QUIET_MAKE=1 $COMMON_FLAGS $BTYPE
env:
- BTYPE="BINARY=64 INTERFACE64=1"
- <<: *test-macos
env:
- BTYPE="BINARY=32"
# whitelist
branches:
only:
- master
- develop
notifications:
webhooks:
urls:
@@ -7,32 +181,3 @@ notifications:
on_success: change # options: [always|never|change] default: always
on_failure: always # options: [always|never|change] default: always
on_start: never # options: [always|never|change] default: always
compiler:
- gcc
env:
- TARGET_BOX=LINUX64 BTYPE="BINARY=64"
- TARGET_BOX=LINUX64 BTYPE="BINARY=64 USE_OPENMP=1"
- TARGET_BOX=LINUX64 BTYPE="BINARY=64 INTERFACE64=1"
- TARGET_BOX=LINUX32 BTYPE="BINARY=32"
- TARGET_BOX=WIN64 BTYPE="BINARY=64 HOSTCC=gcc CC=x86_64-w64-mingw32-gcc FC=x86_64-w64-mingw32-gfortran"
before_install:
- sudo apt-get update -qq
- sudo apt-get install -qq gfortran
- if [[ "$TARGET_BOX" == "WIN64" ]]; then sudo apt-get install -qq binutils-mingw-w64-x86-64 gcc-mingw-w64-x86-64 gfortran-mingw-w64-x86-64; fi
- if [[ "$TARGET_BOX" == "LINUX32" ]]; then sudo apt-get install -qq gcc-multilib gfortran-multilib; fi
script:
- set -e
- make QUIET_MAKE=1 DYNAMIC_ARCH=1 TARGET=NEHALEM NUM_THREADS=32 $BTYPE
- if [ "$TARGET_BOX" == "LINUX32" ] || [ "$TARGET_BOX" == "LINUX64" ]; then make -C test DYNAMIC_ARCH=1 TARGET=NEHALEM NUM_THREADS=32 $BTYPE; fi
- if [ "$TARGET_BOX" == "LINUX32" ] || [ "$TARGET_BOX" == "LINUX64" ]; then make -C ctest DYNAMIC_ARCH=1 TARGET=NEHALEM NUM_THREADS=32 $BTYPE; fi
- if [ "$TARGET_BOX" == "LINUX32" ] || [ "$TARGET_BOX" == "LINUX64" ]; then make -C utest DYNAMIC_ARCH=1 TARGET=NEHALEM NUM_THREADS=32 $BTYPE; fi
# whitelist
branches:
only:
- master
- develop
+126 -127
View File
@@ -3,52 +3,34 @@
##
cmake_minimum_required(VERSION 2.8.5)
project(OpenBLAS)
project(OpenBLAS C ASM)
set(OpenBLAS_MAJOR_VERSION 0)
set(OpenBLAS_MINOR_VERSION 2)
set(OpenBLAS_PATCH_VERSION 20)
set(OpenBLAS_MINOR_VERSION 3)
set(OpenBLAS_PATCH_VERSION 2.dev)
set(OpenBLAS_VERSION "${OpenBLAS_MAJOR_VERSION}.${OpenBLAS_MINOR_VERSION}.${OpenBLAS_PATCH_VERSION}")
enable_language(ASM)
enable_language(C)
# Adhere to GNU filesystem layout conventions
include(GNUInstallDirs)
if(MSVC)
set(OpenBLAS_LIBNAME libopenblas)
else()
include(CMakePackageConfigHelpers)
set(OpenBLAS_LIBNAME openblas)
endif()
#######
if(MSVC)
option(BUILD_WITHOUT_LAPACK "Without LAPACK and LAPACKE (Only BLAS or CBLAS)" ON)
endif()
option(BUILD_WITHOUT_CBLAS "Without CBLAS" OFF)
option(BUILD_DEBUG "Build Debug Version" OFF)
option(DYNAMIC_ARCH "Build with DYNAMIC_ARCH" OFF)
option(DYNAMIC_OLDER "Support older cpus with DYNAMIC_ARCH" OFF)
option(BUILD_RELAPACK "Build with ReLAPACK (recursive LAPACK" OFF)
#######
if(BUILD_WITHOUT_LAPACK)
set(NO_LAPACK 1)
set(NO_LAPACKE 1)
endif()
if(CMAKE_CONFIGURATION_TYPES) # multiconfig generator?
set(CMAKE_CONFIGURATION_TYPES "Debug;Release" CACHE STRING "" FORCE)
set(CMAKE_BUILD_TYPE
Debug Debug
Release Release
)
else()
if( NOT CMAKE_BUILD_TYPE )
if(BUILD_DEBUG)
set(CMAKE_BUILD_TYPE Debug)
else()
set(CMAKE_BUILD_TYPE Release)
endif()
endif()
endif()
if(BUILD_WITHOUT_CBLAS)
set(NO_CBLAS 1)
endif()
@@ -74,6 +56,9 @@ endif ()
set(SUBDIRS ${BLASDIRS})
if (NOT NO_LAPACK)
list(APPEND SUBDIRS lapack)
if(BUILD_RELAPACK)
list(APPEND SUBDIRS relapack/src)
endif()
endif ()
# set which float types we want to build for
@@ -85,6 +70,10 @@ if (NOT DEFINED BUILD_SINGLE AND NOT DEFINED BUILD_DOUBLE AND NOT DEFINED BUILD_
set(BUILD_COMPLEX16 true)
endif ()
if (NOT DEFINED BUILD_MATGEN)
set(BUILD_MATGEN true)
endif()
set(FLOAT_TYPES "")
if (BUILD_SINGLE)
message(STATUS "Building Single Precision")
@@ -106,19 +95,10 @@ if (BUILD_COMPLEX16)
list(APPEND FLOAT_TYPES "ZCOMPLEX") # defines COMPLEX and DOUBLE
endif ()
set(SUBDIRS_ALL ${SUBDIRS} test ctest utest exports benchmark ../laswp ../bench)
# all :: libs netlib tests shared
# libs :
if (NOT DEFINED CORE OR "${CORE}" STREQUAL "UNKNOWN")
message(FATAL_ERROR "Detecting CPU failed. Please set TARGET explicitly, e.g. make TARGET=your_cpu_target. Please read README for details.")
endif ()
if (${NO_STATIC} AND ${NO_SHARED})
message(FATAL_ERROR "Neither static nor shared are enabled.")
endif ()
#Set default output directory
set( CMAKE_LIBRARY_OUTPUT_DIRECTORY ${PROJECT_BINARY_DIR}/lib)
set( CMAKE_ARCHIVE_OUTPUT_DIRECTORY ${PROJECT_BINARY_DIR}/lib)
@@ -140,23 +120,51 @@ endforeach ()
# Not using add_subdirectory here because lapack-netlib already has its own CMakeLists.txt. Instead include a cmake script with the sources we want.
if (NOT NOFORTRAN AND NOT NO_LAPACK)
include("${PROJECT_SOURCE_DIR}/cmake/lapack.cmake")
if (NOT NO_LAPACKE)
include("${PROJECT_SOURCE_DIR}/cmake/lapacke.cmake")
endif ()
if (NOT NO_LAPACKE)
include("${PROJECT_SOURCE_DIR}/cmake/lapacke.cmake")
endif ()
endif ()
# Only generate .def for dll on MSVC and always produce pdb files for debug and release
if(MSVC)
set(OpenBLAS_DEF_FILE "${PROJECT_BINARY_DIR}/openblas.def")
set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /Zi")
set(CMAKE_C_FLAGS_RELEASE "${CMAKE_C_FLAGS_RELEASE} /Zi")
set(CMAKE_SHARED_LINKER_FLAGS_RELEASE "${CMAKE_SHARED_LINKER_FLAGS_RELEASE} /DEBUG /OPT:REF /OPT:ICF")
if (${CMAKE_MAJOR_VERSION}.${CMAKE_MINOR_VERSION} LESS 3.4)
set(OpenBLAS_DEF_FILE "${PROJECT_BINARY_DIR}/openblas.def")
endif()
set(CMAKE_C_FLAGS_RELEASE "${CMAKE_C_FLAGS_RELEASE} /Zi")
set(CMAKE_SHARED_LINKER_FLAGS_RELEASE "${CMAKE_SHARED_LINKER_FLAGS_RELEASE} /DEBUG /OPT:REF /OPT:ICF")
endif()
# add objects to the openblas lib
add_library(${OpenBLAS_LIBNAME} SHARED ${LA_SOURCES} ${LAPACKE_SOURCES} ${TARGET_OBJS} ${OpenBLAS_DEF_FILE})
if (${DYNAMIC_ARCH})
add_subdirectory(kernel)
foreach(TARGET_CORE ${DYNAMIC_CORE})
message("${TARGET_CORE}")
list(APPEND TARGET_OBJS "$<TARGET_OBJECTS:kernel_${TARGET_CORE}>")
endforeach()
endif ()
include("${PROJECT_SOURCE_DIR}/cmake/export.cmake")
# Only build shared libs for MSVC
if (MSVC)
set(BUILD_SHARED_LIBS ON)
endif()
# add objects to the openblas lib
add_library(${OpenBLAS_LIBNAME} ${LA_SOURCES} ${LAPACKE_SOURCES} ${RELA_SOURCES} ${TARGET_OBJS} ${OpenBLAS_DEF_FILE})
# Android needs to explicitly link against libm
if(ANDROID)
target_link_libraries(${OpenBLAS_LIBNAME} m)
endif()
# Handle MSVC exports
if(MSVC AND BUILD_SHARED_LIBS)
if (${CMAKE_MAJOR_VERSION}.${CMAKE_MINOR_VERSION} LESS 3.4)
include("${PROJECT_SOURCE_DIR}/cmake/export.cmake")
else()
# Creates verbose .def file (51KB vs 18KB)
set_target_properties(${OpenBLAS_LIBNAME} PROPERTIES WINDOWS_EXPORT_ALL_SYMBOLS true)
endif()
endif()
# Set output for libopenblas
set_target_properties( ${OpenBLAS_LIBNAME} PROPERTIES RUNTIME_OUTPUT_DIRECTORY ${PROJECT_BINARY_DIR}/lib)
@@ -164,110 +172,87 @@ set_target_properties( ${OpenBLAS_LIBNAME} PROPERTIES LIBRARY_OUTPUT_NAME_DEBUG
foreach (OUTPUTCONFIG ${CMAKE_CONFIGURATION_TYPES})
string( TOUPPER ${OUTPUTCONFIG} OUTPUTCONFIG )
set_target_properties( ${OpenBLAS_LIBNAME} PROPERTIES RUNTIME_OUTPUT_DIRECTORY_${OUTPUTCONFIG} ${PROJECT_BINARY_DIR}/lib/${OUTPUTCONFIG} )
set_target_properties( ${OpenBLAS_LIBNAME} PROPERTIES LIBRARY_OUTPUT_DIRECTORY_${OUTPUTCONFIG} ${PROJECT_BINARY_DIR}/lib/${OUTPUTCONFIG} )
set_target_properties( ${OpenBLAS_LIBNAME} PROPERTIES ARCHIVE_OUTPUT_DIRECTORY_${OUTPUTCONFIG} ${PROJECT_BINARY_DIR}/lib/${OUTPUTCONFIG} )
endforeach()
enable_testing()
add_subdirectory(utest)
if (NOT MSVC)
#only build shared library for MSVC
add_library(${OpenBLAS_LIBNAME}_static STATIC ${LA_SOURCES} ${LAPACKE_SOURCES} ${TARGET_OBJS})
set_target_properties(${OpenBLAS_LIBNAME}_static PROPERTIES OUTPUT_NAME ${OpenBLAS_LIBNAME})
set_target_properties(${OpenBLAS_LIBNAME}_static PROPERTIES CLEAN_DIRECT_OUTPUT 1)
if(SMP)
target_link_libraries(${OpenBLAS_LIBNAME} pthread)
target_link_libraries(${OpenBLAS_LIBNAME}_static pthread)
if (USE_THREAD)
# Add threading library to linker
find_package(Threads)
if (THREADS_HAVE_PTHREAD_ARG)
set_property(TARGET ${OpenBLAS_LIBNAME} PROPERTY COMPILE_OPTIONS "-pthread")
set_property(TARGET ${OpenBLAS_LIBNAME} PROPERTY INTERFACE_COMPILE_OPTIONS "-pthread")
endif()
target_link_libraries(${OpenBLAS_LIBNAME} ${CMAKE_THREAD_LIBS_INIT})
endif()
#build test and ctest
add_subdirectory(test)
if(NOT NO_CBLAS)
add_subdirectory(ctest)
endif()
if (MSVC OR NOT NOFORTRAN)
# Broken without fortran on unix
add_subdirectory(utest)
endif()
set_target_properties(${OpenBLAS_LIBNAME} PROPERTIES
if (NOT MSVC AND NOT NOFORTRAN)
# Build test and ctest
add_subdirectory(test)
if(NOT NO_CBLAS)
add_subdirectory(ctest)
endif()
endif()
set_target_properties(${OpenBLAS_LIBNAME} PROPERTIES
VERSION ${OpenBLAS_MAJOR_VERSION}.${OpenBLAS_MINOR_VERSION}
SOVERSION ${OpenBLAS_MAJOR_VERSION}
)
# TODO: Why is the config saved here? Is this necessary with CMake?
#Save the config files for installation
# @cp Makefile.conf Makefile.conf_last
# @cp config.h config_last.h
#ifdef QUAD_PRECISION
# @echo "#define QUAD_PRECISION">> config_last.h
#endif
#ifeq ($(EXPRECISION), 1)
# @echo "#define EXPRECISION">> config_last.h
#endif
###
#ifeq ($(DYNAMIC_ARCH), 1)
# @$(MAKE) -C kernel commonlibs || exit 1
# @for d in $(DYNAMIC_CORE) ; \
# do $(MAKE) GOTOBLAS_MAKEFILE= -C kernel TARGET_CORE=$$d kernel || exit 1 ;\
# done
# @echo DYNAMIC_ARCH=1 >> Makefile.conf_last
#endif
#ifdef USE_THREAD
# @echo USE_THREAD=$(USE_THREAD) >> Makefile.conf_last
#endif
# @touch lib.grd
# Install project
# Install libraries
install(TARGETS ${OpenBLAS_LIBNAME}
EXPORT "OpenBLASTargets"
RUNTIME DESTINATION ${CMAKE_INSTALL_BINDIR}
ARCHIVE DESTINATION ${CMAKE_INSTALL_LIBDIR}
LIBRARY DESTINATION ${CMAKE_INSTALL_LIBDIR} )
LIBRARY DESTINATION ${CMAKE_INSTALL_LIBDIR} )
# Install include files
set (GENCONFIG_BIN ${CMAKE_BINARY_DIR}/gen_config_h${CMAKE_EXECUTABLE_SUFFIX})
ADD_CUSTOM_COMMAND(
OUTPUT ${CMAKE_BINARY_DIR}/openblas_config.h
DEPENDS ${CMAKE_CURRENT_SOURCE_DIR}/config.h
COMMAND ${GENCONFIG_BIN} ${CMAKE_CURRENT_SOURCE_DIR}/config.h ${CMAKE_CURRENT_SOURCE_DIR}/openblas_config_template.h > ${CMAKE_BINARY_DIR}/openblas_config.h
)
message(STATUS "Generating openblas_config.h in ${CMAKE_INSTALL_INCLUDEDIR}")
ADD_CUSTOM_TARGET(genconfig
ALL
DEPENDS openblas_config.h
)
add_dependencies(genconfig ${OpenBLAS_LIBNAME})
set(OPENBLAS_CONFIG_H ${CMAKE_BINARY_DIR}/openblas_config.h)
file(WRITE ${OPENBLAS_CONFIG_H} "#ifndef OPENBLAS_CONFIG_H\n")
file(APPEND ${OPENBLAS_CONFIG_H} "#define OPENBLAS_CONFIG_H\n")
file(STRINGS ${PROJECT_BINARY_DIR}/config.h __lines)
foreach(line ${__lines})
string(REPLACE "#define " "" line ${line})
file(APPEND ${OPENBLAS_CONFIG_H} "#define OPENBLAS_${line}\n")
endforeach()
file(APPEND ${OPENBLAS_CONFIG_H} "#define OPENBLAS_VERSION \"OpenBLAS ${OpenBLAS_VERSION}\"\n")
file(READ ${CMAKE_CURRENT_SOURCE_DIR}/openblas_config_template.h OPENBLAS_CONFIG_TEMPLATE_H_CONTENTS)
file(APPEND ${OPENBLAS_CONFIG_H} "${OPENBLAS_CONFIG_TEMPLATE_H_CONTENTS}\n")
file(APPEND ${OPENBLAS_CONFIG_H} "#endif /* OPENBLAS_CONFIG_H */\n")
install (FILES ${OPENBLAS_CONFIG_H} DESTINATION ${CMAKE_INSTALL_INCLUDEDIR})
install (FILES ${CMAKE_BINARY_DIR}/openblas_config.h DESTINATION ${CMAKE_INSTALL_INCLUDEDIR})
if(NOT NOFORTRAN)
message(STATUS "Generating f77blas.h in ${CMAKE_INSTALL_INCLUDEDIR}")
message(STATUS "Generating f77blas.h in ${CMAKE_INSTALL_INCLUDEDIR}")
ADD_CUSTOM_TARGET(genf77blas
ALL
COMMAND ${AWK} 'BEGIN{print \"\#ifndef OPENBLAS_F77BLAS_H\" \; print \"\#define OPENBLAS_F77BLAS_H\" \; print \"\#include \\"openblas_config.h\\" \"}; NF {print}; END{print \"\#endif\"}' ${CMAKE_CURRENT_SOURCE_DIR}/common_interface.h > ${CMAKE_BINARY_DIR}/f77blas.h
DEPENDS ${CMAKE_CURRENT_SOURCE_DIR}/config.h
)
add_dependencies(genf77blas ${OpenBLAS_LIBNAME})
install (FILES ${CMAKE_BINARY_DIR}/f77blas.h DESTINATION ${CMAKE_INSTALL_INCLUDEDIR})
set(F77BLAS_H ${CMAKE_BINARY_DIR}/f77blas.h)
file(WRITE ${F77BLAS_H} "#ifndef OPENBLAS_F77BLAS_H\n")
file(APPEND ${F77BLAS_H} "#define OPENBLAS_F77BLAS_H\n")
file(APPEND ${F77BLAS_H} "#include \"openblas_config.h\"\n")
file(READ ${CMAKE_CURRENT_SOURCE_DIR}/common_interface.h COMMON_INTERFACE_H_CONTENTS)
file(APPEND ${F77BLAS_H} "${COMMON_INTERFACE_H_CONTENTS}\n")
file(APPEND ${F77BLAS_H} "#endif")
install (FILES ${F77BLAS_H} DESTINATION ${CMAKE_INSTALL_INCLUDEDIR})
endif()
if(NOT NO_CBLAS)
message (STATUS "Generating cblas.h in ${CMAKE_INSTALL_INCLUDEDIR}")
ADD_CUSTOM_TARGET(gencblas
ALL
COMMAND ${SED} 's/common/openblas_config/g' ${CMAKE_CURRENT_SOURCE_DIR}/cblas.h > "${CMAKE_BINARY_DIR}/cblas.tmp"
COMMAND cp "${CMAKE_BINARY_DIR}/cblas.tmp" "${CMAKE_BINARY_DIR}/cblas.h"
DEPENDS ${CMAKE_CURRENT_SOURCE_DIR}/cblas.h
)
add_dependencies(gencblas ${OpenBLAS_LIBNAME})
install (FILES ${CMAKE_BINARY_DIR}/cblas.h DESTINATION ${CMAKE_INSTALL_INCLUDEDIR})
file(READ ${CMAKE_CURRENT_SOURCE_DIR}/cblas.h CBLAS_H_CONTENTS)
string(REPLACE "common" "openblas_config" CBLAS_H_CONTENTS_NEW "${CBLAS_H_CONTENTS}")
file(WRITE ${CMAKE_BINARY_DIR}/cblas.tmp "${CBLAS_H_CONTENTS_NEW}")
install (FILES ${CMAKE_BINARY_DIR}/cblas.tmp DESTINATION ${CMAKE_INSTALL_INCLUDEDIR} RENAME cblas.h)
endif()
if(NOT NO_LAPACKE)
@@ -275,19 +260,33 @@ if(NOT NO_LAPACKE)
add_dependencies( ${OpenBLAS_LIBNAME} genlapacke)
FILE(GLOB_RECURSE INCLUDE_FILES "${CMAKE_CURRENT_SOURCE_DIR}/lapack-netlib/LAPACKE/*.h")
install (FILES ${INCLUDE_FILES} DESTINATION ${CMAKE_INSTALL_INCLUDEDIR})
ADD_CUSTOM_TARGET(genlapacke
COMMAND cp ${CMAKE_CURRENT_SOURCE_DIR}/lapack-netlib/LAPACKE/include/lapacke_mangling_with_flags.h.in "${CMAKE_BINARY_DIR}/lapacke_mangling.h"
COMMAND ${CMAKE_COMMAND} -E copy ${CMAKE_CURRENT_SOURCE_DIR}/lapack-netlib/LAPACKE/include/lapacke_mangling_with_flags.h.in "${CMAKE_BINARY_DIR}/lapacke_mangling.h"
)
install (FILES ${CMAKE_BINARY_DIR}/lapacke_mangling.h DESTINATION ${CMAKE_INSTALL_INCLUDEDIR})
endif()
if(NOT MSVC)
install (TARGETS ${OpenBLAS_LIBNAME}_static DESTINATION ${CMAKE_INSTALL_LIBDIR})
endif()
include(FindPkgConfig QUIET)
if(PKG_CONFIG_FOUND)
configure_file(${PROJECT_SOURCE_DIR}/cmake/openblas.pc.in ${PROJECT_BINARY_DIR}/openblas.pc @ONLY)
install (FILES ${PROJECT_BINARY_DIR}/openblas.pc DESTINATION ${CMAKE_INSTALL_LIBDIR}/pkgconfig/)
endif()
# GNUInstallDirs "DATADIR" wrong here; CMake search path wants "share".
set(PN OpenBLAS)
set(CMAKECONFIG_INSTALL_DIR "share/cmake/${PN}")
configure_package_config_file(cmake/${PN}Config.cmake.in
"${CMAKE_CURRENT_BINARY_DIR}/${PN}Config.cmake"
INSTALL_DESTINATION ${CMAKECONFIG_INSTALL_DIR})
write_basic_package_version_file(${CMAKE_CURRENT_BINARY_DIR}/${PN}ConfigVersion.cmake
VERSION ${${PN}_VERSION}
COMPATIBILITY AnyNewerVersion)
install(FILES ${CMAKE_CURRENT_BINARY_DIR}/${PN}Config.cmake
${CMAKE_CURRENT_BINARY_DIR}/${PN}ConfigVersion.cmake
DESTINATION ${CMAKECONFIG_INSTALL_DIR})
install(EXPORT "${PN}Targets"
NAMESPACE "${PN}::"
DESTINATION ${CMAKECONFIG_INSTALL_DIR})
+1 -1
View File
@@ -166,5 +166,5 @@ In chronological order:
* [2017-01-01] dgemm and dtrmm kernels for IBM z13
* [2017-02-26] ztrmm kernel for IBM z13
* [2017-03-13] strmm and ctrmm kernel for IBM z13
* [2017-09-01] initial Blas Level-1,2 (double precision) for IBM z13
+29 -17
View File
@@ -21,6 +21,17 @@ ifeq ($(BUILD_RELAPACK), 1)
RELA = re_lapack
endif
ifeq ($(NO_FORTRAN), 1)
define NOFORTRAN
1
endef
define NO_LAPACK
1
endef
export NOFORTRAN
export NO_LAPACK
endif
LAPACK_NOOPT := $(filter-out -O0 -O1 -O2 -O3 -Ofast,$(LAPACK_FFLAGS))
SUBDIRS_ALL = $(SUBDIRS) test ctest utest exports benchmark ../laswp ../bench
@@ -47,7 +58,7 @@ endif
endif
@echo " C compiler ... $(C_COMPILER) (command line : $(CC))"
ifndef NOFORTRAN
ifeq ($(NOFORTRAN), $(filter 0,$(NOFORTRAN)))
@echo " Fortran compiler ... $(F_COMPILER) (command line : $(FC))"
endif
ifneq ($(OSNAME), AIX)
@@ -91,11 +102,7 @@ ifeq ($(OSNAME), $(filter $(OSNAME),Linux SunOS Android))
@ln -fs $(LIBSONAME) $(LIBPREFIX).so
@ln -fs $(LIBSONAME) $(LIBPREFIX).so.$(MAJOR_VERSION)
endif
ifeq ($(OSNAME), FreeBSD)
@$(MAKE) -C exports so
@ln -fs $(LIBSONAME) $(LIBPREFIX).so
endif
ifeq ($(OSNAME), NetBSD)
ifeq ($(OSNAME), $(filter $(OSNAME),FreeBSD OpenBSD NetBSD DragonFly))
@$(MAKE) -C exports so
@ln -fs $(LIBSONAME) $(LIBPREFIX).so
endif
@@ -112,7 +119,7 @@ endif
endif
tests :
ifndef NOFORTRAN
ifeq ($(NOFORTRAN), $(filter 0,$(NOFORTRAN)))
touch $(LIBNAME)
ifndef NO_FBLAS
$(MAKE) -C test all
@@ -157,6 +164,9 @@ ifeq ($(DYNAMIC_ARCH), 1)
do $(MAKE) GOTOBLAS_MAKEFILE= -C kernel TARGET_CORE=$$d kernel || exit 1 ;\
done
@echo DYNAMIC_ARCH=1 >> Makefile.conf_last
ifeq ($(DYNAMIC_OLDER), 1)
@echo DYNAMIC_OLDER=1 >> Makefile.conf_last
endif
endif
ifdef USE_THREAD
@echo USE_THREAD=$(USE_THREAD) >> Makefile.conf_last
@@ -211,7 +221,7 @@ netlib :
else
netlib : lapack_prebuild
ifndef NOFORTRAN
ifeq ($(NOFORTRAN), $(filter 0,$(NOFORTRAN)))
@$(MAKE) -C $(NETLIB_LAPACK_DIR) lapacklib
@$(MAKE) -C $(NETLIB_LAPACK_DIR) tmglib
endif
@@ -232,7 +242,7 @@ prof_lapack : lapack_prebuild
@$(MAKE) -C $(NETLIB_LAPACK_DIR) lapack_prof
lapack_prebuild :
ifndef NOFORTRAN
ifeq ($(NOFORTRAN), $(filter 0,$(NOFORTRAN)))
-@echo "FORTRAN = $(FC)" > $(NETLIB_LAPACK_DIR)/make.inc
-@echo "OPTS = $(LAPACK_FFLAGS)" >> $(NETLIB_LAPACK_DIR)/make.inc
-@echo "POPTS = $(LAPACK_FPFLAGS)" >> $(NETLIB_LAPACK_DIR)/make.inc
@@ -242,7 +252,7 @@ ifndef NOFORTRAN
-@echo "CC = $(CC)" >> $(NETLIB_LAPACK_DIR)/make.inc
-@echo "override CFLAGS = $(LAPACK_CFLAGS)" >> $(NETLIB_LAPACK_DIR)/make.inc
-@echo "ARCH = $(AR)" >> $(NETLIB_LAPACK_DIR)/make.inc
-@echo "ARCHFLAGS = -ru" >> $(NETLIB_LAPACK_DIR)/make.inc
-@echo "ARCHFLAGS = $(ARFLAGS) -ru" >> $(NETLIB_LAPACK_DIR)/make.inc
-@echo "RANLIB = $(RANLIB)" >> $(NETLIB_LAPACK_DIR)/make.inc
-@echo "LAPACKLIB = ../$(LIBNAME)" >> $(NETLIB_LAPACK_DIR)/make.inc
-@echo "TMGLIB = ../$(LIBNAME)" >> $(NETLIB_LAPACK_DIR)/make.inc
@@ -270,25 +280,26 @@ endif
ifeq ($(BUILD_LAPACK_DEPRECATED), 1)
-@echo "BUILD_DEPRECATED = 1" >> $(NETLIB_LAPACK_DIR)/make.inc
endif
-@echo "LAPACKE_WITH_TMG = 1" >> $(NETLIB_LAPACK_DIR)/make.inc
-@cat make.inc >> $(NETLIB_LAPACK_DIR)/make.inc
endif
large.tgz :
ifndef NOFORTRAN
ifeq ($(NOFORTRAN), $(filter 0,$(NOFORTRAN)))
if [ ! -a $< ]; then
-wget http://www.netlib.org/lapack/timing/large.tgz;
fi
endif
timing.tgz :
ifndef NOFORTRAN
ifeq ($(NOFORTRAN), $(filter 0,$(NOFORTRAN)))
if [ ! -a $< ]; then
-wget http://www.netlib.org/lapack/timing/timing.tgz;
fi
endif
lapack-timing : large.tgz timing.tgz
ifndef NOFORTRAN
ifeq ($(NOFORTRAN), $(filter 0,$(NOFORTRAN)))
(cd $(NETLIB_LAPACK_DIR); $(TAR) zxf ../timing.tgz TIMING)
(cd $(NETLIB_LAPACK_DIR)/TIMING; $(TAR) zxf ../../large.tgz )
$(MAKE) -C $(NETLIB_LAPACK_DIR)/TIMING
@@ -297,9 +308,10 @@ endif
lapack-test :
(cd $(NETLIB_LAPACK_DIR)/TESTING && rm -f x* *.out)
$(MAKE) -j 1 -C $(NETLIB_LAPACK_DIR)/TESTING xeigtstc xeigtstd xeigtsts xeigtstz xlintstc xlintstd xlintstds xlintstrfd xlintstrfz xlintsts xlintstz xlintstzc xlintstrfs xlintstrfc
$(MAKE) -j 1 -C $(NETLIB_LAPACK_DIR)/TESTING/EIG xeigtstc xeigtstd xeigtsts xeigtstz
$(MAKE) -j 1 -C $(NETLIB_LAPACK_DIR)/TESTING/LIN xlintstc xlintstd xlintstds xlintstrfd xlintstrfz xlintsts xlintstz xlintstzc xlintstrfs xlintstrfc
ifneq ($(CROSS), 1)
( cd $(NETLIB_LAPACK_DIR)/INSTALL; ./testlsame; ./testslamch; ./testdlamch; \
( cd $(NETLIB_LAPACK_DIR)/INSTALL; make all; ./testlsame; ./testslamch; ./testdlamch; \
./testsecond; ./testdsecnd; ./testieee; ./testversion )
(cd $(NETLIB_LAPACK_DIR); ./lapack_testing.py -r )
endif
@@ -311,9 +323,9 @@ lapack-runtest:
blas-test:
(cd $(NETLIB_LAPACK_DIR)/BLAS && rm -f x* *.out)
(cd $(NETLIB_LAPACK_DIR)/BLAS/TESTING && rm -f x* *.out)
$(MAKE) -j 1 -C $(NETLIB_LAPACK_DIR) blas_testing
(cd $(NETLIB_LAPACK_DIR)/BLAS && cat *.out)
(cd $(NETLIB_LAPACK_DIR)/BLAS/TESTING && cat *.out)
dummy :
+11 -14
View File
@@ -72,12 +72,7 @@ ifeq ($(OSNAME), $(filter $(OSNAME),Linux SunOS Android))
ln -fs $(LIBSONAME) $(LIBPREFIX).so ; \
ln -fs $(LIBSONAME) $(LIBPREFIX).so.$(MAJOR_VERSION)
endif
ifeq ($(OSNAME), FreeBSD)
@cp $(LIBSONAME) "$(DESTDIR)$(OPENBLAS_LIBRARY_DIR)"
@cd "$(DESTDIR)$(OPENBLAS_LIBRARY_DIR)" ; \
ln -fs $(LIBSONAME) $(LIBPREFIX).so
endif
ifeq ($(OSNAME), NetBSD)
ifeq ($(OSNAME), $(filter $(OSNAME),FreeBSD OpenBSD NetBSD DragonFly))
@cp $(LIBSONAME) "$(DESTDIR)$(OPENBLAS_LIBRARY_DIR)"
@cd "$(DESTDIR)$(OPENBLAS_LIBRARY_DIR)" ; \
ln -fs $(LIBSONAME) $(LIBPREFIX).so
@@ -90,21 +85,23 @@ ifeq ($(OSNAME), Darwin)
endif
ifeq ($(OSNAME), WINNT)
@-cp $(LIBDLLNAME) "$(DESTDIR)$(OPENBLAS_BINARY_DIR)"
@-cp $(LIBDLLNAME).a "$(DESTDIR)$(OPENBLAS_LIBRARY_DIR)"
@-cp $(IMPLIBNAME) "$(DESTDIR)$(OPENBLAS_LIBRARY_DIR)"
endif
ifeq ($(OSNAME), CYGWIN_NT)
@-cp $(LIBDLLNAME) "$(DESTDIR)$(OPENBLAS_BINARY_DIR)"
@-cp $(IMPLIBNAME) "$(DESTDIR)$(OPENBLAS_LIBRARY_DIR)"
endif
endif
#Generating openblas.pc
@echo Generating openblas.pc in $(DESTDIR)$(OPENBLAS_PKGCONFIG_DIR)
@echo 'libdir='$(OPENBLAS_LIBRARY_DIR) >> $(DESTDIR)$(OPENBLAS_PKGCONFIG_DIR)/openblas.pc
@echo 'includedir='$(OPENBLAS_INCLUDE_DIR) >> $(DESTDIR)$(OPENBLAS_PKGCONFIG_DIR)/openblas.pc
@echo 'version='$(VERSION) >> $(DESTDIR)$(OPENBLAS_PKGCONFIG_DIR)/openblas.pc
@echo 'extralib='$(EXTRALIB) >> $(DESTDIR)$(OPENBLAS_PKGCONFIG_DIR)/openblas.pc
@cat openblas.pc.in >> $(DESTDIR)$(OPENBLAS_PKGCONFIG_DIR)/openblas.pc
@echo Generating openblas.pc in "$(DESTDIR)$(OPENBLAS_PKGCONFIG_DIR)"
@echo 'libdir='$(OPENBLAS_LIBRARY_DIR) > "$(DESTDIR)$(OPENBLAS_PKGCONFIG_DIR)/openblas.pc"
@echo 'includedir='$(OPENBLAS_INCLUDE_DIR) >> "$(DESTDIR)$(OPENBLAS_PKGCONFIG_DIR)/openblas.pc"
@echo 'openblas_config= USE_64BITINT='$(USE_64BITINT) 'DYNAMIC_ARCH='$(DYNAMIC_ARCH) 'DYNAMIC_OLDER='$(DYNAMIC_OLDER) 'NO_CBLAS='$(NO_CBLAS) 'NO_LAPACK='$(NO_LAPACK) 'NO_LAPACKE='$(NO_LAPACKE) 'NO_AFFINITY='$(NO_AFFINITY) 'USE_OPENMP='$(USE_OPENMP) $(CORE) 'MAX_THREADS='$(NUM_THREADS)>> "$(DESTDIR)$(OPENBLAS_PKGCONFIG_DIR)/openblas.pc"
@echo 'version='$(VERSION) >> "$(DESTDIR)$(OPENBLAS_PKGCONFIG_DIR)/openblas.pc"
@echo 'extralib='$(EXTRALIB) >> "$(DESTDIR)$(OPENBLAS_PKGCONFIG_DIR)/openblas.pc"
@cat openblas.pc.in >> "$(DESTDIR)$(OPENBLAS_PKGCONFIG_DIR)/openblas.pc"
#Generating OpenBLASConfig.cmake
@@ -114,7 +111,7 @@ endif
ifndef NO_SHARED
#ifeq logical or
ifeq ($(OSNAME), $(filter $(OSNAME),Linux FreeBSD NetBSD))
ifeq ($(OSNAME), $(filter $(OSNAME),Linux FreeBSD NetBSD OpenBSD DragonFly))
@echo "SET(OpenBLAS_LIBRARIES ${OPENBLAS_LIBRARY_DIR}/$(LIBPREFIX).so)" >> "$(DESTDIR)$(OPENBLAS_CMAKE_DIR)/$(OPENBLAS_CMAKE_CONFIG)"
endif
ifeq ($(OSNAME), $(filter $(OSNAME),WINNT CYGWIN_NT))
+8
View File
@@ -17,6 +17,10 @@ ifdef CPUIDEMU
EXFLAGS = -DCPUIDEMU -DVENDOR=99
endif
ifeq ($(TARGET), 1004K)
TARGET_FLAGS = -mips32r2
endif
ifeq ($(TARGET), P5600)
TARGET_FLAGS = -mips32r5
endif
@@ -29,6 +33,10 @@ ifeq ($(TARGET), P6600)
TARGET_FLAGS = -mips64r6
endif
ifeq ($(TARGET), I6500)
TARGET_FLAGS = -mips64r6
endif
all: getarch_2nd
./getarch_2nd 0 >> $(TARGET_MAKE)
./getarch_2nd 1 >> $(TARGET_CONF)
+27 -4
View File
@@ -3,7 +3,7 @@
#
# This library's version
VERSION = 0.2.20
VERSION = 0.3.2.dev
# If you set the suffix, the library name will be libopenblas_$(LIBNAMESUFFIX).a
# and libopenblas_$(LIBNAMESUFFIX).so. Meanwhile, the soname in shared library
@@ -17,6 +17,11 @@ VERSION = 0.2.20
# If you want to support multiple architecture in one binary
# DYNAMIC_ARCH = 1
# If you want the full list of x86_64 architectures supported in DYNAMIC_ARCH
# mode (including individual optimizied codes for PENRYN, DUNNINGTON, OPTERON,
# OPTERON_SSE3, ATOM and NANO rather than fallbacks to older architectures)
# DYNAMIC_OLDER = 1
# C compiler including binary type(32bit / 64bit). Default is gcc.
# Don't use Intel Compiler or PGI, it won't generate right codes as I expect.
# CC = gcc
@@ -55,11 +60,26 @@ VERSION = 0.2.20
# This flag is always set for POWER8. Don't modify the flag
# USE_OPENMP = 1
# The OpenMP scheduler to use - by default this is "static" and you
# will normally not want to change this unless you know that your main
# workload will involve tasks that have highly unbalanced running times
# for individual threads. Changing away from "static" may also adversely
# affect memory access locality in NUMA systems. Setting to "runtime" will
# allow you to select the scheduler from the environment variable OMP_SCHEDULE
# CCOMMON_OPT += -DOMP_SCHED=dynamic
# You can define maximum number of threads. Basically it should be
# less than actual number of cores. If you don't specify one, it's
# automatically detected by the the script.
# NUM_THREADS = 24
# If you have enabled USE_OPENMP and your application would call
# OpenBLAS's calculation API from multi threads, please comment it in.
# This flag defines how many instances of OpenBLAS's calculation API can
# actually run in parallel. If more threads call OpenBLAS's calculation API,
# they need to wait for the preceding API calls to finish or risk data corruption.
# NUM_PARALLEL = 2
# if you don't need to install the static library, please comment it in.
# NO_STATIC = 1
@@ -100,7 +120,7 @@ BUILD_LAPACK_DEPRECATED = 1
NO_WARMUP = 1
# If you want to disable CPU/Memory affinity on Linux.
#NO_AFFINITY = 1
NO_AFFINITY = 1
# if you are compiling for Linux and you have more than 16 numa nodes or more than 256 cpus
# BIGNUMA = 1
@@ -144,8 +164,11 @@ NO_WARMUP = 1
# CONSISTENT_FPCSR = 1
# If any gemm arguement m, n or k is less or equal this threshold, gemm will be execute
# with single thread. You can use this flag to avoid the overhead of multi-threading
# in small matrix sizes. The default value is 4.
# with single thread. (Actually in recent versions this is a factor proportional to the
# number of floating point operations necessary for the given problem size, no longer
# an individual dimension). You can use this setting to avoid the overhead of multi-
# threading in small matrix sizes. The default value is 4, but values as high as 50 have
# been reported to be optimal for certain workloads (50 is the recommended value for Julia).
# GEMM_MULTITHREAD_THRESHOLD = 4
# If you need santy check by comparing reference BLAS. It'll be very
+103 -16
View File
@@ -17,15 +17,24 @@ NETLIB_LAPACK_DIR = $(TOPDIR)/lapack-netlib
# http://stackoverflow.com/questions/4029274/mingw-and-make-variables
# - Default value is 'cc' which is not always a valid command (e.g. MinGW).
ifeq ($(origin CC),default)
# Check if $(CC) refers to a valid command and set the value to gcc if not
ifneq ($(findstring cmd.exe,$(SHELL)),)
ifeq ($(shell where $(CC) 2>NUL),)
CC = gcc
# Change the default compile to clang on Mac OSX.
# http://stackoverflow.com/questions/714100/os-detecting-makefile
UNAME_S := $(shell uname -s)
ifeq ($(UNAME_S),Darwin)
CC = clang
# EXTRALIB += -Wl,-no_compact_unwind
endif
endif
else # POSIX-ish
ifeq ($(shell command -v $(CC) 2>/dev/null),)
ifeq ($(shell uname -s),Darwin)
CC = clang
# EXTRALIB += -Wl,-no_compact_unwind
else
CC = gcc
endif # Darwin
endif # CC exists
endif # Shell is sane
endif # CC is set to default
# Default Fortran compiler (FC) is selected by f_check.
@@ -53,6 +62,9 @@ ifeq ($(BINARY), 32)
ifeq ($(TARGET), HASWELL)
GETARCH_FLAGS := -DFORCE_NEHALEM
endif
ifeq ($(TARGET), SKYLAKEX)
GETARCH_FLAGS := -DFORCE_NEHALEM
endif
ifeq ($(TARGET), SANDYBRIDGE)
GETARCH_FLAGS := -DFORCE_NEHALEM
endif
@@ -86,6 +98,9 @@ ifeq ($(BINARY), 32)
ifeq ($(TARGET_CORE), HASWELL)
GETARCH_FLAGS := -DFORCE_NEHALEM
endif
ifeq ($(TARGET_CORE), SKYLAKEX)
GETARCH_FLAGS := -DFORCE_NEHALEM
endif
ifeq ($(TARGET_CORE), SANDYBRIDGE)
GETARCH_FLAGS := -DFORCE_NEHALEM
endif
@@ -132,6 +147,10 @@ ifeq ($(NO_AVX2), 1)
GETARCH_FLAGS += -DNO_AVX2
endif
ifeq ($(NO_AVX512), 1)
GETARCH_FLAGS += -DNO_AVX512
endif
ifeq ($(DEBUG), 1)
GETARCH_FLAGS += -g
endif
@@ -175,12 +194,17 @@ endif
endif
ifndef NUM_PARALLEL
NUM_PARALLEL = 1
endif
ifndef NUM_THREADS
NUM_THREADS = $(NUM_CORES)
endif
ifeq ($(NUM_THREADS), 1)
override USE_THREAD = 0
override USE_OPENMP = 0
endif
ifdef USE_THREAD
@@ -224,12 +248,12 @@ endif
ifeq ($(OSNAME), Darwin)
ifndef MACOSX_DEPLOYMENT_TARGET
export MACOSX_DEPLOYMENT_TARGET=10.6
export MACOSX_DEPLOYMENT_TARGET=10.8
endif
MD5SUM = md5 -r
endif
ifeq ($(OSNAME), FreeBSD)
ifneq (,$(findstring $(OSNAME), FreeBSD OpenBSD DragonFly))
MD5SUM = md5 -r
endif
@@ -303,6 +327,7 @@ endif
ifeq ($(OSNAME), CYGWIN_NT)
NEED_PIC = 0
NO_EXPRECISION = 1
OS_CYGWIN_NT = 1
endif
ifneq ($(OSNAME), WINNT)
@@ -422,7 +447,7 @@ CCOMMON_OPT += -fopenmp
endif
ifeq ($(C_COMPILER), INTEL)
CCOMMON_OPT += -openmp
CCOMMON_OPT += -fopenmp
endif
ifeq ($(C_COMPILER), PGI)
@@ -447,13 +472,37 @@ DYNAMIC_CORE = KATMAI COPPERMINE NORTHWOOD PRESCOTT BANIAS \
endif
ifeq ($(ARCH), x86_64)
DYNAMIC_CORE = PRESCOTT CORE2 PENRYN DUNNINGTON NEHALEM OPTERON OPTERON_SSE3 BARCELONA BOBCAT ATOM NANO
DYNAMIC_CORE = PRESCOTT CORE2
ifeq ($(DYNAMIC_OLDER), 1)
DYNAMIC_CORE += PENRYN DUNNINGTON
endif
DYNAMIC_CORE += NEHALEM
ifeq ($(DYNAMIC_OLDER), 1)
DYNAMIC_CORE += OPTERON OPTERON_SSE3
endif
DYNAMIC_CORE += BARCELONA
ifeq ($(DYNAMIC_OLDER), 1)
DYNAMIC_CORE += BOBCAT ATOM NANO
endif
ifneq ($(NO_AVX), 1)
DYNAMIC_CORE += SANDYBRIDGE BULLDOZER PILEDRIVER STEAMROLLER EXCAVATOR
endif
ifneq ($(NO_AVX2), 1)
DYNAMIC_CORE += HASWELL ZEN
endif
ifneq ($(NO_AVX512), 1)
ifneq ($(NO_AVX2), 1)
DYNAMIC_CORE += SKYLAKEX
endif
endif
endif
ifdef DYNAMIC_LIST
override DYNAMIC_CORE = PRESCOTT $(DYNAMIC_LIST)
XCCOMMON_OPT = -DDYNAMIC_LIST -DDYN_PRESCOTT
XCCOMMON_OPT += $(foreach dcore,$(DYNAMIC_LIST),-DDYN_$(dcore))
CCOMMON_OPT += $(XCCOMMON_OPT)
#CCOMMON_OPT += -DDYNAMIC_LIST='$(DYNAMIC_LIST)'
endif
# If DYNAMIC_CORE is not set, DYNAMIC_ARCH cannot do anything, so force it to empty
@@ -553,9 +602,14 @@ CCOMMON_OPT += -march=mips64
FCOMMON_OPT += -march=mips64
endif
ifeq ($(CORE), 1004K)
CCOMMON_OPT += -mips32r2 $(MSA_FLAGS)
FCOMMON_OPT += -mips32r2 $(MSA_FLAGS)
endif
ifeq ($(CORE), P5600)
CCOMMON_OPT += -mips32r5 -mnan=2008 -mtune=p5600 $(MSA_FLAGS)
FCOMMON_OPT += -mips32r5 -mnan=2008 -mtune=p5600 $(MSA_FLAGS)
CCOMMON_OPT += -mips32r5 -mnan=2008 -mtune=p5600 $(MSA_FLAGS)
FCOMMON_OPT += -mips32r5 -mnan=2008 -mtune=p5600 $(MSA_FLAGS)
endif
ifeq ($(CORE), I6400)
@@ -568,6 +622,11 @@ CCOMMON_OPT += -mips64r6 -mnan=2008 -mtune=p6600 $(MSA_FLAGS)
FCOMMON_OPT += -mips64r6 -mnan=2008 -mtune=p6600 $(MSA_FLAGS)
endif
ifeq ($(CORE), I6500)
CCOMMON_OPT += -mips64r6 -mnan=2008 -mtune=i6400 $(MSA_FLAGS)
FCOMMON_OPT += -mips64r6 -mnan=2008 -mtune=i6400 $(MSA_FLAGS)
endif
ifeq ($(OSNAME), AIX)
BINARY_DEFINED = 1
endif
@@ -575,12 +634,14 @@ endif
endif
ifndef BINARY_DEFINED
ifneq ($(OSNAME), AIX)
ifdef BINARY64
CCOMMON_OPT += -m64
else
CCOMMON_OPT += -m32
endif
endif
endif
endif
@@ -625,6 +686,7 @@ ifeq ($(F_COMPILER), G77)
CCOMMON_OPT += -DF_INTERFACE_G77
FCOMMON_OPT += -Wall
ifndef NO_BINARY_MODE
ifneq ($(OSNAME), AIX)
ifdef BINARY64
FCOMMON_OPT += -m64
else
@@ -632,10 +694,12 @@ FCOMMON_OPT += -m32
endif
endif
endif
endif
ifeq ($(F_COMPILER), G95)
CCOMMON_OPT += -DF_INTERFACE_G95
FCOMMON_OPT += -Wall
ifneq ($(OSNAME), AIX)
ifndef NO_BINARY_MODE
ifdef BINARY64
FCOMMON_OPT += -m64
@@ -644,6 +708,7 @@ FCOMMON_OPT += -m32
endif
endif
endif
endif
ifeq ($(F_COMPILER), GFORTRAN)
CCOMMON_OPT += -DF_INTERFACE_GFORT
@@ -664,16 +729,20 @@ FCOMMON_OPT += -mabi=32
endif
else
ifdef BINARY64
ifneq ($(OSNAME), AIX)
FCOMMON_OPT += -m64
endif
ifdef INTERFACE64
ifneq ($(INTERFACE64), 0)
FCOMMON_OPT += -fdefault-integer-8
endif
endif
else
ifneq ($(OSNAME), AIX)
FCOMMON_OPT += -m32
endif
endif
endif
ifeq ($(USE_OPENMP), 1)
FCOMMON_OPT += -fopenmp
endif
@@ -687,7 +756,7 @@ FCOMMON_OPT += -i8
endif
endif
ifeq ($(USE_OPENMP), 1)
FCOMMON_OPT += -openmp
FCOMMON_OPT += -fopenmp
endif
endif
@@ -867,6 +936,10 @@ ifeq ($(DYNAMIC_ARCH), 1)
CCOMMON_OPT += -DDYNAMIC_ARCH
endif
ifeq ($(DYNAMIC_OLDER), 1)
CCOMMON_OPT += -DDYNAMIC_OLDER
endif
ifeq ($(NO_LAPACK), 1)
CCOMMON_OPT += -DNO_LAPACK
#Disable LAPACK C interface
@@ -889,6 +962,10 @@ ifeq ($(NO_AVX2), 1)
CCOMMON_OPT += -DNO_AVX2
endif
ifeq ($(NO_AVX512), 1)
CCOMMON_OPT += -DNO_AVX512
endif
ifdef SMP
CCOMMON_OPT += -DSMP_SERVER
@@ -935,6 +1012,8 @@ endif
CCOMMON_OPT += -DMAX_CPU_NUMBER=$(NUM_THREADS)
CCOMMON_OPT += -DMAX_PARALLEL_NUMBER=$(NUM_PARALLEL)
ifdef USE_SIMPLE_THREADED_LEVEL3
CCOMMON_OPT += -DUSE_SIMPLE_THREADED_LEVEL3
endif
@@ -948,9 +1027,15 @@ SYMBOLSUFFIX =
endif
ifndef LIBNAMESUFFIX
LIBPREFIX = lib$(SYMBOLPREFIX)openblas$(SYMBOLSUFFIX)
LIBNAMEBASE = $(SYMBOLPREFIX)openblas$(SYMBOLSUFFIX)
else
LIBPREFIX = lib$(SYMBOLPREFIX)openblas$(SYMBOLSUFFIX)_$(LIBNAMESUFFIX)
LIBNAMEBASE = $(SYMBOLPREFIX)openblas$(SYMBOLSUFFIX)_$(LIBNAMESUFFIX)
endif
ifeq ($(OSNAME), CYGWIN_NT)
LIBPREFIX = cyg$(LIBNAMEBASE)
else
LIBPREFIX = lib$(LIBNAMEBASE)
endif
KERNELDIR = $(TOPDIR)/kernel/$(ARCH)
@@ -1109,6 +1194,7 @@ endif
LIBDLLNAME = $(LIBPREFIX).dll
IMPLIBNAME = lib$(LIBNAMEBASE).dll.a
LIBSONAME = $(LIBNAME:.$(LIBSUFFIX)=.so)
LIBDYNNAME = $(LIBNAME:.$(LIBSUFFIX)=.dylib)
LIBDEFNAME = $(LIBNAME:.$(LIBSUFFIX)=.def)
@@ -1186,6 +1272,7 @@ export MSA_FLAGS
export KERNELDIR
export FUNCTION_PROFILE
export TARGET_CORE
export NO_AVX512
export SGEMM_UNROLL_M
export SGEMM_UNROLL_N
+7
View File
@@ -8,6 +8,13 @@ endif
endif
endif
ifeq ($(CORE), SKYLAKEX)
ifndef NO_AVX512
CCOMMON_OPT += -march=skylake-avx512
FCOMMON_OPT += -march=skylake-avx512
endif
endif
ifeq ($(OSNAME), Interix)
ARFLAGS = -m x64
endif
+141 -94
View File
@@ -2,175 +2,222 @@
[![Join the chat at https://gitter.im/xianyi/OpenBLAS](https://badges.gitter.im/Join%20Chat.svg)](https://gitter.im/xianyi/OpenBLAS?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge)
Travis CI: [![Build Status](https://travis-ci.org/xianyi/OpenBLAS.png?branch=develop)](https://travis-ci.org/xianyi/OpenBLAS)
Travis CI: [![Build Status](https://travis-ci.org/xianyi/OpenBLAS.svg?branch=develop)](https://travis-ci.org/xianyi/OpenBLAS)
AppVeyor: [![Build status](https://ci.appveyor.com/api/projects/status/09sohd35n8nkkx64/branch/develop?svg=true)](https://ci.appveyor.com/project/xianyi/openblas/branch/develop)
## Introduction
OpenBLAS is an optimized BLAS library based on GotoBLAS2 1.13 BSD version.
Please read the documents on OpenBLAS wiki pages <http://github.com/xianyi/OpenBLAS/wiki>.
Please read the documentation on the OpenBLAS wiki pages: <http://github.com/xianyi/OpenBLAS/wiki>.
## Binary Packages
We provide binary packages for the following platform.
We provide official binary packages for the following platform:
* Windows x86/x86_64
You can download them from [file hosting on sourceforge.net](https://sourceforge.net/projects/openblas/files/).
## Installation from Source
Download from project homepage. http://xianyi.github.com/OpenBLAS/
Or, check out codes from git://github.com/xianyi/OpenBLAS.git
Download from project homepage, http://xianyi.github.com/OpenBLAS/, or check out the code
using Git from https://github.com/xianyi/OpenBLAS.git.
### Dependencies
Building OpenBLAS requires the following to be installed:
* GNU Make
* A C compiler, e.g. GCC or Clang
* A Fortran compiler (optional, for LAPACK)
* IBM MASS (optional, see below)
### Normal compile
* type "make" to detect the CPU automatically.
or
* type "make TARGET=xxx" to set target CPU, e.g. "make TARGET=NEHALEM". The full target list is in file TargetList.txt.
Simply invoking `make` (or `gmake` on BSD) will detect the CPU automatically.
To set a specific target CPU, use `make TARGET=xxx`, e.g. `make TARGET=NEHALEM`.
The full target list is in the file `TargetList.txt`.
### Cross compile
Please set CC and FC with the cross toolchains. Then, set HOSTCC with your host C compiler. At last, set TARGET explicitly.
Set `CC` and `FC` to point to the cross toolchains, and set `HOSTCC` to your host C compiler.
The target must be specified explicitly when cross compiling.
Examples:
On X86 box, compile this library for loongson3a CPU.
* On an x86 box, compile this library for a loongson3a CPU:
```sh
make BINARY=64 CC=mips64el-unknown-linux-gnu-gcc FC=mips64el-unknown-linux-gnu-gfortran HOSTCC=gcc TARGET=LOONGSON3A
```
make BINARY=64 CC=mips64el-unknown-linux-gnu-gcc FC=mips64el-unknown-linux-gnu-gfortran HOSTCC=gcc TARGET=LOONGSON3A
On X86 box, compile this library for loongson3a CPU with loongcc (based on Open64) compiler.
make CC=loongcc FC=loongf95 HOSTCC=gcc TARGET=LOONGSON3A CROSS=1 CROSS_SUFFIX=mips64el-st-linux-gnu- NO_LAPACKE=1 NO_SHARED=1 BINARY=32
* On an x86 box, compile this library for a loongson3a CPU with loongcc (based on Open64) compiler:
```sh
make CC=loongcc FC=loongf95 HOSTCC=gcc TARGET=LOONGSON3A CROSS=1 CROSS_SUFFIX=mips64el-st-linux-gnu- NO_LAPACKE=1 NO_SHARED=1 BINARY=32
```
### Debug version
make DEBUG=1
A debug version can be built using `make DEBUG=1`.
### Compile with MASS Support on Power CPU (Optional dependency)
### Compile with MASS support on Power CPU (optional)
[IBM MASS](http://www-01.ibm.com/software/awdtools/mass/linux/mass-linux.html) library consists of a set of mathematical functions for C, C++, and
Fortran-language applications that are tuned for optimum performance on POWER architectures. OpenBLAS with MASS requires 64-bit, little-endian OS on POWER.
The library can be installed as below -
The [IBM MASS](http://www-01.ibm.com/software/awdtools/mass/linux/mass-linux.html) library
consists of a set of mathematical functions for C, C++, and Fortran applications that are
are tuned for optimum performance on POWER architectures.
OpenBLAS with MASS requires a 64-bit, little-endian OS on POWER.
The library can be installed as shown:
* On Ubuntu:
* On Ubuntu:
```sh
wget -q http://public.dhe.ibm.com/software/server/POWER/Linux/xl-compiler/eval/ppc64le/ubuntu/public.gpg -O- | sudo apt-key add -
echo "deb http://public.dhe.ibm.com/software/server/POWER/Linux/xl-compiler/eval/ppc64le/ubuntu/ trusty main" | sudo tee /etc/apt/sources.list.d/ibm-xl-compiler-eval.list
sudo apt-get update
sudo apt-get install libxlmass-devel.8.1.5
```
wget -q http://public.dhe.ibm.com/software/server/POWER/Linux/xl-compiler/eval/ppc64le/ubuntu/public.gpg -O- | sudo apt-key add -</br>
echo "deb http://public.dhe.ibm.com/software/server/POWER/Linux/xl-compiler/eval/ppc64le/ubuntu/ trusty main" | sudo tee /etc/apt/sources.list.d/ibm-xl-compiler-eval.list</br>
sudo apt-get update</br>
sudo apt-get install libxlmass-devel.8.1.5</br>
* On RHEL/CentOS:
```sh
wget http://public.dhe.ibm.com/software/server/POWER/Linux/xl-compiler/eval/ppc64le/rhel7/repodata/repomd.xml.key
sudo rpm --import repomd.xml.key
wget http://public.dhe.ibm.com/software/server/POWER/Linux/xl-compiler/eval/ppc64le/rhel7/ibm-xl-compiler-eval.repo
sudo cp ibm-xl-compiler-eval.repo /etc/yum.repos.d/
sudo yum install libxlmass-devel.8.1.5
```
* On RHEL/CentOS:
After installing the MASS library, compile OpenBLAS with `USE_MASS=1`.
For example, to compile on Power8 with MASS support: `make USE_MASS=1 TARGET=POWER8`.
wget http://public.dhe.ibm.com/software/server/POWER/Linux/xl-compiler/eval/ppc64le/rhel7/repodata/repomd.xml.key</br>
sudo rpm --import repomd.xml.key</br>
wget http://public.dhe.ibm.com/software/server/POWER/Linux/xl-compiler/eval/ppc64le/rhel7/ibm-xl-compiler-eval.repo</br>
sudo cp ibm-xl-compiler-eval.repo /etc/yum.repos.d/</br>
sudo yum install libxlmass-devel.8.1.5</br>
### Install to a specific directory (optional)
After installing MASS library, compile openblas with USE_MASS=1.
Use `PREFIX=` when invoking `make`, for example
Example:
```sh
make install PREFIX=your_installation_directory
```
Compiling on Power8 with MASS support -
The default installation directory is `/opt/OpenBLAS`.
make USE_MASS=1 TARGET=POWER8
## Supported CPUs and Operating Systems
### Install to the directory (optional)
Please read `GotoBLAS_01Readme.txt`.
Example:
### Additional supported CPUs
make install PREFIX=your_installation_directory
#### x86/x86-64
The default directory is /opt/OpenBLAS
## Support CPU & OS
Please read GotoBLAS_01Readme.txt
### Additional support CPU:
#### x86/x86-64:
- **Intel Xeon 56xx (Westmere)**: Used GotoBLAS2 Nehalem codes.
- **Intel Sandy Bridge**: Optimized Level-3 and Level-2 BLAS with AVX on x86-64.
- **Intel Haswell**: Optimized Level-3 and Level-2 BLAS with AVX2 and FMA on x86-64.
- **AMD Bobcat**: Used GotoBLAS2 Barcelona codes.
- **AMD Bulldozer**: x86-64 ?GEMM FMA4 kernels. (Thank Werner Saar)
- **AMD Bulldozer**: x86-64 ?GEMM FMA4 kernels. (Thanks to Werner Saar)
- **AMD PILEDRIVER**: Uses Bulldozer codes with some optimizations.
- **AMD STEAMROLLER**: Uses Bulldozer codes with some optimizations.
#### MIPS64:
#### MIPS64
- **ICT Loongson 3A**: Optimized Level-3 BLAS and the part of Level-1,2.
- **ICT Loongson 3B**: Experimental
#### ARM:
- **ARMV6**: Optimized BLAS for vfpv2 and vfpv3-d16 ( e.g. BCM2835, Cortex M0+ )
- **ARMV7**: Optimized BLAS for vfpv3-d32 ( e.g. Cortex A8, A9 and A15 )
#### ARM
#### ARM64:
- **ARMV8**: Experimental
- **ARMv6**: Optimized BLAS for vfpv2 and vfpv3-d16 (e.g. BCM2835, Cortex M0+)
- **ARMv7**: Optimized BLAS for vfpv3-d32 (e.g. Cortex A8, A9 and A15)
#### ARM64
- **ARMv8**: Experimental
- **ARM Cortex-A57**: Experimental
#### IBM zEnterprise System:
- **Z13**: Optimized Level-3 BLAS
#### PPC/PPC64
- **POWER8**: Optmized Level-3 BLAS and some Level-1, only with `USE_OPENMP=1`
#### IBM zEnterprise System
- **Z13**: Optimized Level-3 BLAS and Level-1,2 (double precision)
### Supported OS
### Support OS:
- **GNU/Linux**
- **MingWin or Visual Studio(CMake)/Windows**: Please read <https://github.com/xianyi/OpenBLAS/wiki/How-to-use-OpenBLAS-in-Microsoft-Visual-Studio>.
- **Darwin/Mac OS X**: Experimental. Although GotoBLAS2 supports Darwin, we are the beginner on Mac OS X.
- **FreeBSD**: Supported by community. We didn't test the library on this OS.
- **Android**: Supported by community. Please read <https://github.com/xianyi/OpenBLAS/wiki/How-to-build-OpenBLAS-for-Android>.
- **MinGW or Visual Studio (CMake)/Windows**: Please read <https://github.com/xianyi/OpenBLAS/wiki/How-to-use-OpenBLAS-in-Microsoft-Visual-Studio>.
- **Darwin/macOS**: Experimental. Although GotoBLAS2 supports Darwin, we are not macOS experts.
- **FreeBSD**: Supported by the community. We don't actively test the library on this OS.
- **OpenBSD**: Supported by the community. We don't actively test the library on this OS.
- **DragonFly BSD**: Supported by the community. We don't actively test the library on this OS.
- **Android**: Supported by the community. Please read <https://github.com/xianyi/OpenBLAS/wiki/How-to-build-OpenBLAS-for-Android>.
## Usages
Link with libopenblas.a or -lopenblas for shared library.
## Usage
### Set the number of threads with environment variables.
Statically link with `libopenblas.a` or dynamically link with `-lopenblas` if OpenBLAS was
compiled as a shared library.
Examples:
### Setting the number of threads using environment variables
export OPENBLAS_NUM_THREADS=4
Environment variables are used to specify a maximum number of threads.
For example,
or
```sh
export OPENBLAS_NUM_THREADS=4
export GOTO_NUM_THREADS=4
export OMP_NUM_THREADS=4
```
export GOTO_NUM_THREADS=4
The priorities are `OPENBLAS_NUM_THREADS` > `GOTO_NUM_THREADS` > `OMP_NUM_THREADS`.
or
If you compile this library with `USE_OPENMP=1`, you should set the `OMP_NUM_THREADS`
environment variable; OpenBLAS ignores `OPENBLAS_NUM_THREADS` and `GOTO_NUM_THREADS` when
compiled with `USE_OPENMP=1`.
export OMP_NUM_THREADS=4
### Setting the number of threads at runtime
The priorities are OPENBLAS_NUM_THREADS > GOTO_NUM_THREADS > OMP_NUM_THREADS.
We provide the following functions to control the number of threads at runtime:
If you compile this lib with USE_OPENMP=1, you should set OMP_NUM_THREADS environment variable. OpenBLAS ignores OPENBLAS_NUM_THREADS and GOTO_NUM_THREADS with USE_OPENMP=1.
```c
void goto_set_num_threads(int num_threads);
void openblas_set_num_threads(int num_threads);
```
### Set the number of threads on runtime.
If you compile this library with `USE_OPENMP=1`, you should use the above functions too.
We provided the below functions to control the number of threads on runtime.
## Reporting bugs
void goto_set_num_threads(int num_threads);
void openblas_set_num_threads(int num_threads);
If you compile this lib with USE_OPENMP=1, you should use the above functions, too.
## Report Bugs
Please add a issue in https://github.com/xianyi/OpenBLAS/issues
Please submit an issue in https://github.com/xianyi/OpenBLAS/issues.
## Contact
* OpenBLAS users mailing list: https://groups.google.com/forum/#!forum/openblas-users
* OpenBLAS developers mailing list: https://groups.google.com/forum/#!forum/openblas-dev
## ChangeLog
Please see Changelog.txt to obtain the differences between GotoBLAS2 1.13 BSD version.
## Change log
Please see Changelog.txt to view the differences between OpenBLAS and GotoBLAS2 1.13 BSD version.
## Troubleshooting
* Please read [Faq](https://github.com/xianyi/OpenBLAS/wiki/Faq) at first.
* Please use gcc version 4.6 and above to compile Sandy Bridge AVX kernels on Linux/MingW/BSD.
* Please use Clang version 3.1 and above to compile the library on Sandy Bridge microarchitecture. The Clang 3.0 will generate the wrong AVX binary code.
* The number of CPUs/Cores should less than or equal to 256. On Linux x86_64(amd64), there is experimental support for up to 1024 CPUs/Cores and 128 numa nodes if you build the library with BIGNUMA=1.
* OpenBLAS does not set processor affinity by default. On Linux, you can enable processor affinity by commenting the line NO_AFFINITY=1 in Makefile.rule. But this may cause [the conflict with R parallel](https://stat.ethz.ch/pipermail/r-sig-hpc/2012-April/001348.html).
* On Loongson 3A. make test would be failed because of pthread_create error. The error code is EAGAIN. However, it will be OK when you run the same testcase on shell.
* Please read the [FAQ](https://github.com/xianyi/OpenBLAS/wiki/Faq) first.
* Please use GCC version 4.6 and above to compile Sandy Bridge AVX kernels on Linux/MinGW/BSD.
* Please use Clang version 3.1 and above to compile the library on Sandy Bridge microarchitecture.
Clang 3.0 will generate the wrong AVX binary code.
* The number of CPUs/cores should less than or equal to 256. On Linux `x86_64` (`amd64`),
there is experimental support for up to 1024 CPUs/cores and 128 numa nodes if you build
the library with `BIGNUMA=1`.
* OpenBLAS does not set processor affinity by default.
On Linux, you can enable processor affinity by commenting out the line `NO_AFFINITY=1` in
Makefile.rule. However, note that this may cause
[a conflict with R parallel](https://stat.ethz.ch/pipermail/r-sig-hpc/2012-April/001348.html).
* On Loongson 3A, `make test` may fail with a `pthread_create` error (`EAGAIN`).
However, it will be okay when you run the same test case on the shell.
## Contributing
1. [Check for open issues](https://github.com/xianyi/OpenBLAS/issues) or open a fresh issue to start a discussion around a feature idea or a bug.
1. Fork the [OpenBLAS](https://github.com/xianyi/OpenBLAS) repository to start making your changes.
1. Write a test which shows that the bug was fixed or that the feature works as expected.
1. Send a pull request. Make sure to add yourself to `CONTRIBUTORS.md`.
1. [Check for open issues](https://github.com/xianyi/OpenBLAS/issues) or open a fresh issue
to start a discussion around a feature idea or a bug.
2. Fork the [OpenBLAS](https://github.com/xianyi/OpenBLAS) repository to start making your changes.
3. Write a test which shows that the bug was fixed or that the feature works as expected.
4. Send a pull request. Make sure to add yourself to `CONTRIBUTORS.md`.
## Donation
Please read [this wiki page](https://github.com/xianyi/OpenBLAS/wiki/Donation).
+6
View File
@@ -20,6 +20,7 @@ DUNNINGTON
NEHALEM
SANDYBRIDGE
HASWELL
SKYLAKEX
ATOM
b)AMD CPU:
@@ -56,6 +57,7 @@ CELL
3.MIPS CPU:
P5600
1004K
4.MIPS64 CPU:
SICORTEX
@@ -63,6 +65,7 @@ LOONGSON3A
LOONGSON3B
I6400
P6600
I6500
5.IA64 CPU:
ITANIUM2
@@ -85,3 +88,6 @@ VULCAN
THUNDERX
THUNDERX2T99
9.System Z:
ZARCH_GENERIC
Z13
+14
View File
@@ -14,6 +14,20 @@ Please build OpenBLAS with larger `NUM_THREADS`. For example, `make
NUM_THREADS=32` or `make NUM_THREADS=64`. In `Makefile.system`, we will set
`MAX_CPU_NUMBER=NUM_THREADS`.
Despite its name, and due to the use of memory buffers in functions like SGEMM,
the setting of NUM_THREADS can be relevant even for a single-threaded build
of OpenBLAS, if such functions get called by multiple threads of a program
that uses OpenBLAS. In some cases, the affected code may simply crash or throw
a segmentation fault without displaying the above warning first.
Note that the number of threads used at runtime can be altered to differ from the
value NUM_THREADS was set to at build time. At runtime, the actual number of
threads can be set anywhere from 1 to the build's NUM_THREADS (note however,
that this does not change the number of memory buffers that will be allocated,
which is set at build time). The number of threads for a process can be set by
using the mechanisms described below.
#### How can I use OpenBLAS in multi-threaded applications?
If your application is already multi-threaded, it will conflict with OpenBLAS
+37 -15
View File
@@ -5,6 +5,8 @@ version: 0.2.19.{build}
platform:
- x64
os: Visual Studio 2017
configuration: Release
clone_folder: c:\projects\OpenBLAS
@@ -12,33 +14,53 @@ clone_folder: c:\projects\OpenBLAS
init:
- git config --global core.autocrlf input
build:
project: OpenBLAS.sln
clone_depth: 5
#branches to build
branches:
only:
- master
- develop
- cmake
skip_tags: true
matrix:
fast_finish: true
fast_finish: false
skip_commits:
# Add [av skip] to commit messages
message: /\[av skip\]/
environment:
global:
CONDA_INSTALL_LOCN: C:\\Miniconda36-x64
matrix:
- COMPILER: clang-cl
WITH_FORTRAN: yes
- COMPILER: clang-cl
DYNAMIC_ARCH: ON
WITH_FORTRAN: no
- COMPILER: cl
install:
- if [%COMPILER%]==[clang-cl] call %CONDA_INSTALL_LOCN%\Scripts\activate.bat
- if [%COMPILER%]==[clang-cl] conda config --add channels conda-forge --force
- if [%COMPILER%]==[clang-cl] conda install --yes --quiet clangdev cmake
- if [%WITH_FORTRAN%]==[no] conda install --yes --quiet ninja
- if [%WITH_FORTRAN%]==[yes] conda install --yes --quiet -c isuruf kitware-ninja
- if [%WITH_FORTRAN%]==[yes] conda install --yes --quiet flang
- if [%COMPILER%]==[clang-cl] call "C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Auxiliary\Build\vcvarsall.bat" x64
- if [%COMPILER%]==[clang-cl] set "LIB=%CONDA_INSTALL_LOCN%\Library\lib;%LIB%"
- if [%COMPILER%]==[clang-cl] set "CPATH=%CONDA_INSTALL_LOCN%\Library\include;%CPATH%"
before_build:
- echo Running cmake...
- cd c:\projects\OpenBLAS
- cmake -G "Visual Studio 12 Win64" .
- ps: if (-Not (Test-Path .\build)) { mkdir build }
- cd build
- if [%COMPILER%]==[cl] cmake -G "Visual Studio 15 2017 Win64" ..
- if [%WITH_FORTRAN%]==[no] cmake -G "Ninja" -DCMAKE_CXX_COMPILER=clang-cl -DCMAKE_C_COMPILER=clang-cl ..
- if [%WITH_FORTRAN%]==[yes] cmake -G "Ninja" -DCMAKE_CXX_COMPILER=clang-cl -DCMAKE_C_COMPILER=clang-cl -DCMAKE_Fortran_COMPILER=flang -DBUILD_WITHOUT_LAPACK=no -DNOFORTRAN=0 ..
- if [%DYNAMIC_ARCH%]==[ON] cmake -DDYNAMIC_ARCH=ON ..
build_script:
- cmake --build .
test_script:
- echo Running Test
- cd c:\projects\OpenBLAS\utest
- cd utest
- openblas_utest
+80 -56
View File
@@ -121,13 +121,15 @@ static void *huge_malloc(BLASLONG size){
int main(int argc, char *argv[]){
FLOAT *a, *b, *c;
FLOAT alpha[] = {1.0, 1.0};
FLOAT alpha[] = {1.0, 0.0};
FLOAT beta [] = {0.0, 0.0};
char trans='N';
blasint m, n, i, j;
char transa = 'N';
char transb = 'N';
blasint m, n, k, i, j, lda, ldb, ldc;
int loops = 1;
int has_param_n=0;
int l;
int has_param_m = 0;
int has_param_n = 0;
int has_param_k = 0;
char *p;
int from = 1;
@@ -135,86 +137,108 @@ int main(int argc, char *argv[]){
int step = 1;
struct timeval start, stop;
double time1,timeg;
double time1, timeg;
argc--;argv++;
if (argc > 0) { from = atol(*argv); argc--; argv++;}
if (argc > 0) { to = MAX(atol(*argv), from); argc--; argv++;}
if (argc > 0) { step = atol(*argv); argc--; argv++;}
if (argc > 0) { from = atol(*argv); argc--; argv++; }
if (argc > 0) { to = MAX(atol(*argv), from); argc--; argv++; }
if (argc > 0) { step = atol(*argv); argc--; argv++; }
if ((p = getenv("OPENBLAS_TRANS"))) trans=*p;
fprintf(stderr, "From : %3d To : %3d Step=%d : Trans=%c\n", from, to, step, trans);
if (( a = (FLOAT *)malloc(sizeof(FLOAT) * to * to * COMPSIZE)) == NULL){
fprintf(stderr,"Out of Memory!!\n");exit(1);
if ((p = getenv("OPENBLAS_TRANS"))) {
transa=*p;
transb=*p;
}
if (( b = (FLOAT *)malloc(sizeof(FLOAT) * to * to * COMPSIZE)) == NULL){
fprintf(stderr,"Out of Memory!!\n");exit(1);
if ((p = getenv("OPENBLAS_TRANSA"))) {
transa=*p;
}
if (( c = (FLOAT *)malloc(sizeof(FLOAT) * to * to * COMPSIZE)) == NULL){
fprintf(stderr,"Out of Memory!!\n");exit(1);
if ((p = getenv("OPENBLAS_TRANSB"))) {
transb=*p;
}
TOUPPER(transa);
TOUPPER(transb);
fprintf(stderr, "From : %3d To : %3d Step=%d : Transa=%c : Transb=%c\n", from, to, step, transa, transb);
p = getenv("OPENBLAS_LOOPS");
if ( p != NULL )
loops = atoi(p);
if ( p != NULL ) {
loops = atoi(p);
}
if ((p = getenv("OPENBLAS_PARAM_M"))) {
m = atoi(p);
has_param_m=1;
} else {
m = to;
}
if ((p = getenv("OPENBLAS_PARAM_N"))) {
n = atoi(p);
has_param_n=1;
n = atoi(p);
has_param_n=1;
} else {
n = to;
}
if ((p = getenv("OPENBLAS_PARAM_K"))) {
k = atoi(p);
has_param_k=1;
} else {
k = to;
}
if (( a = (FLOAT *)malloc(sizeof(FLOAT) * m * k * COMPSIZE)) == NULL) {
fprintf(stderr,"Out of Memory!!\n");exit(1);
}
if (( b = (FLOAT *)malloc(sizeof(FLOAT) * k * n * COMPSIZE)) == NULL) {
fprintf(stderr,"Out of Memory!!\n");exit(1);
}
if (( c = (FLOAT *)malloc(sizeof(FLOAT) * m * n * COMPSIZE)) == NULL) {
fprintf(stderr,"Out of Memory!!\n");exit(1);
}
#ifdef linux
srandom(getpid());
#endif
for (i = 0; i < m * k * COMPSIZE; i++) {
a[i] = ((FLOAT) rand() / (FLOAT) RAND_MAX) - 0.5;
}
for (i = 0; i < k * n * COMPSIZE; i++) {
b[i] = ((FLOAT) rand() / (FLOAT) RAND_MAX) - 0.5;
}
for (i = 0; i < m * n * COMPSIZE; i++) {
c[i] = ((FLOAT) rand() / (FLOAT) RAND_MAX) - 0.5;
}
for(j = 0; j < to; j++){
for(i = 0; i < to * COMPSIZE; i++){
a[i + j * to * COMPSIZE] = ((FLOAT) rand() / (FLOAT) RAND_MAX) - 0.5;
b[i + j * to * COMPSIZE] = ((FLOAT) rand() / (FLOAT) RAND_MAX) - 0.5;
c[i + j * to * COMPSIZE] = ((FLOAT) rand() / (FLOAT) RAND_MAX) - 0.5;
}
}
fprintf(stderr, " SIZE Flops Time\n");
for(m = from; m <= to; m += step)
{
fprintf(stderr, " SIZE Flops Time\n");
for (i = from; i <= to; i += step) {
timeg=0;
if ( has_param_n == 1 && n <= m )
n=n;
else
n=m;
if (!has_param_m) { m = i; }
if (!has_param_n) { n = i; }
if (!has_param_k) { k = i; }
if (transa == 'N') { lda = m; }
else { lda = k; }
if (transb == 'N') { ldb = k; }
else { ldb = n; }
ldc = m;
fprintf(stderr, " %6dx%d : ", (int)m, (int)n);
fprintf(stderr, " M=%4d, N=%4d, K=%4d : ", (int)m, (int)n, (int)k);
gettimeofday( &start, (struct timezone *)0);
for (l=0; l<loops; l++)
{
GEMM (&trans, &trans, &m, &n, &m, alpha, a, &m, b, &m, beta, c, &m );
for (j=0; j<loops; j++) {
GEMM (&transa, &transb, &m, &n, &k, alpha, a, &lda, b, &ldb, beta, c, &ldc);
}
gettimeofday( &stop, (struct timezone *)0);
time1 = (double)(stop.tv_sec - start.tv_sec) + (double)((stop.tv_usec - start.tv_usec)) * 1.e-6;
gettimeofday( &stop, (struct timezone *)0);
time1 = (double)(stop.tv_sec - start.tv_sec) + (double)((stop.tv_usec - start.tv_usec)) * 1.e-6;
timeg = time1/loops;
fprintf(stderr,
" %10.2f MFlops %10.6f sec\n",
COMPSIZE * COMPSIZE * 2. * (double)m * (double)m * (double)n / timeg * 1.e-6, time1);
COMPSIZE * COMPSIZE * 2. * (double)k * (double)m * (double)n / timeg * 1.e-6, time1);
}
return 0;
+18
View File
@@ -54,6 +54,8 @@ $compiler = GCC if ($compiler eq "");
$os = Linux if ($data =~ /OS_LINUX/);
$os = FreeBSD if ($data =~ /OS_FREEBSD/);
$os = NetBSD if ($data =~ /OS_NETBSD/);
$os = OpenBSD if ($data =~ /OS_OPENBSD/);
$os = DragonFly if ($data =~ /OS_DRAGONFLY/);
$os = Darwin if ($data =~ /OS_DARWIN/);
$os = SunOS if ($data =~ /OS_SUNOS/);
$os = AIX if ($data =~ /OS_AIX/);
@@ -199,6 +201,21 @@ $architecture = zarch if ($data =~ /ARCH_ZARCH/);
$binformat = bin32;
$binformat = bin64 if ($data =~ /BINARY_64/);
$no_avx512= 0;
if (($architecture eq "x86") || ($architecture eq "x86_64")) {
$code = '"vbroadcastss -4 * 4(%rsi), %zmm2"';
print $tmpf "int main(void){ __asm__ volatile($code); }\n";
$args = " -march=skylake-avx512 -o $tmpf.o -x c $tmpf";
my @cmd = ("$compiler_name $args >/dev/null 2>/dev/null");
system(@cmd) == 0;
if ($? != 0) {
$no_avx512 = 1;
} else {
$no_avx512 = 0;
}
unlink("tmpf.o");
}
$data = `$compiler_name -S ctest1.c && grep globl ctest1.s | head -n 1 && rm -f ctest1.s`;
$data =~ /globl\s([_\.]*)(.*)/;
@@ -286,6 +303,7 @@ print MAKEFILE "CROSS=1\n" if $cross != 0;
print MAKEFILE "CEXTRALIB=$linker_L $linker_l $linker_a\n";
print MAKEFILE "HAVE_MSA=1\n" if $have_msa eq 1;
print MAKEFILE "MSA_FLAGS=$msa_flags\n" if $have_msa eq 1;
print MAKEFILE "NO_AVX512=1\n" if $no_avx512 eq 1;
$os =~ tr/[a-z]/[A-Z]/;
$architecture =~ tr/[a-z]/[A-Z]/;
+89 -84
View File
@@ -57,45 +57,50 @@ double cblas_dsdot (OPENBLAS_CONST blasint n, OPENBLAS_CONST float *x, OPENBLAS_
float cblas_sdot(OPENBLAS_CONST blasint n, OPENBLAS_CONST float *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST float *y, OPENBLAS_CONST blasint incy);
double cblas_ddot(OPENBLAS_CONST blasint n, OPENBLAS_CONST double *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST double *y, OPENBLAS_CONST blasint incy);
openblas_complex_float cblas_cdotu(OPENBLAS_CONST blasint n, OPENBLAS_CONST float *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST float *y, OPENBLAS_CONST blasint incy);
openblas_complex_float cblas_cdotc(OPENBLAS_CONST blasint n, OPENBLAS_CONST float *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST float *y, OPENBLAS_CONST blasint incy);
openblas_complex_double cblas_zdotu(OPENBLAS_CONST blasint n, OPENBLAS_CONST double *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST double *y, OPENBLAS_CONST blasint incy);
openblas_complex_double cblas_zdotc(OPENBLAS_CONST blasint n, OPENBLAS_CONST double *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST double *y, OPENBLAS_CONST blasint incy);
openblas_complex_float cblas_cdotu(OPENBLAS_CONST blasint n, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST void *y, OPENBLAS_CONST blasint incy);
openblas_complex_float cblas_cdotc(OPENBLAS_CONST blasint n, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST void *y, OPENBLAS_CONST blasint incy);
openblas_complex_double cblas_zdotu(OPENBLAS_CONST blasint n, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST void *y, OPENBLAS_CONST blasint incy);
openblas_complex_double cblas_zdotc(OPENBLAS_CONST blasint n, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST void *y, OPENBLAS_CONST blasint incy);
void cblas_cdotu_sub(OPENBLAS_CONST blasint n, OPENBLAS_CONST float *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST float *y, OPENBLAS_CONST blasint incy, openblas_complex_float *ret);
void cblas_cdotc_sub(OPENBLAS_CONST blasint n, OPENBLAS_CONST float *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST float *y, OPENBLAS_CONST blasint incy, openblas_complex_float *ret);
void cblas_zdotu_sub(OPENBLAS_CONST blasint n, OPENBLAS_CONST double *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST double *y, OPENBLAS_CONST blasint incy, openblas_complex_double *ret);
void cblas_zdotc_sub(OPENBLAS_CONST blasint n, OPENBLAS_CONST double *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST double *y, OPENBLAS_CONST blasint incy, openblas_complex_double *ret);
void cblas_cdotu_sub(OPENBLAS_CONST blasint n, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST void *y, OPENBLAS_CONST blasint incy, void *ret);
void cblas_cdotc_sub(OPENBLAS_CONST blasint n, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST void *y, OPENBLAS_CONST blasint incy, void *ret);
void cblas_zdotu_sub(OPENBLAS_CONST blasint n, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST void *y, OPENBLAS_CONST blasint incy, void *ret);
void cblas_zdotc_sub(OPENBLAS_CONST blasint n, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST void *y, OPENBLAS_CONST blasint incy, void *ret);
float cblas_sasum (OPENBLAS_CONST blasint n, OPENBLAS_CONST float *x, OPENBLAS_CONST blasint incx);
double cblas_dasum (OPENBLAS_CONST blasint n, OPENBLAS_CONST double *x, OPENBLAS_CONST blasint incx);
float cblas_scasum(OPENBLAS_CONST blasint n, OPENBLAS_CONST float *x, OPENBLAS_CONST blasint incx);
double cblas_dzasum(OPENBLAS_CONST blasint n, OPENBLAS_CONST double *x, OPENBLAS_CONST blasint incx);
float cblas_scasum(OPENBLAS_CONST blasint n, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx);
double cblas_dzasum(OPENBLAS_CONST blasint n, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx);
float cblas_snrm2 (OPENBLAS_CONST blasint N, OPENBLAS_CONST float *X, OPENBLAS_CONST blasint incX);
double cblas_dnrm2 (OPENBLAS_CONST blasint N, OPENBLAS_CONST double *X, OPENBLAS_CONST blasint incX);
float cblas_scnrm2(OPENBLAS_CONST blasint N, OPENBLAS_CONST float *X, OPENBLAS_CONST blasint incX);
double cblas_dznrm2(OPENBLAS_CONST blasint N, OPENBLAS_CONST double *X, OPENBLAS_CONST blasint incX);
float cblas_scnrm2(OPENBLAS_CONST blasint N, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX);
double cblas_dznrm2(OPENBLAS_CONST blasint N, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX);
CBLAS_INDEX cblas_isamax(OPENBLAS_CONST blasint n, OPENBLAS_CONST float *x, OPENBLAS_CONST blasint incx);
CBLAS_INDEX cblas_idamax(OPENBLAS_CONST blasint n, OPENBLAS_CONST double *x, OPENBLAS_CONST blasint incx);
CBLAS_INDEX cblas_icamax(OPENBLAS_CONST blasint n, OPENBLAS_CONST float *x, OPENBLAS_CONST blasint incx);
CBLAS_INDEX cblas_izamax(OPENBLAS_CONST blasint n, OPENBLAS_CONST double *x, OPENBLAS_CONST blasint incx);
CBLAS_INDEX cblas_icamax(OPENBLAS_CONST blasint n, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx);
CBLAS_INDEX cblas_izamax(OPENBLAS_CONST blasint n, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx);
CBLAS_INDEX cblas_isamin(OPENBLAS_CONST blasint n, OPENBLAS_CONST float *x, OPENBLAS_CONST blasint incx);
CBLAS_INDEX cblas_idamin(OPENBLAS_CONST blasint n, OPENBLAS_CONST double *x, OPENBLAS_CONST blasint incx);
CBLAS_INDEX cblas_icamin(OPENBLAS_CONST blasint n, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx);
CBLAS_INDEX cblas_izamin(OPENBLAS_CONST blasint n, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx);
void cblas_saxpy(OPENBLAS_CONST blasint n, OPENBLAS_CONST float alpha, OPENBLAS_CONST float *x, OPENBLAS_CONST blasint incx, float *y, OPENBLAS_CONST blasint incy);
void cblas_daxpy(OPENBLAS_CONST blasint n, OPENBLAS_CONST double alpha, OPENBLAS_CONST double *x, OPENBLAS_CONST blasint incx, double *y, OPENBLAS_CONST blasint incy);
void cblas_caxpy(OPENBLAS_CONST blasint n, OPENBLAS_CONST float *alpha, OPENBLAS_CONST float *x, OPENBLAS_CONST blasint incx, float *y, OPENBLAS_CONST blasint incy);
void cblas_zaxpy(OPENBLAS_CONST blasint n, OPENBLAS_CONST double *alpha, OPENBLAS_CONST double *x, OPENBLAS_CONST blasint incx, double *y, OPENBLAS_CONST blasint incy);
void cblas_caxpy(OPENBLAS_CONST blasint n, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx, void *y, OPENBLAS_CONST blasint incy);
void cblas_zaxpy(OPENBLAS_CONST blasint n, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx, void *y, OPENBLAS_CONST blasint incy);
void cblas_scopy(OPENBLAS_CONST blasint n, OPENBLAS_CONST float *x, OPENBLAS_CONST blasint incx, float *y, OPENBLAS_CONST blasint incy);
void cblas_dcopy(OPENBLAS_CONST blasint n, OPENBLAS_CONST double *x, OPENBLAS_CONST blasint incx, double *y, OPENBLAS_CONST blasint incy);
void cblas_ccopy(OPENBLAS_CONST blasint n, OPENBLAS_CONST float *x, OPENBLAS_CONST blasint incx, float *y, OPENBLAS_CONST blasint incy);
void cblas_zcopy(OPENBLAS_CONST blasint n, OPENBLAS_CONST double *x, OPENBLAS_CONST blasint incx, double *y, OPENBLAS_CONST blasint incy);
void cblas_ccopy(OPENBLAS_CONST blasint n, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx, void *y, OPENBLAS_CONST blasint incy);
void cblas_zcopy(OPENBLAS_CONST blasint n, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx, void *y, OPENBLAS_CONST blasint incy);
void cblas_sswap(OPENBLAS_CONST blasint n, float *x, OPENBLAS_CONST blasint incx, float *y, OPENBLAS_CONST blasint incy);
void cblas_dswap(OPENBLAS_CONST blasint n, double *x, OPENBLAS_CONST blasint incx, double *y, OPENBLAS_CONST blasint incy);
void cblas_cswap(OPENBLAS_CONST blasint n, float *x, OPENBLAS_CONST blasint incx, float *y, OPENBLAS_CONST blasint incy);
void cblas_zswap(OPENBLAS_CONST blasint n, double *x, OPENBLAS_CONST blasint incx, double *y, OPENBLAS_CONST blasint incy);
void cblas_cswap(OPENBLAS_CONST blasint n, void *x, OPENBLAS_CONST blasint incx, void *y, OPENBLAS_CONST blasint incy);
void cblas_zswap(OPENBLAS_CONST blasint n, void *x, OPENBLAS_CONST blasint incx, void *y, OPENBLAS_CONST blasint incy);
void cblas_srot(OPENBLAS_CONST blasint N, float *X, OPENBLAS_CONST blasint incX, float *Y, OPENBLAS_CONST blasint incY, OPENBLAS_CONST float c, OPENBLAS_CONST float s);
void cblas_drot(OPENBLAS_CONST blasint N, double *X, OPENBLAS_CONST blasint incX, double *Y, OPENBLAS_CONST blasint incY, OPENBLAS_CONST double c, OPENBLAS_CONST double s);
@@ -111,59 +116,59 @@ void cblas_drotmg(double *d1, double *d2, double *b1, OPENBLAS_CONST double b2,
void cblas_sscal(OPENBLAS_CONST blasint N, OPENBLAS_CONST float alpha, float *X, OPENBLAS_CONST blasint incX);
void cblas_dscal(OPENBLAS_CONST blasint N, OPENBLAS_CONST double alpha, double *X, OPENBLAS_CONST blasint incX);
void cblas_cscal(OPENBLAS_CONST blasint N, OPENBLAS_CONST float *alpha, float *X, OPENBLAS_CONST blasint incX);
void cblas_zscal(OPENBLAS_CONST blasint N, OPENBLAS_CONST double *alpha, double *X, OPENBLAS_CONST blasint incX);
void cblas_csscal(OPENBLAS_CONST blasint N, OPENBLAS_CONST float alpha, float *X, OPENBLAS_CONST blasint incX);
void cblas_zdscal(OPENBLAS_CONST blasint N, OPENBLAS_CONST double alpha, double *X, OPENBLAS_CONST blasint incX);
void cblas_cscal(OPENBLAS_CONST blasint N, OPENBLAS_CONST void *alpha, void *X, OPENBLAS_CONST blasint incX);
void cblas_zscal(OPENBLAS_CONST blasint N, OPENBLAS_CONST void *alpha, void *X, OPENBLAS_CONST blasint incX);
void cblas_csscal(OPENBLAS_CONST blasint N, OPENBLAS_CONST float alpha, void *X, OPENBLAS_CONST blasint incX);
void cblas_zdscal(OPENBLAS_CONST blasint N, OPENBLAS_CONST double alpha, void *X, OPENBLAS_CONST blasint incX);
void cblas_sgemv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_TRANSPOSE trans, OPENBLAS_CONST blasint m, OPENBLAS_CONST blasint n,
OPENBLAS_CONST float alpha, OPENBLAS_CONST float *a, OPENBLAS_CONST blasint lda, OPENBLAS_CONST float *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST float beta, float *y, OPENBLAS_CONST blasint incy);
void cblas_dgemv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_TRANSPOSE trans, OPENBLAS_CONST blasint m, OPENBLAS_CONST blasint n,
OPENBLAS_CONST double alpha, OPENBLAS_CONST double *a, OPENBLAS_CONST blasint lda, OPENBLAS_CONST double *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST double beta, double *y, OPENBLAS_CONST blasint incy);
void cblas_cgemv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_TRANSPOSE trans, OPENBLAS_CONST blasint m, OPENBLAS_CONST blasint n,
OPENBLAS_CONST float *alpha, OPENBLAS_CONST float *a, OPENBLAS_CONST blasint lda, OPENBLAS_CONST float *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST float *beta, float *y, OPENBLAS_CONST blasint incy);
OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *a, OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST void *beta, void *y, OPENBLAS_CONST blasint incy);
void cblas_zgemv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_TRANSPOSE trans, OPENBLAS_CONST blasint m, OPENBLAS_CONST blasint n,
OPENBLAS_CONST double *alpha, OPENBLAS_CONST double *a, OPENBLAS_CONST blasint lda, OPENBLAS_CONST double *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST double *beta, double *y, OPENBLAS_CONST blasint incy);
OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *a, OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx, OPENBLAS_CONST void *beta, void *y, OPENBLAS_CONST blasint incy);
void cblas_sger (OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST float alpha, OPENBLAS_CONST float *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST float *Y, OPENBLAS_CONST blasint incY, float *A, OPENBLAS_CONST blasint lda);
void cblas_dger (OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST double alpha, OPENBLAS_CONST double *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST double *Y, OPENBLAS_CONST blasint incY, double *A, OPENBLAS_CONST blasint lda);
void cblas_cgeru(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST float *alpha, OPENBLAS_CONST float *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST float *Y, OPENBLAS_CONST blasint incY, float *A, OPENBLAS_CONST blasint lda);
void cblas_cgerc(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST float *alpha, OPENBLAS_CONST float *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST float *Y, OPENBLAS_CONST blasint incY, float *A, OPENBLAS_CONST blasint lda);
void cblas_zgeru(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST double *alpha, OPENBLAS_CONST double *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST double *Y, OPENBLAS_CONST blasint incY, double *A, OPENBLAS_CONST blasint lda);
void cblas_zgerc(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST double *alpha, OPENBLAS_CONST double *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST double *Y, OPENBLAS_CONST blasint incY, double *A, OPENBLAS_CONST blasint lda);
void cblas_cgeru(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST void *Y, OPENBLAS_CONST blasint incY, void *A, OPENBLAS_CONST blasint lda);
void cblas_cgerc(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST void *Y, OPENBLAS_CONST blasint incY, void *A, OPENBLAS_CONST blasint lda);
void cblas_zgeru(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST void *Y, OPENBLAS_CONST blasint incY, void *A, OPENBLAS_CONST blasint lda);
void cblas_zgerc(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST void *Y, OPENBLAS_CONST blasint incY, void *A, OPENBLAS_CONST blasint lda);
void cblas_strsv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint N, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, float *X, OPENBLAS_CONST blasint incX);
void cblas_dtrsv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint N, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, double *X, OPENBLAS_CONST blasint incX);
void cblas_ctrsv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint N, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, float *X, OPENBLAS_CONST blasint incX);
void cblas_ztrsv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint N, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, double *X, OPENBLAS_CONST blasint incX);
void cblas_ctrsv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint N, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, void *X, OPENBLAS_CONST blasint incX);
void cblas_ztrsv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint N, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, void *X, OPENBLAS_CONST blasint incX);
void cblas_strmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint N, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, float *X, OPENBLAS_CONST blasint incX);
void cblas_dtrmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint N, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, double *X, OPENBLAS_CONST blasint incX);
void cblas_ctrmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint N, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, float *X, OPENBLAS_CONST blasint incX);
void cblas_ztrmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint N, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, double *X, OPENBLAS_CONST blasint incX);
void cblas_ctrmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint N, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, void *X, OPENBLAS_CONST blasint incX);
void cblas_ztrmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint N, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, void *X, OPENBLAS_CONST blasint incX);
void cblas_ssyr(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST float alpha, OPENBLAS_CONST float *X, OPENBLAS_CONST blasint incX, float *A, OPENBLAS_CONST blasint lda);
void cblas_dsyr(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST double alpha, OPENBLAS_CONST double *X, OPENBLAS_CONST blasint incX, double *A, OPENBLAS_CONST blasint lda);
void cblas_cher(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST float alpha, OPENBLAS_CONST float *X, OPENBLAS_CONST blasint incX, float *A, OPENBLAS_CONST blasint lda);
void cblas_zher(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST double alpha, OPENBLAS_CONST double *X, OPENBLAS_CONST blasint incX, double *A, OPENBLAS_CONST blasint lda);
void cblas_cher(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST float alpha, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX, void *A, OPENBLAS_CONST blasint lda);
void cblas_zher(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST double alpha, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX, void *A, OPENBLAS_CONST blasint lda);
void cblas_ssyr2(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo,OPENBLAS_CONST blasint N, OPENBLAS_CONST float alpha, OPENBLAS_CONST float *X,
OPENBLAS_CONST blasint incX, OPENBLAS_CONST float *Y, OPENBLAS_CONST blasint incY, float *A, OPENBLAS_CONST blasint lda);
void cblas_dsyr2(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST double alpha, OPENBLAS_CONST double *X,
OPENBLAS_CONST blasint incX, OPENBLAS_CONST double *Y, OPENBLAS_CONST blasint incY, double *A, OPENBLAS_CONST blasint lda);
void cblas_cher2(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST float *alpha, OPENBLAS_CONST float *X, OPENBLAS_CONST blasint incX,
OPENBLAS_CONST float *Y, OPENBLAS_CONST blasint incY, float *A, OPENBLAS_CONST blasint lda);
void cblas_zher2(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST double *alpha, OPENBLAS_CONST double *X, OPENBLAS_CONST blasint incX,
OPENBLAS_CONST double *Y, OPENBLAS_CONST blasint incY, double *A, OPENBLAS_CONST blasint lda);
void cblas_cher2(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX,
OPENBLAS_CONST void *Y, OPENBLAS_CONST blasint incY, void *A, OPENBLAS_CONST blasint lda);
void cblas_zher2(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX,
OPENBLAS_CONST void *Y, OPENBLAS_CONST blasint incY, void *A, OPENBLAS_CONST blasint lda);
void cblas_sgbmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N,
OPENBLAS_CONST blasint KL, OPENBLAS_CONST blasint KU, OPENBLAS_CONST float alpha, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST float *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST float beta, float *Y, OPENBLAS_CONST blasint incY);
void cblas_dgbmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N,
OPENBLAS_CONST blasint KL, OPENBLAS_CONST blasint KU, OPENBLAS_CONST double alpha, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST double *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST double beta, double *Y, OPENBLAS_CONST blasint incY);
void cblas_cgbmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N,
OPENBLAS_CONST blasint KL, OPENBLAS_CONST blasint KU, OPENBLAS_CONST float *alpha, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST float *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST float *beta, float *Y, OPENBLAS_CONST blasint incY);
OPENBLAS_CONST blasint KL, OPENBLAS_CONST blasint KU, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST void *beta, void *Y, OPENBLAS_CONST blasint incY);
void cblas_zgbmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N,
OPENBLAS_CONST blasint KL, OPENBLAS_CONST blasint KU, OPENBLAS_CONST double *alpha, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST double *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST double *beta, double *Y, OPENBLAS_CONST blasint incY);
OPENBLAS_CONST blasint KL, OPENBLAS_CONST blasint KU, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST void *beta, void *Y, OPENBLAS_CONST blasint incY);
void cblas_ssbmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST float alpha, OPENBLAS_CONST float *A,
OPENBLAS_CONST blasint lda, OPENBLAS_CONST float *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST float beta, float *Y, OPENBLAS_CONST blasint incY);
@@ -176,45 +181,45 @@ void cblas_stbmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLA
void cblas_dtbmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag,
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, double *X, OPENBLAS_CONST blasint incX);
void cblas_ctbmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag,
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, float *X, OPENBLAS_CONST blasint incX);
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, void *X, OPENBLAS_CONST blasint incX);
void cblas_ztbmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag,
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, double *X, OPENBLAS_CONST blasint incX);
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, void *X, OPENBLAS_CONST blasint incX);
void cblas_stbsv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag,
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, float *X, OPENBLAS_CONST blasint incX);
void cblas_dtbsv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag,
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, double *X, OPENBLAS_CONST blasint incX);
void cblas_ctbsv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag,
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, float *X, OPENBLAS_CONST blasint incX);
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, void *X, OPENBLAS_CONST blasint incX);
void cblas_ztbsv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag,
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, double *X, OPENBLAS_CONST blasint incX);
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, void *X, OPENBLAS_CONST blasint incX);
void cblas_stpmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag,
OPENBLAS_CONST blasint N, OPENBLAS_CONST float *Ap, float *X, OPENBLAS_CONST blasint incX);
void cblas_dtpmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag,
OPENBLAS_CONST blasint N, OPENBLAS_CONST double *Ap, double *X, OPENBLAS_CONST blasint incX);
void cblas_ctpmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag,
OPENBLAS_CONST blasint N, OPENBLAS_CONST float *Ap, float *X, OPENBLAS_CONST blasint incX);
OPENBLAS_CONST blasint N, OPENBLAS_CONST void *Ap, void *X, OPENBLAS_CONST blasint incX);
void cblas_ztpmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag,
OPENBLAS_CONST blasint N, OPENBLAS_CONST double *Ap, double *X, OPENBLAS_CONST blasint incX);
OPENBLAS_CONST blasint N, OPENBLAS_CONST void *Ap, void *X, OPENBLAS_CONST blasint incX);
void cblas_stpsv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag,
OPENBLAS_CONST blasint N, OPENBLAS_CONST float *Ap, float *X, OPENBLAS_CONST blasint incX);
void cblas_dtpsv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag,
OPENBLAS_CONST blasint N, OPENBLAS_CONST double *Ap, double *X, OPENBLAS_CONST blasint incX);
void cblas_ctpsv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag,
OPENBLAS_CONST blasint N, OPENBLAS_CONST float *Ap, float *X, OPENBLAS_CONST blasint incX);
OPENBLAS_CONST blasint N, OPENBLAS_CONST void *Ap, void *X, OPENBLAS_CONST blasint incX);
void cblas_ztpsv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_DIAG Diag,
OPENBLAS_CONST blasint N, OPENBLAS_CONST double *Ap, double *X, OPENBLAS_CONST blasint incX);
OPENBLAS_CONST blasint N, OPENBLAS_CONST void *Ap, void *X, OPENBLAS_CONST blasint incX);
void cblas_ssymv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST float alpha, OPENBLAS_CONST float *A,
OPENBLAS_CONST blasint lda, OPENBLAS_CONST float *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST float beta, float *Y, OPENBLAS_CONST blasint incY);
void cblas_dsymv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST double alpha, OPENBLAS_CONST double *A,
OPENBLAS_CONST blasint lda, OPENBLAS_CONST double *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST double beta, double *Y, OPENBLAS_CONST blasint incY);
void cblas_chemv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST float *alpha, OPENBLAS_CONST float *A,
OPENBLAS_CONST blasint lda, OPENBLAS_CONST float *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST float *beta, float *Y, OPENBLAS_CONST blasint incY);
void cblas_zhemv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST double *alpha, OPENBLAS_CONST double *A,
OPENBLAS_CONST blasint lda, OPENBLAS_CONST double *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST double *beta, double *Y, OPENBLAS_CONST blasint incY);
void cblas_chemv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A,
OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST void *beta, void *Y, OPENBLAS_CONST blasint incY);
void cblas_zhemv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A,
OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST void *beta, void *Y, OPENBLAS_CONST blasint incY);
void cblas_sspmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST float alpha, OPENBLAS_CONST float *Ap,
@@ -225,36 +230,36 @@ void cblas_dspmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLA
void cblas_sspr(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST float alpha, OPENBLAS_CONST float *X, OPENBLAS_CONST blasint incX, float *Ap);
void cblas_dspr(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST double alpha, OPENBLAS_CONST double *X, OPENBLAS_CONST blasint incX, double *Ap);
void cblas_chpr(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST float alpha, OPENBLAS_CONST float *X, OPENBLAS_CONST blasint incX, float *A);
void cblas_zhpr(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST double alpha, OPENBLAS_CONST double *X,OPENBLAS_CONST blasint incX, double *A);
void cblas_chpr(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST float alpha, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX, void *A);
void cblas_zhpr(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST double alpha, OPENBLAS_CONST void *X,OPENBLAS_CONST blasint incX, void *A);
void cblas_sspr2(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST float alpha, OPENBLAS_CONST float *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST float *Y, OPENBLAS_CONST blasint incY, float *A);
void cblas_dspr2(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST double alpha, OPENBLAS_CONST double *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST double *Y, OPENBLAS_CONST blasint incY, double *A);
void cblas_chpr2(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST float *alpha, OPENBLAS_CONST float *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST float *Y, OPENBLAS_CONST blasint incY, float *Ap);
void cblas_zhpr2(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST double *alpha, OPENBLAS_CONST double *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST double *Y, OPENBLAS_CONST blasint incY, double *Ap);
void cblas_chpr2(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST void *Y, OPENBLAS_CONST blasint incY, void *Ap);
void cblas_zhpr2(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST void *Y, OPENBLAS_CONST blasint incY, void *Ap);
void cblas_chbmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K,
OPENBLAS_CONST float *alpha, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST float *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST float *beta, float *Y, OPENBLAS_CONST blasint incY);
OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST void *beta, void *Y, OPENBLAS_CONST blasint incY);
void cblas_zhbmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K,
OPENBLAS_CONST double *alpha, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST double *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST double *beta, double *Y, OPENBLAS_CONST blasint incY);
OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST void *beta, void *Y, OPENBLAS_CONST blasint incY);
void cblas_chpmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N,
OPENBLAS_CONST float *alpha, OPENBLAS_CONST float *Ap, OPENBLAS_CONST float *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST float *beta, float *Y, OPENBLAS_CONST blasint incY);
OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *Ap, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST void *beta, void *Y, OPENBLAS_CONST blasint incY);
void cblas_zhpmv(OPENBLAS_CONST enum CBLAS_ORDER order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint N,
OPENBLAS_CONST double *alpha, OPENBLAS_CONST double *Ap, OPENBLAS_CONST double *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST double *beta, double *Y, OPENBLAS_CONST blasint incY);
OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *Ap, OPENBLAS_CONST void *X, OPENBLAS_CONST blasint incX, OPENBLAS_CONST void *beta, void *Y, OPENBLAS_CONST blasint incY);
void cblas_sgemm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransB, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K,
OPENBLAS_CONST float alpha, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST float *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST float beta, float *C, OPENBLAS_CONST blasint ldc);
void cblas_dgemm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransB, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K,
OPENBLAS_CONST double alpha, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST double *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST double beta, double *C, OPENBLAS_CONST blasint ldc);
void cblas_cgemm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransB, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K,
OPENBLAS_CONST float *alpha, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST float *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST float *beta, float *C, OPENBLAS_CONST blasint ldc);
OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST void *beta, void *C, OPENBLAS_CONST blasint ldc);
void cblas_cgemm3m(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransB, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K,
OPENBLAS_CONST float *alpha, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST float *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST float *beta, float *C, OPENBLAS_CONST blasint ldc);
OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST void *beta, void *C, OPENBLAS_CONST blasint ldc);
void cblas_zgemm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransB, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K,
OPENBLAS_CONST double *alpha, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST double *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST double *beta, double *C, OPENBLAS_CONST blasint ldc);
OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST void *beta, void *C, OPENBLAS_CONST blasint ldc);
void cblas_zgemm3m(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransB, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K,
OPENBLAS_CONST double *alpha, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST double *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST double *beta, double *C, OPENBLAS_CONST blasint ldc);
OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST void *beta, void *C, OPENBLAS_CONST blasint ldc);
void cblas_ssymm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_SIDE Side, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N,
@@ -262,60 +267,60 @@ void cblas_ssymm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLA
void cblas_dsymm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_SIDE Side, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N,
OPENBLAS_CONST double alpha, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST double *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST double beta, double *C, OPENBLAS_CONST blasint ldc);
void cblas_csymm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_SIDE Side, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N,
OPENBLAS_CONST float *alpha, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST float *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST float *beta, float *C, OPENBLAS_CONST blasint ldc);
OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST void *beta, void *C, OPENBLAS_CONST blasint ldc);
void cblas_zsymm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_SIDE Side, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N,
OPENBLAS_CONST double *alpha, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST double *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST double *beta, double *C, OPENBLAS_CONST blasint ldc);
OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST void *beta, void *C, OPENBLAS_CONST blasint ldc);
void cblas_ssyrk(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE Trans,
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST float alpha, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST float beta, float *C, OPENBLAS_CONST blasint ldc);
void cblas_dsyrk(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE Trans,
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST double alpha, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST double beta, double *C, OPENBLAS_CONST blasint ldc);
void cblas_csyrk(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE Trans,
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST float *alpha, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST float *beta, float *C, OPENBLAS_CONST blasint ldc);
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *beta, void *C, OPENBLAS_CONST blasint ldc);
void cblas_zsyrk(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE Trans,
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST double *alpha, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST double *beta, double *C, OPENBLAS_CONST blasint ldc);
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *beta, void *C, OPENBLAS_CONST blasint ldc);
void cblas_ssyr2k(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE Trans,
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST float alpha, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST float *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST float beta, float *C, OPENBLAS_CONST blasint ldc);
void cblas_dsyr2k(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE Trans,
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST double alpha, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST double *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST double beta, double *C, OPENBLAS_CONST blasint ldc);
void cblas_csyr2k(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE Trans,
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST float *alpha, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST float *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST float *beta, float *C, OPENBLAS_CONST blasint ldc);
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST void *beta, void *C, OPENBLAS_CONST blasint ldc);
void cblas_zsyr2k(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE Trans,
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST double *alpha, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST double *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST double *beta, double *C, OPENBLAS_CONST blasint ldc);
OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST void *beta, void *C, OPENBLAS_CONST blasint ldc);
void cblas_strmm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_SIDE Side, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA,
OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST float alpha, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, float *B, OPENBLAS_CONST blasint ldb);
void cblas_dtrmm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_SIDE Side, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA,
OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST double alpha, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, double *B, OPENBLAS_CONST blasint ldb);
void cblas_ctrmm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_SIDE Side, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA,
OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST float *alpha, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, float *B, OPENBLAS_CONST blasint ldb);
OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, void *B, OPENBLAS_CONST blasint ldb);
void cblas_ztrmm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_SIDE Side, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA,
OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST double *alpha, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, double *B, OPENBLAS_CONST blasint ldb);
OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, void *B, OPENBLAS_CONST blasint ldb);
void cblas_strsm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_SIDE Side, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA,
OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST float alpha, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, float *B, OPENBLAS_CONST blasint ldb);
void cblas_dtrsm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_SIDE Side, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA,
OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST double alpha, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, double *B, OPENBLAS_CONST blasint ldb);
void cblas_ctrsm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_SIDE Side, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA,
OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST float *alpha, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, float *B, OPENBLAS_CONST blasint ldb);
OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, void *B, OPENBLAS_CONST blasint ldb);
void cblas_ztrsm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_SIDE Side, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA,
OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST double *alpha, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, double *B, OPENBLAS_CONST blasint ldb);
OPENBLAS_CONST enum CBLAS_DIAG Diag, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, void *B, OPENBLAS_CONST blasint ldb);
void cblas_chemm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_SIDE Side, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N,
OPENBLAS_CONST float *alpha, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST float *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST float *beta, float *C, OPENBLAS_CONST blasint ldc);
OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST void *beta, void *C, OPENBLAS_CONST blasint ldc);
void cblas_zhemm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_SIDE Side, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N,
OPENBLAS_CONST double *alpha, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST double *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST double *beta, double *C, OPENBLAS_CONST blasint ldc);
OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST void *beta, void *C, OPENBLAS_CONST blasint ldc);
void cblas_cherk(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE Trans, OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K,
OPENBLAS_CONST float alpha, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST float beta, float *C, OPENBLAS_CONST blasint ldc);
OPENBLAS_CONST float alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST float beta, void *C, OPENBLAS_CONST blasint ldc);
void cblas_zherk(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE Trans, OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K,
OPENBLAS_CONST double alpha, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST double beta, double *C, OPENBLAS_CONST blasint ldc);
OPENBLAS_CONST double alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST double beta, void *C, OPENBLAS_CONST blasint ldc);
void cblas_cher2k(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE Trans, OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K,
OPENBLAS_CONST float *alpha, OPENBLAS_CONST float *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST float *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST float beta, float *C, OPENBLAS_CONST blasint ldc);
OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST float beta, void *C, OPENBLAS_CONST blasint ldc);
void cblas_zher2k(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_UPLO Uplo, OPENBLAS_CONST enum CBLAS_TRANSPOSE Trans, OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K,
OPENBLAS_CONST double *alpha, OPENBLAS_CONST double *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST double *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST double beta, double *C, OPENBLAS_CONST blasint ldc);
OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST void *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST double beta, void *C, OPENBLAS_CONST blasint ldc);
void cblas_xerbla(blasint p, char *rout, char *form, ...);
@@ -325,9 +330,9 @@ void cblas_saxpby(OPENBLAS_CONST blasint n, OPENBLAS_CONST float alpha, OPENBLAS
void cblas_daxpby(OPENBLAS_CONST blasint n, OPENBLAS_CONST double alpha, OPENBLAS_CONST double *x, OPENBLAS_CONST blasint incx,OPENBLAS_CONST double beta, double *y, OPENBLAS_CONST blasint incy);
void cblas_caxpby(OPENBLAS_CONST blasint n, OPENBLAS_CONST float *alpha, OPENBLAS_CONST float *x, OPENBLAS_CONST blasint incx,OPENBLAS_CONST float *beta, float *y, OPENBLAS_CONST blasint incy);
void cblas_caxpby(OPENBLAS_CONST blasint n, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx,OPENBLAS_CONST void *beta, void *y, OPENBLAS_CONST blasint incy);
void cblas_zaxpby(OPENBLAS_CONST blasint n, OPENBLAS_CONST double *alpha, OPENBLAS_CONST double *x, OPENBLAS_CONST blasint incx,OPENBLAS_CONST double *beta, double *y, OPENBLAS_CONST blasint incy);
void cblas_zaxpby(OPENBLAS_CONST blasint n, OPENBLAS_CONST void *alpha, OPENBLAS_CONST void *x, OPENBLAS_CONST blasint incx,OPENBLAS_CONST void *beta, void *y, OPENBLAS_CONST blasint incy);
void cblas_somatcopy(OPENBLAS_CONST enum CBLAS_ORDER CORDER, OPENBLAS_CONST enum CBLAS_TRANSPOSE CTRANS, OPENBLAS_CONST blasint crows, OPENBLAS_CONST blasint ccols, OPENBLAS_CONST float calpha, OPENBLAS_CONST float *a,
OPENBLAS_CONST blasint clda, float *b, OPENBLAS_CONST blasint cldb);
+79
View File
@@ -0,0 +1,79 @@
# OpenBLASConfig.cmake
# --------------------
#
# OpenBLAS cmake module.
# This module sets the following variables in your project::
#
# OpenBLAS_FOUND - true if OpenBLAS and all required components found on the system
# OpenBLAS_VERSION - OpenBLAS version in format Major.Minor.Release
# OpenBLAS_INCLUDE_DIRS - Directory where OpenBLAS header is located.
# OpenBLAS_INCLUDE_DIR - same as DIRS
# OpenBLAS_LIBRARIES - OpenBLAS library to link against.
# OpenBLAS_LIBRARY - same as LIBRARIES
#
#
# Available components::
#
## shared - search for only shared library
## static - search for only static library
# serial - search for unthreaded library
# pthread - search for native pthread threaded library
# openmp - search for OpenMP threaded library
#
#
# Exported targets::
#
# If OpenBLAS is found, this module defines the following :prop_tgt:`IMPORTED`
## target. Target is shared _or_ static, so, for both, use separate, not
## overlapping, installations. ::
#
# OpenBLAS::OpenBLAS - the main OpenBLAS library #with header & defs attached.
#
#
# Suggested usage::
#
# find_package(OpenBLAS)
# find_package(OpenBLAS 0.2.20 EXACT CONFIG REQUIRED COMPONENTS pthread)
#
#
# The following variables can be set to guide the search for this package::
#
# OpenBLAS_DIR - CMake variable, set to directory containing this Config file
# CMAKE_PREFIX_PATH - CMake variable, set to root directory of this package
# PATH - environment variable, set to bin directory of this package
# CMAKE_DISABLE_FIND_PACKAGE_OpenBLAS - CMake variable, disables
# find_package(OpenBLAS) when not REQUIRED, perhaps to force internal build
@PACKAGE_INIT@
set(PN OpenBLAS)
# need to check that the @USE_*@ evaluate to something cmake can perform boolean logic upon
if(@USE_OPENMP@)
set(${PN}_openmp_FOUND 1)
elseif(@USE_THREAD@)
set(${PN}_pthread_FOUND 1)
else()
set(${PN}_serial_FOUND 1)
endif()
check_required_components(${PN})
#-----------------------------------------------------------------------------
# Don't include targets if this file is being picked up by another
# project which has already built this as a subproject
#-----------------------------------------------------------------------------
if(NOT TARGET ${PN}::OpenBLAS)
include("${CMAKE_CURRENT_LIST_DIR}/${PN}Targets.cmake")
get_property(_loc TARGET ${PN}::OpenBLAS PROPERTY LOCATION)
set(${PN}_LIBRARY ${_loc})
get_property(_ill TARGET ${PN}::OpenBLAS PROPERTY INTERFACE_LINK_LIBRARIES)
set(${PN}_LIBRARIES ${_ill})
get_property(_id TARGET ${PN}::OpenBLAS PROPERTY INCLUDE_DIRECTORIES)
set(${PN}_INCLUDE_DIR ${_id})
get_property(_iid TARGET ${PN}::OpenBLAS PROPERTY INTERFACE_INCLUDE_DIRECTORIES)
set(${PN}_INCLUDE_DIRS ${_iid})
endif()
+32 -37
View File
@@ -3,9 +3,9 @@
## Description: Ported from portion of OpenBLAS/Makefile.system
## Sets various variables based on architecture.
if (${ARCH} STREQUAL "x86" OR ${ARCH} STREQUAL "x86_64")
if (X86 OR X86_64)
if (${ARCH} STREQUAL "x86")
if (X86)
if (NOT BINARY)
set(NO_BINARY_MODE 1)
endif ()
@@ -33,47 +33,42 @@ if (${CMAKE_C_COMPILER_ID} STREQUAL "Intel")
endif ()
if (USE_OPENMP)
if (${CMAKE_C_COMPILER_ID} STREQUAL "GNU" OR ${CMAKE_C_COMPILER_ID} STREQUAL "LSB")
set(CCOMMON_OPT "${CCOMMON_OPT} -fopenmp")
endif ()
if (${CMAKE_C_COMPILER_ID} STREQUAL "Clang")
message(WARNING "Clang doesn't support OpenMP yet.")
set(CCOMMON_OPT "${CCOMMON_OPT} -fopenmp")
endif ()
if (${CMAKE_C_COMPILER_ID} STREQUAL "Intel")
set(CCOMMON_OPT "${CCOMMON_OPT} -openmp")
endif ()
if (${CMAKE_C_COMPILER_ID} STREQUAL "PGI")
set(CCOMMON_OPT "${CCOMMON_OPT} -mp")
endif ()
if (${CMAKE_C_COMPILER_ID} STREQUAL "OPEN64")
set(CCOMMON_OPT "${CCOMMON_OPT} -mp")
set(CEXTRALIB "${CEXTRALIB} -lstdc++")
endif ()
if (${CMAKE_C_COMPILER_ID} STREQUAL "PATHSCALE")
set(CCOMMON_OPT "${CCOMMON_OPT} -mp")
endif ()
# USE_SIMPLE_THREADED_LEVEL3 = 1
# NO_AFFINITY = 1
find_package(OpenMP REQUIRED)
if (OpenMP_FOUND)
set(CCOMMON_OPT "${CCOMMON_OPT} ${OpenMP_C_FLAGS} -DUSE_OPENMP")
set(FCOMMON_OPT "${FCOMMON_OPT} ${OpenMP_Fortran_FLAGS}")
endif()
endif ()
if (DYNAMIC_ARCH)
if (${ARCH} STREQUAL "x86")
set(DYNAMIC_CORE "KATMAI COPPERMINE NORTHWOOD PRESCOTT BANIAS CORE2 PENRYN DUNNINGTON NEHALEM ATHLON OPTERON OPTERON_SSE3 BARCELONA BOBCAT ATOM NANO")
if (X86)
set(DYNAMIC_CORE KATMAI COPPERMINE NORTHWOOD PRESCOTT BANIAS CORE2 PENRYN DUNNINGTON NEHALEM ATHLON OPTERON OPTERON_SSE3 BARCELONA BOBCAT ATOM NANO)
endif ()
if (${ARCH} STREQUAL "x86_64")
set(DYNAMIC_CORE "PRESCOTT CORE2 PENRYN DUNNINGTON NEHALEM OPTERON OPTERON_SSE3 BARCELONA BOBCAT ATOM NANO")
if (X86_64)
set(DYNAMIC_CORE PRESCOTT CORE2)
if (DYNAMIC_OLDER)
set (DYNAMIC_CORE ${DYNAMIC_CORE} PENRYN DUNNINGTON)
endif ()
set (DYNAMIC_CORE ${DYNAMIC_CORE} NEHALEM)
if (DYNAMIC_OLDER)
set (DYNAMIC_CORE ${DYNAMIC_CORE} OPTERON OPTERON_SSE3)
endif ()
set (DYNAMIC_CORE ${DYNAMIC_CORE} BARCELONA)
if (DYNAMIC_OLDER)
set (DYNAMIC_CORE ${DYNAMIC_CORE} BOBCAT ATOM NANO)
endif ()
if (NOT NO_AVX)
set(DYNAMIC_CORE "${DYNAMIC_CORE} SANDYBRIDGE BULLDOZER PILEDRIVER STEAMROLLER")
set(DYNAMIC_CORE ${DYNAMIC_CORE} SANDYBRIDGE BULLDOZER PILEDRIVER STEAMROLLER EXCAVATOR)
endif ()
if (NOT NO_AVX2)
set(DYNAMIC_CORE "${DYNAMIC_CORE} HASWELL ZEN")
set(DYNAMIC_CORE ${DYNAMIC_CORE} HASWELL ZEN)
endif ()
if (NOT NO_AVX512)
set(DYNAMIC_CORE ${DYNAMIC_CORE} SKYLAKEX)
endif ()
endif ()
@@ -94,7 +89,7 @@ if (${ARCH} STREQUAL "ia64")
endif ()
endif ()
if (${ARCH} STREQUAL "mips64")
if (MIPS64)
set(NO_BINARY_MODE 1)
endif ()
@@ -103,12 +98,12 @@ if (${ARCH} STREQUAL "alpha")
set(BINARY_DEFINED 1)
endif ()
if (${ARCH} STREQUAL "arm")
if (ARM)
set(NO_BINARY_MODE 1)
set(BINARY_DEFINED 1)
endif ()
if (${ARCH} STREQUAL "arm64")
if (ARM64)
set(NO_BINARY_MODE 1)
set(BINARY_DEFINED 1)
endif ()
-98
View File
@@ -1,98 +0,0 @@
##
## Author: Hank Anderson <hank@statease.com>
## Description: Ported from the OpenBLAS/c_check perl script.
## This is triggered by prebuild.cmake and runs before any of the code is built.
## Creates config.h and Makefile.conf.
# CMake vars set by this file:
# OSNAME (use CMAKE_SYSTEM_NAME)
# ARCH
# C_COMPILER (use CMAKE_C_COMPILER)
# BINARY32
# BINARY64
# FU
# CROSS_SUFFIX
# CROSS
# CEXTRALIB
# Defines set by this file:
# OS_
# ARCH_
# C_
# __32BIT__
# __64BIT__
# FUNDERSCORE
# PTHREAD_CREATE_FUNC
# N.B. c_check (and ctest.c) is not cross-platform, so instead try to use CMake variables.
set(FU "")
if(APPLE)
set(FU "_")
elseif(MSVC)
set(FU "_")
elseif(UNIX)
set(FU "")
endif()
# Convert CMake vars into the format that OpenBLAS expects
string(TOUPPER ${CMAKE_SYSTEM_NAME} HOST_OS)
if (${HOST_OS} STREQUAL "WINDOWS")
set(HOST_OS WINNT)
endif ()
# added by hpa - check size of void ptr to detect 64-bit compile
if (NOT DEFINED BINARY)
set(BINARY 32)
if (CMAKE_SIZEOF_VOID_P EQUAL 8)
set(BINARY 64)
endif ()
endif ()
if (BINARY EQUAL 64)
set(BINARY64 1)
else ()
set(BINARY32 1)
endif ()
# CMake docs define these:
# CMAKE_SYSTEM_PROCESSOR - The name of the CPU CMake is building for.
# CMAKE_HOST_SYSTEM_PROCESSOR - The name of the CPU CMake is running on.
#
# TODO: CMAKE_SYSTEM_PROCESSOR doesn't seem to be correct - instead get it from the compiler a la c_check
set(ARCH ${CMAKE_SYSTEM_PROCESSOR})
if (${ARCH} STREQUAL "AMD64")
set(ARCH "x86_64")
endif ()
# If you are using a 32-bit compiler on a 64-bit system CMAKE_SYSTEM_PROCESSOR will be wrong
if (${ARCH} STREQUAL "x86_64" AND BINARY EQUAL 32)
set(ARCH x86)
endif ()
if (${ARCH} STREQUAL "X86")
set(ARCH x86)
endif ()
if (${ARCH} MATCHES "ppc")
set(ARCH power)
endif ()
set(COMPILER_ID ${CMAKE_CXX_COMPILER_ID})
if (${COMPILER_ID} STREQUAL "GNU")
set(COMPILER_ID "GCC")
endif ()
string(TOUPPER ${ARCH} UC_ARCH)
file(WRITE ${TARGET_CONF}
"#define OS_${HOST_OS}\t1\n"
"#define ARCH_${UC_ARCH}\t1\n"
"#define C_${COMPILER_ID}\t1\n"
"#define __${BINARY}BIT__\t1\n"
"#define FUNDERSCORE\t${FU}\n")
if (${HOST_OS} STREQUAL "WINDOWSSTORE")
file(APPEND ${TARGET_CONF}
"#define OS_WINNT\t1\n")
endif ()
+6 -11
View File
@@ -15,7 +15,7 @@ if (${CMAKE_C_COMPILER} STREQUAL "GNU" OR ${CMAKE_C_COMPILER} STREQUAL "LSB" OR
if (NO_BINARY_MODE)
if (${ARCH} STREQUAL "mips64")
if (MIPS64)
if (BINARY64)
set(CCOMMON_OPT "${CCOMMON_OPT} -mabi=64")
else ()
@@ -24,17 +24,12 @@ if (${CMAKE_C_COMPILER} STREQUAL "GNU" OR ${CMAKE_C_COMPILER} STREQUAL "LSB" OR
set(BINARY_DEFINED 1)
endif ()
if (${CORE} STREQUAL "LOONGSON3A")
if (${CORE} STREQUAL "LOONGSON3A" OR ${CORE} STREQUAL "LOONGSON3B")
set(CCOMMON_OPT "${CCOMMON_OPT} -march=mips64")
set(FCOMMON_OPT "${FCOMMON_OPT} -march=mips64")
endif ()
if (${CORE} STREQUAL "LOONGSON3B")
set(CCOMMON_OPT "${CCOMMON_OPT} -march=mips64")
set(FCOMMON_OPT "${FCOMMON_OPT} -march=mips64")
endif ()
if (${OSNAME} STREQUAL "AIX")
if (CMAKE_SYSTEM_NAME STREQUAL "AIX")
set(BINARY_DEFINED 1)
endif ()
endif ()
@@ -66,7 +61,7 @@ endif ()
if (${CMAKE_C_COMPILER} STREQUAL "OPEN64")
if (${ARCH} STREQUAL "mips64")
if (MIPS64)
if (NOT BINARY64)
set(CCOMMON_OPT "${CCOMMON_OPT} -n32")
@@ -94,10 +89,10 @@ endif ()
if (${CMAKE_C_COMPILER} STREQUAL "SUN")
set(CCOMMON_OPT "${CCOMMON_OPT} -w")
if (${ARCH} STREQUAL "x86")
if (X86)
set(CCOMMON_OPT "${CCOMMON_OPT} -m32")
else ()
set(FCOMMON_OPT "${FCOMMON_OPT} -m64")
set(CCOMMON_OPT "${CCOMMON_OPT} -m64")
endif ()
endif ()
+2 -1
View File
@@ -51,7 +51,8 @@ else()
endif()
add_custom_command(
TARGET ${OpenBLAS_LIBNAME} PRE_LINK
OUTPUT ${PROJECT_BINARY_DIR}/openblas.def
#TARGET ${OpenBLAS_LIBNAME} PRE_LINK
COMMAND perl
ARGS "${PROJECT_SOURCE_DIR}/exports/gensymbol" "win2k" "${ARCH_IN}" "dummy" "${EXPRECISION_IN}" "${NO_CBLAS_IN}" "${NO_LAPACK_IN}" "${NO_LAPACKE_IN}" "${NEED2UNDERSCORES_IN}" "${ONLY_CBLAS_IN}" "${SYMBOLPREFIX}" "${SYMBOLSUFFIX}" > "${PROJECT_BINARY_DIR}/openblas.def"
COMMENT "Create openblas.def file"
+2 -12
View File
@@ -20,12 +20,6 @@
# NEEDBUNDERSCORE
# NEED2UNDERSCORES
if (MSVC)
# had to do this for MSVC, else CMake automatically assumes I have ifort... -hpa
include(CMakeForceCompiler)
CMAKE_FORCE_Fortran_COMPILER(gfortran GNU)
endif ()
if (NOT NO_LAPACK)
enable_language(Fortran)
else()
@@ -34,17 +28,13 @@ else()
endif()
if (NOT ONLY_CBLAS)
# N.B. f_check is not cross-platform, so instead try to use CMake variables
# run f_check (appends to TARGET files)
# message(STATUS "Running f_check...")
# execute_process(COMMAND perl f_check ${TARGET_MAKE} ${TARGET_CONF} ${CMAKE_Fortran_COMPILER}
# WORKING_DIRECTORY ${PROJECT_SOURCE_DIR})
# TODO: detect whether underscore needed, set #defines and BU appropriately - use try_compile
# TODO: set FEXTRALIB flags a la f_check?
set(BU "_")
file(APPEND ${TARGET_CONF}
file(APPEND ${TARGET_CONF_TEMP}
"#define BUNDERSCORE _\n"
"#define NEEDBUNDERSCORE 1\n"
"#define NEED2UNDERSCORES 0\n")
@@ -56,7 +46,7 @@ else ()
set(NO_FBLAS 1)
#set(F_COMPILER GFORTRAN) # CMake handles the fortran compiler
set(BU "_")
file(APPEND ${TARGET_CONF}
file(APPEND ${TARGET_CONF_TEMP}
"#define BUNDERSCORE _\n"
"#define NEEDBUNDERSCORE 1\n")
endif()
+6 -11
View File
@@ -5,13 +5,8 @@
if (${F_COMPILER} STREQUAL "FLANG")
set(CCOMMON_OPT "${CCOMMON_OPT} -DF_INTERFACE_FLANG")
if (BINARY64)
if (INTERFACE64)
set(FCOMMON_OPT "${FCOMMON_OPT} -i8")
endif ()
set(FCOMMON_OPT "${FCOMMON_OPT} -Wall")
else ()
set(FCOMMON_OPT "${FCOMMON_OPT} -Wall")
if (BINARY64 AND INTERFACE64)
set(FCOMMON_OPT "${FCOMMON_OPT} -i8")
endif ()
if (USE_OPENMP)
set(FCOMMON_OPT "${FCOMMON_OPT} -fopenmp")
@@ -50,7 +45,7 @@ if (${F_COMPILER} STREQUAL "GFORTRAN")
set(EXTRALIB "{EXTRALIB} -lgfortran")
endif ()
if (NO_BINARY_MODE)
if (${ARCH} STREQUAL "mips64")
if (MIPS64)
if (BINARY64)
set(FCOMMON_OPT "${FCOMMON_OPT} -mabi=64")
else ()
@@ -130,7 +125,7 @@ if (${F_COMPILER} STREQUAL "PATHSCALE")
endif ()
endif ()
if (NOT ${ARCH} STREQUAL "mips64")
if (NOT MIPS64)
if (NOT BINARY64)
set(FCOMMON_OPT "${FCOMMON_OPT} -m32")
else ()
@@ -158,7 +153,7 @@ if (${F_COMPILER} STREQUAL "OPEN64")
endif ()
endif ()
if (${ARCH} STREQUAL "mips64")
if (MIPS64)
if (NOT BINARY64)
set(FCOMMON_OPT "${FCOMMON_OPT} -n32")
@@ -189,7 +184,7 @@ endif ()
if (${F_COMPILER} STREQUAL "SUN")
set(CCOMMON_OPT "${CCOMMON_OPT} -DF_INTERFACE_SUN")
if (${ARCH} STREQUAL "x86")
if (X86)
set(FCOMMON_OPT "${FCOMMON_OPT} -m32")
else ()
set(FCOMMON_OPT "${FCOMMON_OPT} -m64")
+459 -361
View File
@@ -1,387 +1,485 @@
# Sources for compiling lapack-netlib. Can't use CMakeLists.txt because lapack-netlib already has its own cmake files.
set(ALLAUX
ilaenv.f ieeeck.f lsamen.f xerbla_array.f iparmq.f
ilaprec.f ilatrans.f ilauplo.f iladiag.f iparam2stage.F chla_transtype.f
../INSTALL/ilaver.f ../INSTALL/slamch.f
)
set(ALLAUX ilaenv.f ilaenv2stage.f ieeeck.f lsamen.f iparmq.f iparam2stage.F
ilaprec.f ilatrans.f ilauplo.f iladiag.f chla_transtype.f
../INSTALL/ilaver.f xerbla_array.f
../INSTALL/slamch.f)
set(SCLAUX
sbdsdc.f
sbdsqr.f sdisna.f slabad.f slacpy.f sladiv.f slae2.f slaebz.f
slaed0.f slaed1.f slaed2.f slaed3.f slaed4.f slaed5.f slaed6.f
slaed7.f slaed8.f slaed9.f slaeda.f slaev2.f slagtf.f
slagts.f slamrg.f slanst.f
slapy2.f slapy3.f slarnv.f
slarra.f slarrb.f slarrc.f slarrd.f slarre.f slarrf.f slarrj.f
slarrk.f slarrr.f slaneg.f
slartg.f slaruv.f slas2.f slascl.f
slasd0.f slasd1.f slasd2.f slasd3.f slasd4.f slasd5.f slasd6.f
slasd7.f slasd8.f slasda.f slasdq.f slasdt.f
slaset.f slasq1.f slasq2.f slasq3.f slasq4.f slasq5.f slasq6.f
slasr.f slasrt.f slassq.f slasv2.f spttrf.f sstebz.f sstedc.f
ssteqr.f ssterf.f slaisnan.f sisnan.f
slartgp.f slartgs.f
../INSTALL/second_${TIMER}.f
)
sbdsdc.f
sbdsqr.f sdisna.f slabad.f slacpy.f sladiv.f slae2.f slaebz.f
slaed0.f slaed1.f slaed2.f slaed3.f slaed4.f slaed5.f slaed6.f
slaed7.f slaed8.f slaed9.f slaeda.f slaev2.f slagtf.f
slagts.f slamrg.f slanst.f
slapy2.f slapy3.f slarnv.f
slarra.f slarrb.f slarrc.f slarrd.f slarre.f slarrf.f slarrj.f
slarrk.f slarrr.f slaneg.f
slartg.f slaruv.f slas2.f slascl.f
slasd0.f slasd1.f slasd2.f slasd3.f slasd4.f slasd5.f slasd6.f
slasd7.f slasd8.f slasda.f slasdq.f slasdt.f
slaset.f slasq1.f slasq2.f slasq3.f slasq4.f slasq5.f slasq6.f
slasr.f slasrt.f slassq.f slasv2.f spttrf.f sstebz.f sstedc.f
ssteqr.f ssterf.f slaisnan.f sisnan.f
slartgp.f slartgs.f
../INSTALL/second_${TIMER}.f)
set(DZLAUX
dbdsdc.f dbdsvdx.f
dbdsqr.f ddisna.f dlabad.f dlacpy.f dladiv.f dlae2.f dlaebz.f
dlaed0.f dlaed1.f dlaed2.f dlaed3.f dlaed4.f dlaed5.f dlaed6.f
dlaed7.f dlaed8.f dlaed9.f dlaeda.f dlaev2.f dlagtf.f
dlagts.f dlamrg.f dlanst.f
dlapy2.f dlapy3.f dlarnv.f
dlarra.f dlarrb.f dlarrc.f dlarrd.f dlarre.f dlarrf.f dlarrj.f
dlarrk.f dlarrr.f dlaneg.f
dlartg.f dlaruv.f dlas2.f dlascl.f
dlasd0.f dlasd1.f dlasd2.f dlasd3.f dlasd4.f dlasd5.f dlasd6.f
dlasd7.f dlasd8.f dlasda.f dlasdq.f dlasdt.f
dlaset.f dlasq1.f dlasq2.f dlasq3.f dlasq4.f dlasq5.f dlasq6.f
dlasr.f dlasrt.f dlassq.f dlasv2.f dpttrf.f dstebz.f dstedc.f
dsteqr.f dsterf.f dlaisnan.f disnan.f
dlartgp.f dlartgs.f
../INSTALL/dlamch.f ../INSTALL/dsecnd_${TIMER}.f
dgelq.f dgelqt.f dgelqt3.f dgemlq.f dgemlqt.f dgemqr.f dgeqr.f
dgetsls.f dlamswlq.f dlamtsqr.f dlaswlq.f dlatsqr.f dtplqt.f
dtplqt2.f dtpmlqt.f dsysv_aa.f dsytrf_aa.f dsytrs_aa.f dlasyf_aa.f
dsytf2_rk.f dlasyf_rk.f dsytrf_rk.f dsytrs_3.f dsycon_3.f dsytri_3.f
dsytri_3x.f dsysv_rk.f dsb2st_kernels.f dsbev_2stage.f dsbevd_2stage.f
dsbevx_2stage.f dsyev_2stage.f dsyevd_2stage.f dsyevr_2stage.f
dsyevx_2stage.f dsygv_2stage.f dsytrd_2stage.f dsytrd_sb2st.F
dsytrd_sy2sb.f dlarfy.f
)
dbdsdc.f
dbdsqr.f ddisna.f dlabad.f dlacpy.f dladiv.f dlae2.f dlaebz.f
dlaed0.f dlaed1.f dlaed2.f dlaed3.f dlaed4.f dlaed5.f dlaed6.f
dlaed7.f dlaed8.f dlaed9.f dlaeda.f dlaev2.f dlagtf.f
dlagts.f dlamrg.f dlanst.f
dlapy2.f dlapy3.f dlarnv.f
dlarra.f dlarrb.f dlarrc.f dlarrd.f dlarre.f dlarrf.f dlarrj.f
dlarrk.f dlarrr.f dlaneg.f
dlartg.f dlaruv.f dlas2.f dlascl.f
dlasd0.f dlasd1.f dlasd2.f dlasd3.f dlasd4.f dlasd5.f dlasd6.f
dlasd7.f dlasd8.f dlasda.f dlasdq.f dlasdt.f
dlaset.f dlasq1.f dlasq2.f dlasq3.f dlasq4.f dlasq5.f dlasq6.f
dlasr.f dlasrt.f dlassq.f dlasv2.f dpttrf.f dstebz.f dstedc.f
dsteqr.f dsterf.f dlaisnan.f disnan.f
dlartgp.f dlartgs.f
../INSTALL/dlamch.f ../INSTALL/dsecnd_${TIMER}.f)
set(SLASRC
sbdsvdx.f sgbbrd.f sgbcon.f sgbequ.f sgbrfs.f sgbsv.f
sgbsvx.f sgbtf2.f sgbtrf.f sgbtrs.f sgebak.f sgebal.f sgebd2.f
sgebrd.f sgecon.f sgeequ.f sgees.f sgeesx.f sgeev.f sgeevx.f
DEPRECATED/sgegs.f DEPRECATED/sgegv.f sgehd2.f sgehrd.f sgelq2.f sgelqf.f
sgels.f sgelsd.f sgelss.f DEPRECATED/sgelsx.f sgelsy.f sgeql2.f sgeqlf.f
sgeqp3.f DEPRECATED/sgeqpf.f sgeqr2.f sgeqr2p.f sgeqrf.f sgeqrfp.f sgerfs.f
sgerq2.f sgerqf.f sgesc2.f sgesdd.f sgesvd.f sgesvdx.f sgesvx.f
sgetc2.f sgetri.f sgetrf2.f
sggbak.f sggbal.f sgghd3.f sgges.f sgges3.f sggesx.f sggev.f sggev3.f sggevx.f
sggglm.f sgghrd.f sgglse.f sggqrf.f
sggrqf.f DEPRECATED/sggsvd.f sggsvd3.f DEPRECATED/sggsvp.f sggsvp3.f sgtcon.f sgtrfs.f sgtsv.f
sgtsvx.f sgttrf.f sgttrs.f sgtts2.f shgeqz.f
shsein.f shseqr.f slabrd.f slacon.f slacn2.f
slaein.f slaexc.f slag2.f slags2.f slagtm.f slagv2.f slahqr.f
DEPRECATED/slahrd.f slahr2.f slaic1.f slaln2.f slals0.f slalsa.f slalsd.f
slangb.f slange.f slangt.f slanhs.f slansb.f slansp.f
slansy.f slantb.f slantp.f slantr.f slanv2.f
slapll.f slapmt.f
slaqgb.f slaqge.f slaqp2.f slaqps.f slaqsb.f slaqsp.f slaqsy.f
slaqr0.f slaqr1.f slaqr2.f slaqr3.f slaqr4.f slaqr5.f
slaqtr.f slar1v.f slar2v.f ilaslr.f ilaslc.f
slarf.f slarfb.f slarfg.f slarfgp.f slarft.f slarfx.f slargv.f
slarrv.f slartv.f
slarz.f slarzb.f slarzt.f slasy2.f slasyf.f slasyf_rook.f
slatbs.f slatdf.f slatps.f slatrd.f slatrs.f slatrz.f DEPRECATED/slatzm.f
sopgtr.f sopmtr.f sorg2l.f sorg2r.f
sorgbr.f sorghr.f sorgl2.f sorglq.f sorgql.f sorgqr.f sorgr2.f
sorgrq.f sorgtr.f sorm2l.f sorm2r.f sorm22.f
sormbr.f sormhr.f sorml2.f sormlq.f sormql.f sormqr.f sormr2.f
sormr3.f sormrq.f sormrz.f sormtr.f spbcon.f spbequ.f spbrfs.f
spbstf.f spbsv.f spbsvx.f
spbtf2.f spbtrf.f spbtrs.f spocon.f spoequ.f sporfs.f sposv.f
sposvx.f spstrf.f spstf2.f
sppcon.f sppequ.f
spprfs.f sppsv.f sppsvx.f spptrf.f spptri.f spptrs.f sptcon.f
spteqr.f sptrfs.f sptsv.f sptsvx.f spttrs.f sptts2.f srscl.f
ssbev.f ssbevd.f ssbevx.f ssbgst.f ssbgv.f ssbgvd.f ssbgvx.f
ssbtrd.f sspcon.f sspev.f sspevd.f sspevx.f sspgst.f
sspgv.f sspgvd.f sspgvx.f ssprfs.f sspsv.f sspsvx.f ssptrd.f
ssptrf.f ssptri.f ssptrs.f sstegr.f sstein.f sstev.f sstevd.f sstevr.f
sstevx.f
ssycon.f ssyev.f ssyevd.f ssyevr.f ssyevx.f ssygs2.f
ssygst.f ssygv.f ssygvd.f ssygvx.f ssyrfs.f ssysv.f ssysvx.f
ssytd2.f ssytf2.f ssytrd.f ssytrf.f ssytri.f ssytri2.f ssytri2x.f
ssyswapr.f ssytrs.f ssytrs2.f ssyconv.f
ssytf2_rook.f ssytrf_rook.f ssytrs_rook.f
ssytri_rook.f ssycon_rook.f ssysv_rook.f
stbcon.f
stbrfs.f stbtrs.f stgevc.f stgex2.f stgexc.f stgsen.f
stgsja.f stgsna.f stgsy2.f stgsyl.f stpcon.f stprfs.f stptri.f
stptrs.f
strcon.f strevc.f strevc3.f strexc.f strrfs.f strsen.f strsna.f strsyl.f
strtrs.f DEPRECATED/stzrqf.f stzrzf.f sstemr.f
slansf.f spftrf.f spftri.f spftrs.f ssfrk.f stfsm.f stftri.f stfttp.f
stfttr.f stpttf.f stpttr.f strttf.f strttp.f
sgejsv.f sgesvj.f sgsvj0.f sgsvj1.f
sgeequb.f ssyequb.f spoequb.f sgbequb.f
sbbcsd.f slapmr.f sorbdb.f sorbdb1.f sorbdb2.f sorbdb3.f sorbdb4.f
sorbdb5.f sorbdb6.f sorcsd.f sorcsd2by1.f
sgeqrt.f sgeqrt2.f sgeqrt3.f sgemqrt.f
stpqrt.f stpqrt2.f stpmqrt.f stprfb.f spotri.f
sgelq.f sgelqt.f sgelqt3.f sgemlq.f sgemlqt.f sgemqr.f sgeqr.f sgetsls.f
slamswlq.f slamtsqr.f slaswlq.f slatsqr.f stplqt.f stplqt2.f stpmlqt.f
ssysv_aa.f ssytrf_aa.f ssytrs_aa.f slasyf_aa.f ssytf2_rk.f slasyf_rk.f
ssytrf_rk.f ssytrs_3.f ssycon_3.f ssytri_3.f ssytri_3x.f ssysv_rk.f
ssb2st_kernels.f ssbev_2stage.f ssbevd_2stage.f ssbevx_2stage.f
ssyev_2stage.f ssyevd_2stage.f ssyevr_2stage.f ssyevx_2stage.f
ssygv_2stage.f ssytrd_2stage.f ssytrd_sb2st.F ssytrd_sy2sb.f slarfy.f
)
sbdsvdx.f sgbbrd.f sgbcon.f sgbequ.f sgbrfs.f sgbsv.f
sgbsvx.f sgbtf2.f sgbtrf.f sgbtrs.f sgebak.f sgebal.f sgebd2.f
sgebrd.f sgecon.f sgeequ.f sgees.f sgeesx.f sgeev.f sgeevx.f
sgehd2.f sgehrd.f sgelq2.f sgelqf.f
sgels.f sgelsd.f sgelss.f sgelsy.f sgeql2.f sgeqlf.f
sgeqp3.f sgeqr2.f sgeqr2p.f sgeqrf.f sgeqrfp.f sgerfs.f sgerq2.f sgerqf.f
sgesc2.f sgesdd.f sgesvd.f sgesvdx.f sgesvx.f sgetc2.f
sgetrf2.f sgetri.f
sggbak.f sggbal.f
sgges.f sgges3.f sggesx.f sggev.f sggev3.f sggevx.f
sggglm.f sgghrd.f sgghd3.f sgglse.f sggqrf.f
sggrqf.f sggsvd3.f sggsvp3.f sgtcon.f sgtrfs.f sgtsv.f
sgtsvx.f sgttrf.f sgttrs.f sgtts2.f shgeqz.f
shsein.f shseqr.f slabrd.f slacon.f slacn2.f
slaein.f slaexc.f slag2.f slags2.f slagtm.f slagv2.f slahqr.f
slahr2.f slaic1.f slaln2.f slals0.f slalsa.f slalsd.f
slangb.f slange.f slangt.f slanhs.f slansb.f slansp.f
slansy.f slantb.f slantp.f slantr.f slanv2.f
slapll.f slapmt.f
slaqgb.f slaqge.f slaqp2.f slaqps.f slaqsb.f slaqsp.f slaqsy.f
slaqr0.f slaqr1.f slaqr2.f slaqr3.f slaqr4.f slaqr5.f
slaqtr.f slar1v.f slar2v.f ilaslr.f ilaslc.f
slarf.f slarfb.f slarfg.f slarfgp.f slarft.f slarfx.f slarfy.f slargv.f
slarrv.f slartv.f
slarz.f slarzb.f slarzt.f slasy2.f
slasyf.f slasyf_rook.f slasyf_rk.f slasyf_aa.f
slatbs.f slatdf.f slatps.f slatrd.f slatrs.f slatrz.f
sopgtr.f sopmtr.f sorg2l.f sorg2r.f
sorgbr.f sorghr.f sorgl2.f sorglq.f sorgql.f sorgqr.f sorgr2.f
sorgrq.f sorgtr.f sorm2l.f sorm2r.f sorm22.f
sormbr.f sormhr.f sorml2.f sormlq.f sormql.f sormqr.f sormr2.f
sormr3.f sormrq.f sormrz.f sormtr.f spbcon.f spbequ.f spbrfs.f
spbstf.f spbsv.f spbsvx.f
spbtf2.f spbtrf.f spbtrs.f spocon.f spoequ.f sporfs.f sposv.f
sposvx.f spotrf2.f spotri.f spstrf.f spstf2.f
sppcon.f sppequ.f
spprfs.f sppsv.f sppsvx.f spptrf.f spptri.f spptrs.f sptcon.f
spteqr.f sptrfs.f sptsv.f sptsvx.f spttrs.f sptts2.f srscl.f
ssbev.f ssbevd.f ssbevx.f ssbgst.f ssbgv.f ssbgvd.f ssbgvx.f
ssbtrd.f sspcon.f sspev.f sspevd.f sspevx.f sspgst.f
sspgv.f sspgvd.f sspgvx.f ssprfs.f sspsv.f sspsvx.f ssptrd.f
ssptrf.f ssptri.f ssptrs.f sstegr.f sstein.f sstev.f sstevd.f sstevr.f
sstevx.f ssycon.f ssyev.f ssyevd.f ssyevr.f ssyevx.f ssygs2.f
ssygst.f ssygv.f ssygvd.f ssygvx.f ssyrfs.f ssysv.f ssysvx.f
ssytd2.f ssytf2.f ssytrd.f ssytrf.f ssytri.f ssytri2.f ssytri2x.f
ssyswapr.f ssytrs.f ssytrs2.f
ssyconv.f ssyconvf.f ssyconvf_rook.f
ssysv_aa.f ssysv_aa_2stage.f ssytrf_aa.f ssytrf_aa_2stage.f ssytrs_aa.f ssytrs_aa_2stage.f
ssytf2_rook.f ssytrf_rook.f ssytrs_rook.f
ssytri_rook.f ssycon_rook.f ssysv_rook.f
ssytf2_rk.f ssytrf_rk.f ssytrs_3.f
ssytri_3.f ssytri_3x.f ssycon_3.f ssysv_rk.f
ssysv_aa.f ssytrf_aa.f ssytrs_aa.f
stbcon.f
stbrfs.f stbtrs.f stgevc.f stgex2.f stgexc.f stgsen.f
stgsja.f stgsna.f stgsy2.f stgsyl.f stpcon.f stprfs.f stptri.f
stptrs.f
strcon.f strevc.f strevc3.f strexc.f strrfs.f strsen.f strsna.f strsyl.f
strtrs.f stzrzf.f sstemr.f
slansf.f spftrf.f spftri.f spftrs.f ssfrk.f stfsm.f stftri.f stfttp.f
stfttr.f stpttf.f stpttr.f strttf.f strttp.f
sgejsv.f sgesvj.f sgsvj0.f sgsvj1.f
sgeequb.f ssyequb.f spoequb.f sgbequb.f
sbbcsd.f slapmr.f sorbdb.f sorbdb1.f sorbdb2.f sorbdb3.f sorbdb4.f
sorbdb5.f sorbdb6.f sorcsd.f sorcsd2by1.f
sgeqrt.f sgeqrt2.f sgeqrt3.f sgemqrt.f
stpqrt.f stpqrt2.f stpmqrt.f stprfb.f
sgelqt.f sgelqt3.f sgemlqt.f
sgetsls.f sgeqr.f slatsqr.f slamtsqr.f sgemqr.f
sgelq.f slaswlq.f slamswlq.f sgemlq.f
stplqt.f stplqt2.f stpmlqt.f
ssytrd_2stage.f ssytrd_sy2sb.f ssytrd_sb2st.F ssb2st_kernels.f
ssyevd_2stage.f ssyev_2stage.f ssyevx_2stage.f ssyevr_2stage.f
ssbev_2stage.f ssbevx_2stage.f ssbevd_2stage.f ssygv_2stage.f)
set(DSLASRC spotrs.f spotrf2.f)
set(SXLASRC sgesvxx.f sgerfsx.f sla_gerfsx_extended.f sla_geamv.f
sla_gercond.f sla_gerpvgrw.f ssysvxx.f ssyrfsx.f
sla_syrfsx_extended.f sla_syamv.f sla_syrcond.f sla_syrpvgrw.f
sposvxx.f sporfsx.f sla_porfsx_extended.f sla_porcond.f
sla_porpvgrw.f sgbsvxx.f sgbrfsx.f sla_gbrfsx_extended.f
sla_gbamv.f sla_gbrcond.f sla_gbrpvgrw.f sla_lin_berr.f slarscl2.f
slascl2.f sla_wwaddw.f)
set(CLASRC
cbdsqr.f cgbbrd.f cgbcon.f cgbequ.f cgbrfs.f cgbsv.f cgbsvx.f
cgbtf2.f cgbtrf.f cgbtrs.f cgebak.f cgebal.f cgebd2.f cgebrd.f
cgecon.f cgeequ.f cgees.f cgeesx.f cgeev.f cgeevx.f
DEPRECATED/cgegs.f DEPRECATED/cgegv.f cgehd2.f cgehrd.f cgelq2.f cgelqf.f
cgels.f cgelsd.f cgelss.f DEPRECATED/cgelsx.f cgelsy.f cgeql2.f cgeqlf.f cgeqp3.f
DEPRECATED/cgeqpf.f cgeqr2.f cgeqr2p.f cgeqrf.f cgeqrfp.f cgerfs.f
cgerq2.f cgerqf.f cgesc2.f cgesdd.f cgesvd.f
cgesvx.f cgetc2.f cgetri.f
cggbak.f cggbal.f cgges.f cggesx.f cggev.f cggevx.f cggglm.f
cgghrd.f cgglse.f cggqrf.f cggrqf.f
DEPRECATED/cggsvd.f DEPRECATED/cggsvp.f
cgtcon.f cgtrfs.f cgtsv.f cgtsvx.f cgttrf.f cgttrs.f cgtts2.f chbev.f
chbevd.f chbevx.f chbgst.f chbgv.f chbgvd.f chbgvx.f chbtrd.f
checon.f cheev.f cheevd.f cheevr.f cheevx.f chegs2.f chegst.f
chegv.f chegvd.f chegvx.f cherfs.f chesv.f chesvx.f chetd2.f
chetf2.f chetrd.f
chetrf.f chetri.f chetri2.f chetri2x.f cheswapr.f
chetrs.f chetrs2.f
chetf2_rook.f chetrf_rook.f chetri_rook.f chetrs_rook.f checon_rook.f chesv_rook.f
chgeqz.f chpcon.f chpev.f chpevd.f
chpevx.f chpgst.f chpgv.f chpgvd.f chpgvx.f chprfs.f chpsv.f
chpsvx.f
chptrd.f chptrf.f chptri.f chptrs.f chsein.f chseqr.f clabrd.f
clacgv.f clacon.f clacn2.f clacp2.f clacpy.f clacrm.f clacrt.f cladiv.f
claed0.f claed7.f claed8.f
claein.f claesy.f claev2.f clags2.f clagtm.f
clahef.f clahef_rook.f clahqr.f
DEPRECATED/clahrd.f clahr2.f claic1.f clals0.f clalsa.f clalsd.f clangb.f clange.f clangt.f
clanhb.f clanhe.f
clanhp.f clanhs.f clanht.f clansb.f clansp.f clansy.f clantb.f
clantp.f clantr.f clapll.f clapmt.f clarcm.f claqgb.f claqge.f
claqhb.f claqhe.f claqhp.f claqp2.f claqps.f claqsb.f
claqr0.f claqr1.f claqr2.f claqr3.f claqr4.f claqr5.f
claqsp.f claqsy.f clar1v.f clar2v.f ilaclr.f ilaclc.f
clarf.f clarfb.f clarfg.f clarft.f clarfgp.f
clarfx.f clargv.f clarnv.f clarrv.f clartg.f clartv.f
clarz.f clarzb.f clarzt.f clascl.f claset.f clasr.f classq.f
clasyf.f clasyf_rook.f clatbs.f clatdf.f clatps.f clatrd.f clatrs.f clatrz.f
DEPRECATED/clatzm.f cpbcon.f cpbequ.f cpbrfs.f cpbstf.f cpbsv.f
cpbsvx.f cpbtf2.f cpbtrf.f cpbtrs.f cpocon.f cpoequ.f cporfs.f
cposv.f cposvx.f cpstrf.f cpstf2.f
cppcon.f cppequ.f cpprfs.f cppsv.f cppsvx.f cpptrf.f cpptri.f cpptrs.f
cptcon.f cpteqr.f cptrfs.f cptsv.f cptsvx.f cpttrf.f cpttrs.f cptts2.f
crot.f cspcon.f csprfs.f cspsv.f
cspsvx.f csptrf.f csptri.f csptrs.f csrscl.f cstedc.f
cstegr.f cstein.f csteqr.f
csycon.f
csyrfs.f csysv.f csysvx.f csytf2.f csytrf.f csytri.f csytri2.f csytri2x.f
csyswapr.f csytrs.f csytrs2.f csyconv.f
csytf2_rook.f csytrf_rook.f csytrs_rook.f
csytri_rook.f csycon_rook.f csysv_rook.f
ctbcon.f ctbrfs.f ctbtrs.f ctgevc.f ctgex2.f
ctgexc.f ctgsen.f ctgsja.f ctgsna.f ctgsy2.f ctgsyl.f ctpcon.f
ctprfs.f ctptri.f
ctptrs.f ctrcon.f ctrevc.f ctrevc3.f ctrexc.f ctrrfs.f ctrsen.f ctrsna.f
ctrsyl.f ctrtrs.f DEPRECATED/ctzrqf.f ctzrzf.f cung2l.f cung2r.f
cungbr.f cunghr.f cungl2.f cunglq.f cungql.f cungqr.f cungr2.f
cungrq.f cungtr.f cunm2l.f cunm2r.f cunmbr.f cunmhr.f cunml2.f
cunmlq.f cunmql.f cunmqr.f cunmr2.f cunmr3.f cunmrq.f cunmrz.f
cunmtr.f cupgtr.f cupmtr.f icmax1.f scsum1.f cstemr.f
chfrk.f ctfttp.f clanhf.f cpftrf.f cpftri.f cpftrs.f ctfsm.f ctftri.f
ctfttr.f ctpttf.f ctpttr.f ctrttf.f ctrttp.f
cgeequb.f cgbequb.f csyequb.f cpoequb.f cheequb.f
cbbcsd.f clapmr.f cunbdb.f cunbdb1.f cunbdb2.f cunbdb3.f cunbdb4.f
cunbdb5.f cunbdb6.f cuncsd.f cuncsd2by1.f
cgeqrt.f cgeqrt2.f cgeqrt3.f cgemqrt.f
ctpqrt.f ctpqrt2.f ctpmqrt.f ctprfb.f cpotri.f
cgelq.f cgelqt.f cgelqt3.f cgemlq.f cgemlqt.f cgemqr.f cgeqr.f cgetsls.f
clamswlq.f clamtsqr.f claswlq.f clatsqr.f ctplqt.f ctplqt2.f ctpmlqt.f
chesv_aa.f chetrf_aa.f chetrs_aa.f clahef_aa.f csytf2_rk.f clasyf_rk.f
csytrf_rk.f csytrs_3.f csycon_3.f csytri_3.f csytri_3x.f csysv_rk.f
chetf2_rk.f clahef_rk.f chetrf_rk.f chetrs_3.f checon_3.f chetri_3.f
chetri_3x.f chesv_rk.f chb2st_kernels.f chbev_2stage.f chbevd_2stage.f
chbevx_2stage.f cheev_2stage.f cheevd_2stage.f cheevr_2stage.f cheevx_2stage.f
chegv_2stage.f chetrd_2stage.f chetrd_hb2st.F chetrd_he2hb.f clarfy.f
)
cbdsqr.f cgbbrd.f cgbcon.f cgbequ.f cgbrfs.f cgbsv.f cgbsvx.f
cgbtf2.f cgbtrf.f cgbtrs.f cgebak.f cgebal.f cgebd2.f cgebrd.f
cgecon.f cgeequ.f cgees.f cgeesx.f cgeev.f cgeevx.f
cgehd2.f cgehrd.f cgelq2.f cgelqf.f
cgels.f cgelsd.f cgelss.f cgelsy.f cgeql2.f cgeqlf.f cgeqp3.f
cgeqr2.f cgeqr2p.f cgeqrf.f cgeqrfp.f cgerfs.f cgerq2.f cgerqf.f
cgesc2.f cgesdd.f cgesvd.f cgesvdx.f
cgesvj.f cgejsv.f cgsvj0.f cgsvj1.f
cgesvx.f cgetc2.f cgetrf2.f
cgetri.f
cggbak.f cggbal.f
cgges.f cgges3.f cggesx.f cggev.f cggev3.f cggevx.f
cggglm.f cgghrd.f cgghd3.f cgglse.f cggqrf.f cggrqf.f
cggsvd3.f cggsvp3.f
cgtcon.f cgtrfs.f cgtsv.f cgtsvx.f cgttrf.f cgttrs.f cgtts2.f chbev.f
chbevd.f chbevx.f chbgst.f chbgv.f chbgvd.f chbgvx.f chbtrd.f
checon.f cheev.f cheevd.f cheevr.f cheevx.f chegs2.f chegst.f
chegv.f chegvd.f chegvx.f cherfs.f chesv.f chesvx.f chetd2.f
chetf2.f chetrd.f
chetrf.f chetri.f chetri2.f chetri2x.f cheswapr.f
chetrs.f chetrs2.f
chetf2_rook.f chetrf_rook.f chetri_rook.f
chetrs_rook.f checon_rook.f chesv_rook.f
chetf2_rk.f chetrf_rk.f chetri_3.f chetri_3x.f
chetrs_3.f checon_3.f chesv_rk.f
chesv_aa.f chesv_aa_2stage.f chetrf_aa.f chetrf_aa_2stage.f chetrs_aa.f chetrs_aa_2stage.f
chgeqz.f chpcon.f chpev.f chpevd.f
chpevx.f chpgst.f chpgv.f chpgvd.f chpgvx.f chprfs.f chpsv.f
chpsvx.f
chptrd.f chptrf.f chptri.f chptrs.f chsein.f chseqr.f clabrd.f
clacgv.f clacon.f clacn2.f clacp2.f clacpy.f clacrm.f clacrt.f cladiv.f
claed0.f claed7.f claed8.f
claein.f claesy.f claev2.f clags2.f clagtm.f
clahef.f clahef_rook.f clahef_rk.f clahef_aa.f clahqr.f
clahr2.f claic1.f clals0.f clalsa.f clalsd.f clangb.f clange.f clangt.f
clanhb.f clanhe.f
clanhp.f clanhs.f clanht.f clansb.f clansp.f clansy.f clantb.f
clantp.f clantr.f clapll.f clapmt.f clarcm.f claqgb.f claqge.f
claqhb.f claqhe.f claqhp.f claqp2.f claqps.f claqsb.f
claqr0.f claqr1.f claqr2.f claqr3.f claqr4.f claqr5.f
claqsp.f claqsy.f clar1v.f clar2v.f ilaclr.f ilaclc.f
clarf.f clarfb.f clarfg.f clarfgp.f clarft.f
clarfx.f clarfy.f clargv.f clarnv.f clarrv.f clartg.f clartv.f
clarz.f clarzb.f clarzt.f clascl.f claset.f clasr.f classq.f
clasyf.f clasyf_rook.f clasyf_rk.f clasyf_aa.f
clatbs.f clatdf.f clatps.f clatrd.f clatrs.f clatrz.f
cpbcon.f cpbequ.f cpbrfs.f cpbstf.f cpbsv.f
cpbsvx.f cpbtf2.f cpbtrf.f cpbtrs.f cpocon.f cpoequ.f cporfs.f
cposv.f cposvx.f cpotrf2.f cpotri.f cpstrf.f cpstf2.f
cppcon.f cppequ.f cpprfs.f cppsv.f cppsvx.f cpptrf.f cpptri.f cpptrs.f
cptcon.f cpteqr.f cptrfs.f cptsv.f cptsvx.f cpttrf.f cpttrs.f cptts2.f
crot.f cspcon.f csprfs.f cspsv.f
cspsvx.f csptrf.f csptri.f csptrs.f csrscl.f cstedc.f
cstegr.f cstein.f csteqr.f csycon.f
csyrfs.f csysv.f csysvx.f csytf2.f csytrf.f csytri.f
csytri2.f csytri2x.f csyswapr.f
csytrs.f csytrs2.f
csyconv.f csyconvf.f csyconvf_rook.f
csytf2_rook.f csytrf_rook.f csytrs_rook.f
csytri_rook.f csycon_rook.f csysv_rook.f
csytf2_rk.f csytrf_rk.f csytrf_aa.f csytrf_aa_2stage.f csytrs_3.f csytrs_aa.f csytrs_aa_2stage.f
csytri_3.f csytri_3x.f csycon_3.f csysv_rk.f csysv_aa.f csysv_aa_2stage.f
ctbcon.f ctbrfs.f ctbtrs.f ctgevc.f ctgex2.f
ctgexc.f ctgsen.f ctgsja.f ctgsna.f ctgsy2.f ctgsyl.f ctpcon.f
ctprfs.f ctptri.f
ctptrs.f ctrcon.f ctrevc.f ctrevc3.f ctrexc.f ctrrfs.f ctrsen.f ctrsna.f
ctrsyl.f ctrtrs.f ctzrzf.f cung2l.f cung2r.f
cungbr.f cunghr.f cungl2.f cunglq.f cungql.f cungqr.f cungr2.f
cungrq.f cungtr.f cunm2l.f cunm2r.f cunmbr.f cunmhr.f cunml2.f cunm22.f
cunmlq.f cunmql.f cunmqr.f cunmr2.f cunmr3.f cunmrq.f cunmrz.f
cunmtr.f cupgtr.f cupmtr.f icmax1.f scsum1.f cstemr.f
chfrk.f ctfttp.f clanhf.f cpftrf.f cpftri.f cpftrs.f ctfsm.f ctftri.f
ctfttr.f ctpttf.f ctpttr.f ctrttf.f ctrttp.f
cgeequb.f cgbequb.f csyequb.f cpoequb.f cheequb.f
cbbcsd.f clapmr.f cunbdb.f cunbdb1.f cunbdb2.f cunbdb3.f cunbdb4.f
cunbdb5.f cunbdb6.f cuncsd.f cuncsd2by1.f
cgeqrt.f cgeqrt2.f cgeqrt3.f cgemqrt.f
ctpqrt.f ctpqrt2.f ctpmqrt.f ctprfb.f
cgelqt.f cgelqt3.f cgemlqt.f
cgetsls.f cgeqr.f clatsqr.f clamtsqr.f cgemqr.f
cgelq.f claswlq.f clamswlq.f cgemlq.f
ctplqt.f ctplqt2.f ctpmlqt.f
chetrd_2stage.f chetrd_he2hb.f chetrd_hb2st.F chb2st_kernels.f
cheevd_2stage.f cheev_2stage.f cheevx_2stage.f cheevr_2stage.f
chbev_2stage.f chbevx_2stage.f chbevd_2stage.f chegv_2stage.f)
set(CXLASRC cgesvxx.f cgerfsx.f cla_gerfsx_extended.f cla_geamv.f
cla_gercond_c.f cla_gercond_x.f cla_gerpvgrw.f
csysvxx.f csyrfsx.f cla_syrfsx_extended.f cla_syamv.f
cla_syrcond_c.f cla_syrcond_x.f cla_syrpvgrw.f
cposvxx.f cporfsx.f cla_porfsx_extended.f
cla_porcond_c.f cla_porcond_x.f cla_porpvgrw.f
cgbsvxx.f cgbrfsx.f cla_gbrfsx_extended.f cla_gbamv.f
cla_gbrcond_c.f cla_gbrcond_x.f cla_gbrpvgrw.f
chesvxx.f cherfsx.f cla_herfsx_extended.f cla_heamv.f
cla_hercond_c.f cla_hercond_x.f cla_herpvgrw.f
cla_lin_berr.f clarscl2.f clascl2.f cla_wwaddw.f)
set(DLASRC
dbdsvdx.f dgbbrd.f dgbcon.f dgbequ.f dgbrfs.f dgbsv.f
dgbsvx.f dgbtf2.f dgbtrf.f dgbtrs.f dgebak.f dgebal.f dgebd2.f
dgebrd.f dgecon.f dgeequ.f dgees.f dgeesx.f dgeev.f dgeevx.f
dgehd2.f dgehrd.f dgelq2.f dgelqf.f
dgels.f dgelsd.f dgelss.f dgelsy.f dgeql2.f dgeqlf.f
dgeqp3.f dgeqr2.f dgeqr2p.f dgeqrf.f dgeqrfp.f dgerfs.f dgerq2.f dgerqf.f
dgesc2.f dgesdd.f dgesvd.f dgesvdx.f dgesvx.f dgetc2.f
dgetrf2.f dgetri.f
dggbak.f dggbal.f
dgges.f dgges3.f dggesx.f dggev.f dggev3.f dggevx.f
dggglm.f dgghrd.f dgghd3.f dgglse.f dggqrf.f
dggrqf.f dggsvd3.f dggsvp3.f dgtcon.f dgtrfs.f dgtsv.f
dgtsvx.f dgttrf.f dgttrs.f dgtts2.f dhgeqz.f
dhsein.f dhseqr.f dlabrd.f dlacon.f dlacn2.f
dlaein.f dlaexc.f dlag2.f dlags2.f dlagtm.f dlagv2.f dlahqr.f
dlahr2.f dlaic1.f dlaln2.f dlals0.f dlalsa.f dlalsd.f
dlangb.f dlange.f dlangt.f dlanhs.f dlansb.f dlansp.f
dlansy.f dlantb.f dlantp.f dlantr.f dlanv2.f
dlapll.f dlapmt.f
dlaqgb.f dlaqge.f dlaqp2.f dlaqps.f dlaqsb.f dlaqsp.f dlaqsy.f
dlaqr0.f dlaqr1.f dlaqr2.f dlaqr3.f dlaqr4.f dlaqr5.f
dlaqtr.f dlar1v.f dlar2v.f iladlr.f iladlc.f
dlarf.f dlarfb.f dlarfg.f dlarfgp.f dlarft.f dlarfx.f dlarfy.f
dlargv.f dlarrv.f dlartv.f
dlarz.f dlarzb.f dlarzt.f dlasy2.f
dlasyf.f dlasyf_rook.f dlasyf_rk.f dlasyf_aa.f
dlatbs.f dlatdf.f dlatps.f dlatrd.f dlatrs.f dlatrz.f
dopgtr.f dopmtr.f dorg2l.f dorg2r.f
dorgbr.f dorghr.f dorgl2.f dorglq.f dorgql.f dorgqr.f dorgr2.f
dorgrq.f dorgtr.f dorm2l.f dorm2r.f dorm22.f
dormbr.f dormhr.f dorml2.f dormlq.f dormql.f dormqr.f dormr2.f
dormr3.f dormrq.f dormrz.f dormtr.f dpbcon.f dpbequ.f dpbrfs.f
dpbstf.f dpbsv.f dpbsvx.f
dpbtf2.f dpbtrf.f dpbtrs.f dpocon.f dpoequ.f dporfs.f dposv.f
dposvx.f dpotrf2.f dpotri.f dpotrs.f dpstrf.f dpstf2.f
dppcon.f dppequ.f
dpprfs.f dppsv.f dppsvx.f dpptrf.f dpptri.f dpptrs.f dptcon.f
dpteqr.f dptrfs.f dptsv.f dptsvx.f dpttrs.f dptts2.f drscl.f
dsbev.f dsbevd.f dsbevx.f dsbgst.f dsbgv.f dsbgvd.f dsbgvx.f
dsbtrd.f dspcon.f dspev.f dspevd.f dspevx.f dspgst.f
dspgv.f dspgvd.f dspgvx.f dsprfs.f dspsv.f dspsvx.f dsptrd.f
dsptrf.f dsptri.f dsptrs.f dstegr.f dstein.f dstev.f dstevd.f dstevr.f
dstevx.f dsycon.f dsyev.f dsyevd.f dsyevr.f
dsyevx.f dsygs2.f dsygst.f dsygv.f dsygvd.f dsygvx.f dsyrfs.f
dsysv.f dsysvx.f
dsytd2.f dsytf2.f dsytrd.f dsytrf.f dsytri.f dsytrs.f dsytrs2.f
dsytri2.f dsytri2x.f dsyswapr.f
dsyconv.f dsyconvf.f dsyconvf_rook.f
dsytf2_rook.f dsytrf_rook.f dsytrs_rook.f
dsytri_rook.f dsycon_rook.f dsysv_rook.f
dsytf2_rk.f dsytrf_rk.f dsytrs_3.f
dsytri_3.f dsytri_3x.f dsycon_3.f dsysv_rk.f
dsysv_aa.f dsysv_aa_2stage.f dsytrf_aa.f dsytrf_aa_2stage.f dsytrs_aa.f dsytrs_aa_2stage.f
dtbcon.f
dtbrfs.f dtbtrs.f dtgevc.f dtgex2.f dtgexc.f dtgsen.f
dtgsja.f dtgsna.f dtgsy2.f dtgsyl.f dtpcon.f dtprfs.f dtptri.f
dtptrs.f
dtrcon.f dtrevc.f dtrevc3.f dtrexc.f dtrrfs.f dtrsen.f dtrsna.f dtrsyl.f
dtrtrs.f dtzrzf.f dstemr.f
dsgesv.f dsposv.f dlag2s.f slag2d.f dlat2s.f
dlansf.f dpftrf.f dpftri.f dpftrs.f dsfrk.f dtfsm.f dtftri.f dtfttp.f
dtfttr.f dtpttf.f dtpttr.f dtrttf.f dtrttp.f
dgejsv.f dgesvj.f dgsvj0.f dgsvj1.f
dgeequb.f dsyequb.f dpoequb.f dgbequb.f
dbbcsd.f dlapmr.f dorbdb.f dorbdb1.f dorbdb2.f dorbdb3.f dorbdb4.f
dorbdb5.f dorbdb6.f dorcsd.f dorcsd2by1.f
dgeqrt.f dgeqrt2.f dgeqrt3.f dgemqrt.f
dtpqrt.f dtpqrt2.f dtpmqrt.f dtprfb.f
dgelqt.f dgelqt3.f dgemlqt.f
dgetsls.f dgeqr.f dlatsqr.f dlamtsqr.f dgemqr.f
dgelq.f dlaswlq.f dlamswlq.f dgemlq.f
dtplqt.f dtplqt2.f dtpmlqt.f
dsytrd_2stage.f dsytrd_sy2sb.f dsytrd_sb2st.F dsb2st_kernels.f
dsyevd_2stage.f dsyev_2stage.f dsyevx_2stage.f dsyevr_2stage.f
dsbev_2stage.f dsbevx_2stage.f dsbevd_2stage.f dsygv_2stage.f)
set(DXLASRC dgesvxx.f dgerfsx.f dla_gerfsx_extended.f dla_geamv.f
dla_gercond.f dla_gerpvgrw.f dsysvxx.f dsyrfsx.f
dla_syrfsx_extended.f dla_syamv.f dla_syrcond.f dla_syrpvgrw.f
dposvxx.f dporfsx.f dla_porfsx_extended.f dla_porcond.f
dla_porpvgrw.f dgbsvxx.f dgbrfsx.f dla_gbrfsx_extended.f
dla_gbamv.f dla_gbrcond.f dla_gbrpvgrw.f dla_lin_berr.f dlarscl2.f
dlascl2.f dla_wwaddw.f)
set(ZLASRC
zbdsqr.f zgbbrd.f zgbcon.f zgbequ.f zgbrfs.f zgbsv.f zgbsvx.f
zgbtf2.f zgbtrf.f zgbtrs.f zgebak.f zgebal.f zgebd2.f zgebrd.f
zgecon.f zgeequ.f zgees.f zgeesx.f zgeev.f zgeevx.f
zgehd2.f zgehrd.f zgelq2.f zgelqf.f
zgels.f zgelsd.f zgelss.f zgelsy.f zgeql2.f zgeqlf.f zgeqp3.f
zgeqr2.f zgeqr2p.f zgeqrf.f zgeqrfp.f zgerfs.f zgerq2.f zgerqf.f
zgesc2.f zgesdd.f zgesvd.f zgesvdx.f zgesvx.f
zgesvj.f zgejsv.f zgsvj0.f zgsvj1.f
zgetc2.f zgetrf2.f
zgetri.f
zggbak.f zggbal.f
zgges.f zgges3.f zggesx.f zggev.f zggev3.f zggevx.f
zggglm.f zgghrd.f zgghd3.f zgglse.f zggqrf.f zggrqf.f
zggsvd3.f zggsvp3.f
zgtcon.f zgtrfs.f zgtsv.f zgtsvx.f zgttrf.f zgttrs.f zgtts2.f zhbev.f
zhbevd.f zhbevx.f zhbgst.f zhbgv.f zhbgvd.f zhbgvx.f zhbtrd.f
zhecon.f zheev.f zheevd.f zheevr.f zheevx.f zhegs2.f zhegst.f
zhegv.f zhegvd.f zhegvx.f zherfs.f zhesv.f zhesvx.f zhetd2.f
zhetf2.f zhetrd.f
zhetrf.f zhetri.f zhetri2.f zhetri2x.f zheswapr.f
zhetrs.f zhetrs2.f
zhetf2_rook.f zhetrf_rook.f zhetri_rook.f
zhetrs_rook.f zhecon_rook.f zhesv_rook.f
zhetf2_rk.f zhetrf_rk.f zhetri_3.f zhetri_3x.f
zhetrs_3.f zhecon_3.f zhesv_rk.f
zhesv_aa.f zhesv_aa_2stage.f zhetrf_aa.f zhetrf_aa_2stage.f zhetrs_aa.f zhetrs_aa_2stage.f
zhgeqz.f zhpcon.f zhpev.f zhpevd.f
zhpevx.f zhpgst.f zhpgv.f zhpgvd.f zhpgvx.f zhprfs.f zhpsv.f
zhpsvx.f
zhptrd.f zhptrf.f zhptri.f zhptrs.f zhsein.f zhseqr.f zlabrd.f
zlacgv.f zlacon.f zlacn2.f zlacp2.f zlacpy.f zlacrm.f zlacrt.f zladiv.f
zlaed0.f zlaed7.f zlaed8.f
zlaein.f zlaesy.f zlaev2.f zlags2.f zlagtm.f
zlahef.f zlahef_rook.f zlahef_rk.f zlahef_aa.f zlahqr.f
zlahr2.f zlaic1.f zlals0.f zlalsa.f zlalsd.f zlangb.f zlange.f
zlangt.f zlanhb.f
zlanhe.f
zlanhp.f zlanhs.f zlanht.f zlansb.f zlansp.f zlansy.f zlantb.f
zlantp.f zlantr.f zlapll.f zlapmt.f zlaqgb.f zlaqge.f
zlaqhb.f zlaqhe.f zlaqhp.f zlaqp2.f zlaqps.f zlaqsb.f
zlaqr0.f zlaqr1.f zlaqr2.f zlaqr3.f zlaqr4.f zlaqr5.f
zlaqsp.f zlaqsy.f zlar1v.f zlar2v.f ilazlr.f ilazlc.f
zlarcm.f zlarf.f zlarfb.f
zlarfg.f zlarfgp.f zlarft.f
zlarfx.f zlarfy.f zlargv.f zlarnv.f zlarrv.f zlartg.f zlartv.f
zlarz.f zlarzb.f zlarzt.f zlascl.f zlaset.f zlasr.f
zlassq.f zlasyf.f zlasyf_rook.f zlasyf_rk.f zlasyf_aa.f
zlatbs.f zlatdf.f zlatps.f zlatrd.f zlatrs.f zlatrz.f
zpbcon.f zpbequ.f zpbrfs.f zpbstf.f zpbsv.f
zpbsvx.f zpbtf2.f zpbtrf.f zpbtrs.f zpocon.f zpoequ.f zporfs.f
zposv.f zposvx.f zpotrf2.f zpotri.f zpotrs.f zpstrf.f zpstf2.f
zppcon.f zppequ.f zpprfs.f zppsv.f zppsvx.f zpptrf.f zpptri.f zpptrs.f
zptcon.f zpteqr.f zptrfs.f zptsv.f zptsvx.f zpttrf.f zpttrs.f zptts2.f
zrot.f zspcon.f zsprfs.f zspsv.f
zspsvx.f zsptrf.f zsptri.f zsptrs.f zdrscl.f zstedc.f
zstegr.f zstein.f zsteqr.f zsycon.f
zsyrfs.f zsysv.f zsysvx.f zsytf2.f zsytrf.f zsytri.f
zsytri2.f zsytri2x.f zsyswapr.f
zsytrs.f zsytrs2.f
zsyconv.f zsyconvf.f zsyconvf_rook.f
zsytf2_rook.f zsytrf_rook.f zsytrs_rook.f zsytrs_aa.f zsytrs_aa_2stage.f
zsytri_rook.f zsycon_rook.f zsysv_rook.f
zsytf2_rk.f zsytrf_rk.f zsytrf_aa.f zsytrf_aa_2stage.f zsytrs_3.f
zsytri_3.f zsytri_3x.f zsycon_3.f zsysv_rk.f zsysv_aa.f zsysv_aa_2stage.f
ztbcon.f ztbrfs.f ztbtrs.f ztgevc.f ztgex2.f
ztgexc.f ztgsen.f ztgsja.f ztgsna.f ztgsy2.f ztgsyl.f ztpcon.f
ztprfs.f ztptri.f
ztptrs.f ztrcon.f ztrevc.f ztrevc3.f ztrexc.f ztrrfs.f ztrsen.f ztrsna.f
ztrsyl.f ztrtrs.f ztzrzf.f zung2l.f
zung2r.f zungbr.f zunghr.f zungl2.f zunglq.f zungql.f zungqr.f zungr2.f
zungrq.f zungtr.f zunm2l.f zunm2r.f zunmbr.f zunmhr.f zunml2.f zunm22.f
zunmlq.f zunmql.f zunmqr.f zunmr2.f zunmr3.f zunmrq.f zunmrz.f
zunmtr.f zupgtr.f
zupmtr.f izmax1.f dzsum1.f zstemr.f
zcgesv.f zcposv.f zlag2c.f clag2z.f zlat2c.f
zhfrk.f ztfttp.f zlanhf.f zpftrf.f zpftri.f zpftrs.f ztfsm.f ztftri.f
ztfttr.f ztpttf.f ztpttr.f ztrttf.f ztrttp.f
zgeequb.f zgbequb.f zsyequb.f zpoequb.f zheequb.f
zbbcsd.f zlapmr.f zunbdb.f zunbdb1.f zunbdb2.f zunbdb3.f zunbdb4.f
zunbdb5.f zunbdb6.f zuncsd.f zuncsd2by1.f
zgeqrt.f zgeqrt2.f zgeqrt3.f zgemqrt.f
ztpqrt.f ztpqrt2.f ztpmqrt.f ztprfb.f
ztplqt.f ztplqt2.f ztpmlqt.f
zgelqt.f zgelqt3.f zgemlqt.f
zgetsls.f zgeqr.f zlatsqr.f zlamtsqr.f zgemqr.f
zgelq.f zlaswlq.f zlamswlq.f zgemlq.f
zhetrd_2stage.f zhetrd_he2hb.f zhetrd_hb2st.F zhb2st_kernels.f
zheevd_2stage.f zheev_2stage.f zheevx_2stage.f zheevr_2stage.f
zhbev_2stage.f zhbevx_2stage.f zhbevd_2stage.f zhegv_2stage.f)
set(ZXLASRC zgesvxx.f zgerfsx.f zla_gerfsx_extended.f zla_geamv.f
zla_gercond_c.f zla_gercond_x.f zla_gerpvgrw.f zsysvxx.f zsyrfsx.f
zla_syrfsx_extended.f zla_syamv.f zla_syrcond_c.f zla_syrcond_x.f
zla_syrpvgrw.f zposvxx.f zporfsx.f zla_porfsx_extended.f
zla_porcond_c.f zla_porcond_x.f zla_porpvgrw.f zgbsvxx.f zgbrfsx.f
zla_gbrfsx_extended.f zla_gbamv.f zla_gbrcond_c.f zla_gbrcond_x.f
zla_gbrpvgrw.f zhesvxx.f zherfsx.f zla_herfsx_extended.f
zla_heamv.f zla_hercond_c.f zla_hercond_x.f zla_herpvgrw.f
zla_lin_berr.f zlarscl2.f zlascl2.f zla_wwaddw.f)
if(USE_XBLAS)
set(ALLXOBJ ${SXLASRC} ${DXLASRC} ${CXLASRC} ${ZXLASRC})
endif()
list(APPEND SLASRC DEPRECATED/sgegs.f DEPRECATED/sgegv.f
DEPRECATED/sgeqpf.f DEPRECATED/sgelsx.f DEPRECATED/sggsvd.f
DEPRECATED/sggsvp.f DEPRECATED/slahrd.f DEPRECATED/slatzm.f DEPRECATED/stzrqf.f)
list(APPEND DLASRC DEPRECATED/dgegs.f DEPRECATED/dgegv.f
DEPRECATED/dgeqpf.f DEPRECATED/dgelsx.f DEPRECATED/dggsvd.f
DEPRECATED/dggsvp.f DEPRECATED/dlahrd.f DEPRECATED/dlatzm.f DEPRECATED/dtzrqf.f)
list(APPEND CLASRC DEPRECATED/cgegs.f DEPRECATED/cgegv.f
DEPRECATED/cgeqpf.f DEPRECATED/cgelsx.f DEPRECATED/cggsvd.f
DEPRECATED/cggsvp.f DEPRECATED/clahrd.f DEPRECATED/clatzm.f DEPRECATED/ctzrqf.f)
list(APPEND ZLASRC DEPRECATED/zgegs.f DEPRECATED/zgegv.f
DEPRECATED/zgeqpf.f DEPRECATED/zgelsx.f DEPRECATED/zggsvd.f
DEPRECATED/zggsvp.f DEPRECATED/zlahrd.f DEPRECATED/zlatzm.f DEPRECATED/ztzrqf.f)
message(STATUS "Building deprecated routines")
set(DSLASRC spotrs.f)
set(ZCLASRC cpotrs.f)
set(DLASRC
dgbbrd.f dgbcon.f dgbequ.f dgbrfs.f dgbsv.f
dgbsvx.f dgbtf2.f dgbtrf.f dgbtrs.f dgebak.f dgebal.f dgebd2.f
dgebrd.f dgecon.f dgeequ.f dgees.f dgeesx.f dgeev.f dgeevx.f
DEPRECATED/dgegs.f DEPRECATED/dgegv.f dgehd2.f dgehrd.f dgelq2.f dgelqf.f
dgels.f dgelsd.f dgelss.f DEPRECATED/dgelsx.f dgelsy.f dgeql2.f dgeqlf.f
dgeqp3.f DEPRECATED/dgeqpf.f dgeqr2.f dgeqr2p.f dgeqrf.f dgeqrfp.f dgerfs.f
dgerq2.f dgerqf.f dgesc2.f dgesdd.f dgesvd.f dgesvdx.f dgesvx.f
dgetc2.f dgetri.f dgetrf2.f
dggbak.f dggbal.f dgges.f dgges3.f dggesx.f dggev.f dggev3.f dggevx.f
dggglm.f dgghd3.f dgghrd.f dgglse.f dggqrf.f
dggrqf.f dggsvd3.f dggsvp3.f DEPRECATED/dggsvd.f DEPRECATED/dggsvp.f dgtcon.f dgtrfs.f dgtsv.f
dgtsvx.f dgttrf.f dgttrs.f dgtts2.f dhgeqz.f
dhsein.f dhseqr.f dlabrd.f dlacon.f dlacn2.f
dlaein.f dlaexc.f dlag2.f dlags2.f dlagtm.f dlagv2.f dlahqr.f
DEPRECATED/dlahrd.f dlahr2.f dlaic1.f dlaln2.f dlals0.f dlalsa.f dlalsd.f
dlangb.f dlange.f dlangt.f dlanhs.f dlansb.f dlansp.f
dlansy.f dlantb.f dlantp.f dlantr.f dlanv2.f
dlapll.f dlapmt.f
dlaqgb.f dlaqge.f dlaqp2.f dlaqps.f dlaqsb.f dlaqsp.f dlaqsy.f
dlaqr0.f dlaqr1.f dlaqr2.f dlaqr3.f dlaqr4.f dlaqr5.f
dlaqtr.f dlar1v.f dlar2v.f iladlr.f iladlc.f
dlarf.f dlarfb.f dlarfg.f dlarfgp.f dlarft.f dlarfx.f
dlargv.f dlarrv.f dlartv.f
dlarz.f dlarzb.f dlarzt.f dlasy2.f dlasyf.f dlasyf_rook.f
dlatbs.f dlatdf.f dlatps.f dlatrd.f dlatrs.f dlatrz.f DEPRECATED/dlatzm.f
dopgtr.f dopmtr.f dorg2l.f dorg2r.f
dorgbr.f dorghr.f dorgl2.f dorglq.f dorgql.f dorgqr.f dorgr2.f
dorgrq.f dorgtr.f dorm2l.f dorm2r.f dorm22.f
dormbr.f dormhr.f dorml2.f dormlq.f dormql.f dormqr.f dormr2.f
dormr3.f dormrq.f dormrz.f dormtr.f dpbcon.f dpbequ.f dpbrfs.f
dpbstf.f dpbsv.f dpbsvx.f
dpbtf2.f dpbtrf.f dpbtrs.f dpocon.f dpoequ.f dporfs.f dposv.f
dposvx.f dpotrf2.f dpotrs.f dpstrf.f dpstf2.f
dppcon.f dppequ.f
dpprfs.f dppsv.f dppsvx.f dpptrf.f dpptri.f dpptrs.f dptcon.f
dpteqr.f dptrfs.f dptsv.f dptsvx.f dpttrs.f dptts2.f drscl.f
dsbev.f dsbevd.f dsbevx.f dsbgst.f dsbgv.f dsbgvd.f dsbgvx.f
dsbtrd.f dspcon.f dspev.f dspevd.f dspevx.f dspgst.f
dspgv.f dspgvd.f dspgvx.f dsprfs.f dspsv.f dspsvx.f dsptrd.f
dsptrf.f dsptri.f dsptrs.f dstegr.f dstein.f dstev.f dstevd.f dstevr.f
dstevx.f
dsycon.f dsyev.f dsyevd.f dsyevr.f
dsyevx.f dsygs2.f dsygst.f dsygv.f dsygvd.f dsygvx.f dsyrfs.f
dsysv.f dsysvx.f
dsytd2.f dsytf2.f dsytrd.f dsytrf.f dsytri.f dsytri2.f dsytri2x.f
dsyswapr.f dsytrs.f dsytrs2.f dsyconv.f
dsytf2_rook.f dsytrf_rook.f dsytrs_rook.f
dsytri_rook.f dsycon_rook.f dsysv_rook.f
dtbcon.f dtbrfs.f dtbtrs.f dtgevc.f dtgex2.f dtgexc.f dtgsen.f
dtgsja.f dtgsna.f dtgsy2.f dtgsyl.f dtpcon.f dtprfs.f dtptri.f
dtptrs.f
dtrcon.f dtrevc.f dtrevc3.f dtrexc.f dtrrfs.f dtrsen.f dtrsna.f dtrsyl.f
dtrtrs.f DEPRECATED/dtzrqf.f dtzrzf.f dstemr.f
dsgesv.f dsposv.f dlag2s.f slag2d.f dlat2s.f
dlansf.f dpftrf.f dpftri.f dpftrs.f dsfrk.f dtfsm.f dtftri.f dtfttp.f
dtfttr.f dtpttf.f dtpttr.f dtrttf.f dtrttp.f
dgejsv.f dgesvj.f dgsvj0.f dgsvj1.f
dgeequb.f dsyequb.f dpoequb.f dgbequb.f
dbbcsd.f dlapmr.f dorbdb.f dorbdb1.f dorbdb2.f dorbdb3.f dorbdb4.f
dorbdb5.f dorbdb6.f dorcsd.f dorcsd2by1.f
dgeqrt.f dgeqrt2.f dgeqrt3.f dgemqrt.f
dtpqrt.f dtpqrt2.f dtpmqrt.f dtprfb.f dpotri.f
dgelq.f dgelqt.f dgelqt3.f dgemlq.f dgemlqt.f dgemqr.f dgeqr.f dgetsls.f
dlamswlq.f dlamtsqr.f dlaswlq.f dlatsqr.f dtplqt.f dtplqt2.f dtpmlqt.f
dsysv_aa.f dsytrf_aa.f dsytrs_aa.f dlasyf_aa.f dsytf2_rk.f dlasyf_rk.f
dsytrf_rk.f dsytrs_3.f dsycon_3.f dsytri_3.f dsytri_3x.f dsysv_rk.f
dsb2st_kernels.f dsbev_2stage.f dsbevd_2stage.f dsbevx_2stage.f
dsyev_2stage.f dsyevd_2stage.f dsyevr_2stage.f dsyevx_2stage.f
dsygv_2stage.f dsytrd_2stage.f dsytrd_sb2st.F dsytrd_sy2sb.f dlarfy.f
)
set(SCATGEN slatm1.f slaran.f slarnd.f)
set(ZLASRC
zbdsqr.f zgbbrd.f zgbcon.f zgbequ.f zgbrfs.f zgbsv.f zgbsvx.f
zgbtf2.f zgbtrf.f zgbtrs.f zgebak.f zgebal.f zgebd2.f zgebrd.f
zgecon.f zgeequ.f zgees.f zgeesx.f zgeev.f zgeevx.f
DEPRECATED/zgegs.f DEPRECATED/zgegv.f zgehd2.f zgehrd.f zgejsv.f zgelq2.f zgelqf.f
zgels.f zgelsd.f zgelss.f DEPRECATED/zgelsx.f zgelsy.f zgeql2.f zgeqlf.f zgeqp3.f
DEPRECATED/zgeqpf.f zgeqr2.f zgeqr2p.f zgeqrf.f zgeqrfp.f zgerfs.f zgerq2.f zgerqf.f
zgesc2.f zgesdd.f zgesvd.f zgesvdx.f zgesvj.f zgesvx.f zgetc2.f
zgetri.f zgetrf2.f
zggbak.f zggbal.f zgges.f zgges3.f zggesx.f zggev.f zggev3.f zggevx.f zggglm.f
zgghd3.f zgghrd.f zgglse.f zggqrf.f zggrqf.f
DEPRECATED/zggsvd.f zggsvd3.f DEPRECATED/zggsvp.f zggsvp3.f
zgsvj0.f zgsvj1.f
zgtcon.f zgtrfs.f zgtsv.f zgtsvx.f zgttrf.f zgttrs.f zgtts2.f zhbev.f
zhbevd.f zhbevx.f zhbgst.f zhbgv.f zhbgvd.f zhbgvx.f zhbtrd.f
zhecon.f zheev.f zheevd.f zheevr.f zheevx.f zhegs2.f zhegst.f
zhegv.f zhegvd.f zhegvx.f zherfs.f zhesv.f zhesvx.f zhetd2.f
zhetf2.f zhetrd.f
zhetrf.f zhetri.f zhetri2.f zhetri2x.f zheswapr.f
zhetrs.f zhetrs2.f
zhetf2_rook.f zhetrf_rook.f zhetri_rook.f zhetrs_rook.f zhecon_rook.f zhesv_rook.f
zhgeqz.f zhpcon.f zhpev.f zhpevd.f
zhpevx.f zhpgst.f zhpgv.f zhpgvd.f zhpgvx.f zhprfs.f zhpsv.f
zhpsvx.f
zhptrd.f zhptrf.f zhptri.f zhptrs.f zhsein.f zhseqr.f zlabrd.f
zlacgv.f zlacon.f zlacn2.f zlacp2.f zlacpy.f zlacrm.f zlacrt.f zladiv.f
zlaed0.f zlaed7.f zlaed8.f
zlaein.f zlaesy.f zlaev2.f zlags2.f zlagtm.f
zlahef.f zlahef_rook.f zlahqr.f
DEPRECATED/zlahrd.f zlahr2.f zlaic1.f zlals0.f zlalsa.f zlalsd.f zlangb.f zlange.f
zlangt.f zlanhb.f
zlanhe.f
zlanhp.f zlanhs.f zlanht.f zlansb.f zlansp.f zlansy.f zlantb.f
zlantp.f zlantr.f zlapll.f zlapmt.f zlaqgb.f zlaqge.f
zlaqhb.f zlaqhe.f zlaqhp.f zlaqp2.f zlaqps.f zlaqsb.f
zlaqr0.f zlaqr1.f zlaqr2.f zlaqr3.f zlaqr4.f zlaqr5.f
zlaqsp.f zlaqsy.f zlar1v.f zlar2v.f ilazlr.f ilazlc.f
zlarcm.f zlarf.f zlarfb.f
zlarfg.f zlarft.f zlarfgp.f
zlarfx.f zlargv.f zlarnv.f zlarrv.f zlartg.f zlartv.f
zlarz.f zlarzb.f zlarzt.f zlascl.f zlaset.f zlasr.f
zlassq.f zlasyf.f zlasyf_rook.f zlasyf_aa.f
zlatbs.f zlatdf.f zlatps.f zlatrd.f zlatrs.f zlatrz.f DEPRECATED/zlatzm.f
zpbcon.f zpbequ.f zpbrfs.f zpbstf.f zpbsv.f
zpbsvx.f zpbtf2.f zpbtrf.f zpbtrs.f zpocon.f zpoequ.f zporfs.f
zposv.f zposvx.f zpotrf2.f zpotrs.f zpstrf.f zpstf2.f
zppcon.f zppequ.f zpprfs.f zppsv.f zppsvx.f zpptrf.f zpptri.f zpptrs.f
zptcon.f zpteqr.f zptrfs.f zptsv.f zptsvx.f zpttrf.f zpttrs.f zptts2.f
zrot.f zspcon.f zsprfs.f zspsv.f
zspsvx.f zsptrf.f zsptri.f zsptrs.f zdrscl.f zstedc.f
zstegr.f zstein.f zsteqr.f
zsycon.f zsysv_aa.f
zsyrfs.f zsysv.f zsysvx.f zsytf2.f zsytrf.f zsytri.f zsytri2.f zsytri2x.f
zsyswapr.f zsytrs.f zsytrs_aa.f zsytrs2.f zsyconv.f
zsytf2_rook.f zsytrf_rook.f zsytrs_rook.f
zsytri_rook.f zsycon_rook.f zsysv_rook.f
ztbcon.f ztbrfs.f ztbtrs.f ztgevc.f ztgex2.f
ztgexc.f ztgsen.f ztgsja.f ztgsna.f ztgsy2.f ztgsyl.f ztpcon.f
ztprfs.f ztptri.f
ztptrs.f ztrcon.f ztrevc.f ztrevc3.f ztrexc.f ztrrfs.f ztrsen.f ztrsna.f
ztrsyl.f ztrtrs.f DEPRECATED/ztzrqf.f ztzrzf.f zung2l.f
zung2r.f zungbr.f zunghr.f zungl2.f zunglq.f zungql.f zungqr.f zungr2.f
zungrq.f zungtr.f zunm2l.f zunm2r.f zunmbr.f zunmhr.f zunm22.f zunml2.f
zunmlq.f zunmql.f zunmqr.f zunmr2.f zunmr3.f zunmrq.f zunmrz.f
zunmtr.f zupgtr.f
zupmtr.f izmax1.f dzsum1.f zstemr.f
zcgesv.f zcposv.f zlag2c.f clag2z.f zlat2c.f
zhfrk.f ztfttp.f zlanhf.f zpftrf.f zpftri.f zpftrs.f ztfsm.f ztftri.f
ztfttr.f ztpttf.f ztpttr.f ztrttf.f ztrttp.f
zgeequb.f zgbequb.f zsyequb.f zpoequb.f zheequb.f
zbbcsd.f zlapmr.f zunbdb.f zunbdb1.f zunbdb2.f zunbdb3.f zunbdb4.f
zunbdb5.f zunbdb6.f zuncsd.f zuncsd2by1.f
zgeqrt.f zgeqrt2.f zgeqrt3.f zgemqrt.f
ztpqrt.f ztpqrt2.f ztpmqrt.f ztprfb.f zpotri.f
zgelq.f zgelqt.f zgelqt3.f zgemlq.f zgemlqt.f zgemqr.f zgeqr.f zgetsls.f
zlamswlq.f zlamtsqr.f zlaswlq.f zlatsqr.f ztplqt.f ztplqt2.f ztpmlqt.f
zhesv_aa.f zhetrf_aa.f zhetrs_aa.f zlahef_aa.f zsytf2_rk.f zlasyf_rk.f
zsytrf_aa.f zsytrf_rk.f zsytrs_3.f zsycon_3.f zsytri_3.f zsytri_3x.f zsysv_rk.f
zhetf2_rk.f zlahef_rk.f zhetrf_rk.f zhetrs_3.f zhecon_3.f zhetri_3.f
zhetri_3x.f zhesv_rk.f zhb2st_kernels.f zhbev_2stage.f zhbevd_2stage.f
zhbevx_2stage.f zheev_2stage.f zheevd_2stage.f zheevr_2stage.f
zheevx_2stage.f zhegv_2stage.f zhetrd_2stage.f zhetrd_hb2st.F zhetrd_he2hb.f
zlarfy.f
)
set(SMATGEN slatms.f slatme.f slatmr.f slatmt.f
slagge.f slagsy.f slakf2.f slarge.f slaror.f slarot.f slatm2.f
slatm3.f slatm5.f slatm6.f slatm7.f slahilb.f)
set(LA_REL_SRC ${ALLAUX})
if (BUILD_SINGLE)
list(APPEND LA_REL_SRC ${SLASRC} ${DSLASRC} ${SCLAUX})
endif ()
set(CMATGEN clatms.f clatme.f clatmr.f clatmt.f
clagge.f claghe.f clagsy.f clakf2.f clarge.f claror.f clarot.f
clatm1.f clarnd.f clatm2.f clatm3.f clatm5.f clatm6.f clahilb.f slatm7.f)
if (BUILD_DOUBLE)
list(APPEND LA_REL_SRC ${DLASRC} ${DSLASRC} ${DZLAUX})
endif ()
set(DZATGEN dlatm1.f dlaran.f dlarnd.f)
if (BUILD_COMPLEX)
list(APPEND LA_REL_SRC ${CLASRC} ${ZCLASRC} ${SCLAUX})
endif ()
set(DMATGEN dlatms.f dlatme.f dlatmr.f dlatmt.f
dlagge.f dlagsy.f dlakf2.f dlarge.f dlaror.f dlarot.f dlatm2.f
dlatm3.f dlatm5.f dlatm6.f dlatm7.f dlahilb.f)
if (BUILD_COMPLEX16)
list(APPEND LA_REL_SRC ${ZLASRC} ${ZCLASRC} ${DZLAUX})
endif ()
set(ZMATGEN zlatms.f zlatme.f zlatmr.f zlatmt.f
zlagge.f zlaghe.f zlagsy.f zlakf2.f zlarge.f zlaror.f zlarot.f
zlatm1.f zlarnd.f zlatm2.f zlatm3.f zlatm5.f zlatm6.f zlahilb.f dlatm7.f)
if(BUILD_SINGLE)
set(LA_REL_SRC ${SLASRC} ${DSLASRC} ${ALLAUX} ${SCLAUX})
set(LA_GEN_SRC ${SMATGEN} ${SCATGEN})
message(STATUS "Building Single Precision")
endif()
if(BUILD_DOUBLE)
set(LA_REL_SRC ${LA_REL_SRC} ${DLASRC} ${DSLASRC} ${ALLAUX} ${DZLAUX})
set(LA_GEN_SRC ${LA_GEN_SRC} ${DMATGEN} ${DZATGEN})
message(STATUS "Building Double Precision")
endif()
if(BUILD_COMPLEX)
set(LA_REL_SRC ${LA_REL_SRC} ${CLASRC} ${ZCLASRC} ${ALLAUX} ${SCLAUX})
SET(LA_GEN_SRC ${LA_GEN_SRC} ${CMATGEN} ${SCATGEN})
message(STATUS "Building Complex Precision")
endif()
if(BUILD_COMPLEX16)
set(LA_REL_SRC ${LA_REL_SRC} ${ZLASRC} ${ZCLASRC} ${ALLAUX} ${DZLAUX})
SET(LA_GEN_SRC ${LA_GEN_SRC} ${ZMATGEN} ${DZATGEN})
message(STATUS "Building Double Complex Precision")
endif()
# add lapack-netlib folder to the sources
set(LA_SOURCES "")
foreach (LA_FILE ${LA_REL_SRC})
list(APPEND LA_SOURCES "${NETLIB_LAPACK_DIR}/SRC/${LA_FILE}")
endforeach ()
foreach (LA_FILE ${LA_GEN_SRC})
list(APPEND LA_SOURCES "${NETLIB_LAPACK_DIR}/TESTING/MATGEN/${LA_FILE}")
endforeach ()
set_source_files_properties(${LA_SOURCES} PROPERTIES COMPILE_FLAGS "${LAPACK_FFLAGS}")
+193 -91
View File
@@ -1,5 +1,5 @@
set(C_SRC
set(CSRC
lapacke_cbbcsd.c
lapacke_cbbcsd_work.c
lapacke_cbdsqr.c
@@ -46,6 +46,8 @@ set(C_SRC
lapacke_cgehrd_work.c
lapacke_cgejsv.c
lapacke_cgejsv_work.c
lapacke_cgelq.c
lapacke_cgelq_work.c
lapacke_cgelq2.c
lapacke_cgelq2_work.c
lapacke_cgelqf.c
@@ -60,12 +62,18 @@ set(C_SRC
lapacke_cgelsy_work.c
lapacke_cgemqr.c
lapacke_cgemqr_work.c
lapacke_cgemlq.c
lapacke_cgemlq_work.c
lapacke_cgemqrt.c
lapacke_cgemqrt_work.c
lapacke_cgeqlf.c
lapacke_cgeqlf_work.c
lapacke_cgeqp3.c
lapacke_cgeqp3_work.c
lapacke_cgeqpf.c
lapacke_cgeqpf_work.c
lapacke_cgeqr.c
lapacke_cgeqr_work.c
lapacke_cgeqr2.c
lapacke_cgeqr2_work.c
lapacke_cgeqrf.c
@@ -134,8 +142,12 @@ set(C_SRC
lapacke_cggqrf_work.c
lapacke_cggrqf.c
lapacke_cggrqf_work.c
lapacke_cggsvd.c
lapacke_cggsvd_work.c
lapacke_cggsvd3.c
lapacke_cggsvd3_work.c
lapacke_cggsvp.c
lapacke_cggsvp_work.c
lapacke_cggsvp3.c
lapacke_cggsvp3_work.c
lapacke_cgtcon.c
@@ -210,6 +222,8 @@ set(C_SRC
lapacke_chesv_work.c
lapacke_chesv_aa.c
lapacke_chesv_aa_work.c
lapacke_chesv_aa_2stage.c
lapacke_chesv_aa_2stage_work.c
lapacke_chesv_rk.c
lapacke_chesv_rk_work.c
lapacke_chesvx.c
@@ -224,6 +238,8 @@ set(C_SRC
lapacke_chetrf_rook_work.c
lapacke_chetrf_aa.c
lapacke_chetrf_aa_work.c
lapacke_chetrf_aa_2stage.c
lapacke_chetrf_aa_2stage_work.c
lapacke_chetrf_rk.c
lapacke_chetrf_rk_work.c
lapacke_chetri.c
@@ -242,6 +258,9 @@ set(C_SRC
lapacke_chetrs_rook_work.c
lapacke_chetrs_aa.c
lapacke_chetrs_aa_work.c
lapacke_chetrs_aa_2stage.c
lapacke_chetrs_aa_2stage_work.c
lapacke_chetrf_rk.c
lapacke_chetrs_3.c
lapacke_chetrs_3_work.c
lapacke_chfrk.c
@@ -290,6 +309,11 @@ set(C_SRC
lapacke_clacp2_work.c
lapacke_clacpy.c
lapacke_clacpy_work.c
lapacke_clacrm.c
lapacke_clacrm_work.c
lapacke_clarcm.c
lapacke_clarcm_work.c
lapacke_clacn2.c
lapacke_clag2z.c
lapacke_clag2z_work.c
lapacke_clange.c
@@ -318,6 +342,8 @@ set(C_SRC
lapacke_clascl_work.c
lapacke_claset.c
lapacke_claset_work.c
lapacke_classq.c
lapacke_classq_work.c
lapacke_claswp.c
lapacke_claswp_work.c
lapacke_clauum.c
@@ -436,6 +462,8 @@ set(C_SRC
lapacke_csysv_work.c
lapacke_csysv_aa.c
lapacke_csysv_aa_work.c
lapacke_csysv_aa_2stage.c
lapacke_csysv_aa_2stage_work.c
lapacke_csysv_rk.c
lapacke_csysv_rk_work.c
lapacke_csysvx.c
@@ -448,6 +476,8 @@ set(C_SRC
lapacke_csytrf_rook_work.c
lapacke_csytrf_aa.c
lapacke_csytrf_aa_work.c
lapacke_csytrf_aa_2stage.c
lapacke_csytrf_aa_2stage_work.c
lapacke_csytrf_rk.c
lapacke_csytrf_rk_work.c
lapacke_csytri.c
@@ -466,6 +496,8 @@ set(C_SRC
lapacke_csytrs_rook_work.c
lapacke_csytrs_aa.c
lapacke_csytrs_aa_work.c
lapacke_csytrs_aa_2stage.c
lapacke_csytrs_aa_2stage_work.c
lapacke_csytrs_3.c
lapacke_csytrs_3_work.c
lapacke_ctbcon.c
@@ -633,6 +665,8 @@ set(DSRC
lapacke_dgehrd_work.c
lapacke_dgejsv.c
lapacke_dgejsv_work.c
lapacke_dgelq.c
lapacke_dgelq_work.c
lapacke_dgelq2.c
lapacke_dgelq2_work.c
lapacke_dgelqf.c
@@ -645,6 +679,8 @@ set(DSRC
lapacke_dgelss_work.c
lapacke_dgelsy.c
lapacke_dgelsy_work.c
lapacke_dgemlq.c
lapacke_dgemlq_work.c
lapacke_dgemqr.c
lapacke_dgemqr_work.c
lapacke_dgemqrt.c
@@ -653,6 +689,10 @@ set(DSRC
lapacke_dgeqlf_work.c
lapacke_dgeqp3.c
lapacke_dgeqp3_work.c
lapacke_dgeqpf.c
lapacke_dgeqpf_work.c
lapacke_dgeqr.c
lapacke_dgeqr_work.c
lapacke_dgeqr2.c
lapacke_dgeqr2_work.c
lapacke_dgeqrf.c
@@ -721,8 +761,12 @@ set(DSRC
lapacke_dggqrf_work.c
lapacke_dggrqf.c
lapacke_dggrqf_work.c
lapacke_dggsvd.c
lapacke_dggsvd_work.c
lapacke_dggsvd3.c
lapacke_dggsvd3_work.c
lapacke_dggsvp.c
lapacke_dggsvp_work.c
lapacke_dggsvp3.c
lapacke_dggsvp3_work.c
lapacke_dgtcon.c
@@ -785,6 +829,9 @@ set(DSRC
lapacke_dlaset_work.c
lapacke_dlasrt.c
lapacke_dlasrt_work.c
lapacke_dlassq.c
lapacke_dlassq_work.c
lapacke_dlaswp.c
lapacke_dlaswp.c
lapacke_dlaswp_work.c
lapacke_dlauum.c
@@ -1025,6 +1072,8 @@ set(DSRC
lapacke_dsysv_work.c
lapacke_dsysv_aa.c
lapacke_dsysv_aa_work.c
lapacke_dsysv_aa_2stage.c
lapacke_dsysv_aa_2stage_work.c
lapacke_dsysv_rk.c
lapacke_dsysv_rk_work.c
lapacke_dsysvx.c
@@ -1039,6 +1088,8 @@ set(DSRC
lapacke_dsytrf_rook_work.c
lapacke_dsytrf_aa.c
lapacke_dsytrf_aa_work.c
lapacke_dsytrf_aa_2stage.c
lapacke_dsytrf_aa_2stage_work.c
lapacke_dsytrf_rk.c
lapacke_dsytrf_rk_work.c
lapacke_dsytri.c
@@ -1055,6 +1106,8 @@ set(DSRC
lapacke_dsytrs2_work.c
lapacke_dsytrs_aa.c
lapacke_dsytrs_aa_work.c
lapacke_dsytrs_aa_2stage.c
lapacke_dsytrs_aa_2stage_work.c
lapacke_dsytrs_3.c
lapacke_dsytrs_3_work.c
lapacke_dsytrs_work.c
@@ -1184,6 +1237,8 @@ set(SSRC
lapacke_sgehrd_work.c
lapacke_sgejsv.c
lapacke_sgejsv_work.c
lapacke_sgelq.c
lapacke_sgelq_work.c
lapacke_sgelq2.c
lapacke_sgelq2_work.c
lapacke_sgelqf.c
@@ -1196,6 +1251,8 @@ set(SSRC
lapacke_sgelss_work.c
lapacke_sgelsy.c
lapacke_sgelsy_work.c
lapacke_sgemlq.c
lapacke_sgemlq_work.c
lapacke_sgemqr.c
lapacke_sgemqr_work.c
lapacke_sgemqrt.c
@@ -1204,6 +1261,10 @@ set(SSRC
lapacke_sgeqlf_work.c
lapacke_sgeqp3.c
lapacke_sgeqp3_work.c
lapacke_sgeqpf.c
lapacke_sgeqpf_work.c
lapacke_sgeqr.c
lapacke_sgeqr_work.c
lapacke_sgeqr2.c
lapacke_sgeqr2_work.c
lapacke_sgeqrf.c
@@ -1272,8 +1333,12 @@ set(SSRC
lapacke_sggqrf_work.c
lapacke_sggrqf.c
lapacke_sggrqf_work.c
lapacke_sggsvd.c
lapacke_sggsvd_work.c
lapacke_sggsvd3.c
lapacke_sggsvd3_work.c
lapacke_sggsvp.c
lapacke_sggsvp_work.c
lapacke_sggsvp3.c
lapacke_sggsvp3_work.c
lapacke_sgtcon.c
@@ -1336,6 +1401,8 @@ set(SSRC
lapacke_slaset_work.c
lapacke_slasrt.c
lapacke_slasrt_work.c
lapacke_slassq.c
lapacke_slassq_work.c
lapacke_slaswp.c
lapacke_slaswp_work.c
lapacke_slauum.c
@@ -1572,6 +1639,8 @@ set(SSRC
lapacke_ssysv_work.c
lapacke_ssysv_aa.c
lapacke_ssysv_aa_work.c
lapacke_ssysv_aa_2stage.c
lapacke_ssysv_aa_2stage_work.c
lapacke_ssysv_rk.c
lapacke_ssysv_rk_work.c
lapacke_ssysvx.c
@@ -1586,6 +1655,9 @@ set(SSRC
lapacke_ssytrf_rook_work.c
lapacke_ssytrf_aa.c
lapacke_ssytrf_aa_work.c
lapacke_ssytrf_aa_2stage.c
lapacke_ssytrf_aa_2stage_work.c
lapacke_ssytrf_rook.c
lapacke_ssytrf_rk.c
lapacke_ssytrf_rk_work.c
lapacke_ssytri.c
@@ -1602,6 +1674,8 @@ set(SSRC
lapacke_ssytrs2_work.c
lapacke_ssytrs_aa.c
lapacke_ssytrs_aa_work.c
lapacke_ssytrs_aa_2stage.c
lapacke_ssytrs_aa_2stage_work.c
lapacke_ssytrs_3.c
lapacke_ssytrs_3_work.c
lapacke_ssytrs_work.c
@@ -1729,6 +1803,8 @@ set(ZSRC
lapacke_zgehrd_work.c
lapacke_zgejsv.c
lapacke_zgejsv_work.c
lapacke_zgelq.c
lapacke_zgelq_work.c
lapacke_zgelq2.c
lapacke_zgelq2_work.c
lapacke_zgelqf.c
@@ -1741,6 +1817,8 @@ set(ZSRC
lapacke_zgelss_work.c
lapacke_zgelsy.c
lapacke_zgelsy_work.c
lapacke_zgemlq.c
lapacke_zgemlq_work.c
lapacke_zgemqr.c
lapacke_zgemqr_work.c
lapacke_zgemqrt.c
@@ -1749,6 +1827,10 @@ set(ZSRC
lapacke_zgeqlf_work.c
lapacke_zgeqp3.c
lapacke_zgeqp3_work.c
lapacke_zgeqpf.c
lapacke_zgeqpf_work.c
lapacke_zgeqr.c
lapacke_zgeqr_work.c
lapacke_zgeqr2.c
lapacke_zgeqr2_work.c
lapacke_zgeqrf.c
@@ -1817,8 +1899,12 @@ set(ZSRC
lapacke_zggqrf_work.c
lapacke_zggrqf.c
lapacke_zggrqf_work.c
lapacke_zggsvd.c
lapacke_zggsvd_work.c
lapacke_zggsvd3.c
lapacke_zggsvd3_work.c
lapacke_zggsvp.c
lapacke_zggsvp_work.c
lapacke_zggsvp3.c
lapacke_zggsvp3_work.c
lapacke_zgtcon.c
@@ -1839,6 +1925,12 @@ set(ZSRC
lapacke_zhbevd_work.c
lapacke_zhbevx.c
lapacke_zhbevx_work.c
lapacke_zhbev_2stage.c
lapacke_zhbev_2stage_work.c
lapacke_zhbevd_2stage.c
lapacke_zhbevd_2stage_work.c
lapacke_zhbevx_2stage.c
lapacke_zhbevx_2stage_work.c
lapacke_zhbgst.c
lapacke_zhbgst_work.c
lapacke_zhbgv.c
@@ -1887,6 +1979,8 @@ set(ZSRC
lapacke_zhesv_work.c
lapacke_zhesv_aa.c
lapacke_zhesv_aa_work.c
lapacke_zhesv_aa_2stage.c
lapacke_zhesv_aa_2stage_work.c
lapacke_zhesv_rk.c
lapacke_zhesv_rk_work.c
lapacke_zhesvx.c
@@ -1901,6 +1995,8 @@ set(ZSRC
lapacke_zhetrf_rook_work.c
lapacke_zhetrf_aa.c
lapacke_zhetrf_aa_work.c
lapacke_zhetrf_aa_2stage.c
lapacke_zhetrf_aa_2stage_work.c
lapacke_zhetrf_rk.c
lapacke_zhetrf_rk_work.c
lapacke_zhetri.c
@@ -1918,6 +2014,8 @@ set(ZSRC
lapacke_zhetrs_work.c
lapacke_zhetrs_aa.c
lapacke_zhetrs_aa_work.c
lapacke_zhetrs_aa_2stage.c
lapacke_zhetrs_aa_2stage_work.c
lapacke_zhetrs_3.c
lapacke_zhetrs_3_work.c
lapacke_zhetrs_rook_work.c
@@ -1967,6 +2065,8 @@ set(ZSRC
lapacke_zlacp2_work.c
lapacke_zlacpy.c
lapacke_zlacpy_work.c
lapacke_zlacrm.c
lapacke_zlacrm_work.c
lapacke_zlag2c.c
lapacke_zlag2c_work.c
lapacke_zlange.c
@@ -1981,6 +2081,8 @@ set(ZSRC
lapacke_zlapmr_work.c
lapacke_zlapmt.c
lapacke_zlapmt_work.c
lapacke_zlarcm.c
lapacke_zlarcm_work.c
lapacke_zlarfb.c
lapacke_zlarfb_work.c
lapacke_zlarfg.c
@@ -1995,6 +2097,8 @@ set(ZSRC
lapacke_zlascl_work.c
lapacke_zlaset.c
lapacke_zlaset_work.c
lapacke_zlassq.c
lapacke_zlassq_work.c
lapacke_zlaswp.c
lapacke_zlaswp_work.c
lapacke_zlauum.c
@@ -2113,6 +2217,8 @@ set(ZSRC
lapacke_zsysv_work.c
lapacke_zsysv_aa.c
lapacke_zsysv_aa_work.c
lapacke_zsysv_aa_2stage.c
lapacke_zsysv_aa_2stage_work.c
lapacke_zsysv_rk.c
lapacke_zsysv_rk_work.c
lapacke_zsysvx.c
@@ -2125,6 +2231,8 @@ set(ZSRC
lapacke_zsytrf_rook_work.c
lapacke_zsytrf_aa.c
lapacke_zsytrf_aa_work.c
lapacke_zsytrf_aa_2stage.c
lapacke_zsytrf_aa_2stage_work.c
lapacke_zsytrf_rk.c
lapacke_zsytrf_rk_work.c
lapacke_zsytri.c
@@ -2143,6 +2251,8 @@ set(ZSRC
lapacke_zsytrs_rook_work.c
lapacke_zsytrs_aa.c
lapacke_zsytrs_aa_work.c
lapacke_zsytrs_aa_2stage.c
lapacke_zsytrs_aa_2stage_work.c
lapacke_zsytrs_3.c
lapacke_zsytrs_3_work.c
lapacke_ztbcon.c
@@ -2263,104 +2373,92 @@ set(ZSRC
)
set(SRCX
lapacke_cgbrfsx.c lapacke_cporfsx.c lapacke_dgerfsx.c lapacke_sgbrfsx.c lapacke_ssyrfsx.c lapacke_zherfsx.c
lapacke_cgbrfsx_work.c lapacke_cporfsx_work.c lapacke_dgerfsx_work.c lapacke_sgbrfsx_work.c lapacke_ssyrfsx_work.c lapacke_zherfsx_work.c
lapacke_cgerfsx.c lapacke_csyrfsx.c lapacke_dporfsx.c lapacke_sgerfsx.c lapacke_zgbrfsx.c lapacke_zporfsx.c
lapacke_cgerfsx_work.c lapacke_csyrfsx_work.c lapacke_dporfsx_work.c lapacke_sgerfsx_work.c lapacke_zgbrfsx_work.c lapacke_zporfsx_work.c
lapacke_cherfsx.c lapacke_dgbrfsx.c lapacke_dsyrfsx.c lapacke_sporfsx.c lapacke_zgerfsx.c lapacke_zsyrfsx.c
lapacke_cherfsx_work.c lapacke_dgbrfsx_work.c lapacke_dsyrfsx_work.c lapacke_sporfsx_work.c lapacke_zgerfsx_work.c lapacke_zsyrfsx_work.c
lapacke_cgbsvxx.c lapacke_cposvxx.c lapacke_dgesvxx.c lapacke_sgbsvxx.c lapacke_ssysvxx.c lapacke_zhesvxx.c
lapacke_cgbsvxx_work.c lapacke_cposvxx_work.c lapacke_dgesvxx_work.c lapacke_sgbsvxx_work.c lapacke_ssysvxx_work.c lapacke_zhesvxx_work.c
lapacke_cgesvxx.c lapacke_csysvxx.c lapacke_dposvxx.c lapacke_sgesvxx.c lapacke_zgbsvxx.c lapacke_zposvxx.c
lapacke_cgesvxx_work.c lapacke_csysvxx_work.c lapacke_dposvxx_work.c lapacke_sgesvxx_work.c lapacke_zgbsvxx_work.c lapacke_zposvxx_work.c
lapacke_chesvxx.c lapacke_dgbsvxx.c lapacke_dsysvxx.c lapacke_sposvxx.c lapacke_zgesvxx.c lapacke_zsysvxx.c
lapacke_cgbrfsx.c lapacke_cporfsx.c lapacke_dgerfsx.c lapacke_sgbrfsx.c lapacke_ssyrfsx.c lapacke_zherfsx.c
lapacke_cgbrfsx_work.c lapacke_cporfsx_work.c lapacke_dgerfsx_work.c lapacke_sgbrfsx_work.c lapacke_ssyrfsx_work.c lapacke_zherfsx_work.c
lapacke_cgerfsx.c lapacke_csyrfsx.c lapacke_dporfsx.c lapacke_sgerfsx.c lapacke_zgbrfsx.c lapacke_zporfsx.c
lapacke_cgerfsx_work.c lapacke_csyrfsx_work.c lapacke_dporfsx_work.c lapacke_sgerfsx_work.c lapacke_zgbrfsx_work.c lapacke_zporfsx_work.c
lapacke_cherfsx.c lapacke_dgbrfsx.c lapacke_dsyrfsx.c lapacke_sporfsx.c lapacke_zgerfsx.c lapacke_zsyrfsx.c
lapacke_cherfsx_work.c lapacke_dgbrfsx_work.c lapacke_dsyrfsx_work.c lapacke_sporfsx_work.c lapacke_zgerfsx_work.c lapacke_zsyrfsx_work.c
lapacke_cgbsvxx.c lapacke_cposvxx.c lapacke_dgesvxx.c lapacke_sgbsvxx.c lapacke_ssysvxx.c lapacke_zhesvxx.c
lapacke_cgbsvxx_work.c lapacke_cposvxx_work.c lapacke_dgesvxx_work.c lapacke_sgbsvxx_work.c lapacke_ssysvxx_work.c lapacke_zhesvxx_work.c
lapacke_cgesvxx.c lapacke_csysvxx.c lapacke_dposvxx.c lapacke_sgesvxx.c lapacke_zgbsvxx.c lapacke_zposvxx.c
lapacke_cgesvxx_work.c lapacke_csysvxx_work.c lapacke_dposvxx_work.c lapacke_sgesvxx_work.c lapacke_zgbsvxx_work.c lapacke_zposvxx_work.c
lapacke_chesvxx.c lapacke_dgbsvxx.c lapacke_dsysvxx.c lapacke_sposvxx.c lapacke_zgesvxx.c lapacke_zsysvxx.c
lapacke_chesvxx_work.c lapacke_dgbsvxx_work.c lapacke_dsysvxx_work.c lapacke_sposvxx_work.c lapacke_zgesvxx_work.c lapacke_zsysvxx_work.c
)
# FILE PARTS OF TMGLIB
# FILE PARTS OF TMGLIB
set(MATGEN
lapacke_clatms.c
lapacke_clatms_work.c
lapacke_dlatms.c
lapacke_dlatms_work.c
lapacke_slatms.c
lapacke_slatms_work.c
lapacke_zlatms.c
lapacke_zlatms_work.c
lapacke_clagge.c
lapacke_clagge_work.c
lapacke_dlagge.c
lapacke_dlagge_work.c
lapacke_slagge.c
lapacke_slagge_work.c
lapacke_zlagge.c
lapacke_zlagge_work.c
lapacke_claghe.c
lapacke_claghe_work.c
lapacke_zlaghe.c
lapacke_zlaghe_work.c
lapacke_clagsy.c
lapacke_clagsy_work.c
lapacke_dlagsy.c
lapacke_dlagsy_work.c
lapacke_slagsy.c
lapacke_slagsy_work.c
lapacke_zlagsy.c
lapacke_clatms.c
lapacke_clatms_work.c
lapacke_dlatms.c
lapacke_dlatms_work.c
lapacke_slatms.c
lapacke_slatms_work.c
lapacke_zlatms.c
lapacke_zlatms_work.c
lapacke_clagge.c
lapacke_clagge_work.c
lapacke_dlagge.c
lapacke_dlagge_work.c
lapacke_slagge.c
lapacke_slagge_work.c
lapacke_zlagge.c
lapacke_zlagge_work.c
lapacke_claghe.c
lapacke_claghe_work.c
lapacke_zlaghe.c
lapacke_zlaghe_work.c
lapacke_clagsy.c
lapacke_clagsy_work.c
lapacke_dlagsy.c
lapacke_dlagsy_work.c
lapacke_slagsy.c
lapacke_slagsy_work.c
lapacke_zlagsy.c
lapacke_zlagsy_work.c
lapacke_nancheck.c
)
set(Utils_SRC
lapacke_cgb_nancheck.c lapacke_dpf_nancheck.c lapacke_ssy_trans.c
lapacke_cgb_trans.c lapacke_dpf_trans.c lapacke_stb_nancheck.c
lapacke_cge_nancheck.c lapacke_dpo_nancheck.c lapacke_stb_trans.c
lapacke_cge_trans.c lapacke_dpo_trans.c lapacke_stf_nancheck.c
lapacke_cgg_nancheck.c lapacke_dpp_nancheck.c lapacke_stf_trans.c
lapacke_cgg_trans.c lapacke_dpp_trans.c lapacke_stp_nancheck.c
lapacke_cgt_nancheck.c lapacke_dpt_nancheck.c lapacke_stp_trans.c
lapacke_chb_nancheck.c lapacke_dsb_nancheck.c lapacke_str_nancheck.c
lapacke_chb_trans.c lapacke_dsb_trans.c lapacke_str_trans.c
lapacke_che_nancheck.c lapacke_dsp_nancheck.c lapacke_xerbla.c
lapacke_che_trans.c lapacke_dsp_trans.c lapacke_zgb_nancheck.c
lapacke_chp_nancheck.c lapacke_dst_nancheck.c lapacke_zgb_trans.c
lapacke_chp_trans.c lapacke_dsy_nancheck.c lapacke_zge_nancheck.c
lapacke_chs_nancheck.c lapacke_dsy_trans.c lapacke_zge_trans.c
lapacke_chs_trans.c lapacke_dtb_nancheck.c lapacke_zgg_nancheck.c
lapacke_c_nancheck.c lapacke_dtb_trans.c lapacke_zgg_trans.c
lapacke_cpb_nancheck.c lapacke_dtf_nancheck.c lapacke_zgt_nancheck.c
lapacke_cpb_trans.c lapacke_dtf_trans.c lapacke_zhb_nancheck.c
lapacke_cpf_nancheck.c lapacke_dtp_nancheck.c lapacke_zhb_trans.c
lapacke_cpf_trans.c lapacke_dtp_trans.c lapacke_zhe_nancheck.c
lapacke_cpo_nancheck.c lapacke_dtr_nancheck.c lapacke_zhe_trans.c
lapacke_cpo_trans.c lapacke_dtr_trans.c lapacke_zhp_nancheck.c
lapacke_cpp_nancheck.c lapacke_lsame.c lapacke_zhp_trans.c
lapacke_cpp_trans.c lapacke_make_complex_double.c lapacke_zhs_nancheck.c
lapacke_cpt_nancheck.c lapacke_make_complex_float.c lapacke_zhs_trans.c
lapacke_csp_nancheck.c lapacke_sgb_nancheck.c lapacke_z_nancheck.c
lapacke_csp_trans.c lapacke_sgb_trans.c lapacke_zpb_nancheck.c
lapacke_cst_nancheck.c lapacke_sge_nancheck.c lapacke_zpb_trans.c
lapacke_csy_nancheck.c lapacke_sge_trans.c lapacke_zpf_nancheck.c
lapacke_csy_trans.c lapacke_sgg_nancheck.c lapacke_zpf_trans.c
lapacke_ctb_nancheck.c lapacke_sgg_trans.c lapacke_zpo_nancheck.c
lapacke_ctb_trans.c lapacke_sgt_nancheck.c lapacke_zpo_trans.c
lapacke_ctf_nancheck.c lapacke_shs_nancheck.c lapacke_zpp_nancheck.c
lapacke_ctf_trans.c lapacke_shs_trans.c lapacke_zpp_trans.c
lapacke_ctp_nancheck.c lapacke_s_nancheck.c lapacke_zpt_nancheck.c
lapacke_ctp_trans.c lapacke_spb_nancheck.c lapacke_zsp_nancheck.c
lapacke_ctr_nancheck.c lapacke_spb_trans.c lapacke_zsp_trans.c
lapacke_ctr_trans.c lapacke_spf_nancheck.c lapacke_zst_nancheck.c
lapacke_dgb_nancheck.c lapacke_spf_trans.c lapacke_zsy_nancheck.c
lapacke_dgb_trans.c lapacke_spo_nancheck.c lapacke_zsy_trans.c
lapacke_dge_nancheck.c lapacke_spo_trans.c lapacke_ztb_nancheck.c
lapacke_dge_trans.c lapacke_spp_nancheck.c lapacke_ztb_trans.c
lapacke_dgg_nancheck.c lapacke_spp_trans.c lapacke_ztf_nancheck.c
lapacke_dgg_trans.c lapacke_spt_nancheck.c lapacke_ztf_trans.c
lapacke_dgt_nancheck.c lapacke_ssb_nancheck.c lapacke_ztp_nancheck.c
lapacke_dhs_nancheck.c lapacke_ssb_trans.c lapacke_ztp_trans.c
lapacke_dhs_trans.c lapacke_ssp_nancheck.c lapacke_ztr_nancheck.c
lapacke_d_nancheck.c lapacke_ssp_trans.c lapacke_ztr_trans.c
lapacke_dpb_nancheck.c lapacke_sst_nancheck.c
lapacke_dpb_trans.c lapacke_ssy_nancheck.c
lapacke_c_nancheck.c lapacke_ctr_trans.c lapacke_make_complex_float.c lapacke_zgb_nancheck.c
lapacke_cgb_nancheck.c lapacke_d_nancheck.c lapacke_s_nancheck.c lapacke_zgb_trans.c
lapacke_cgb_trans.c lapacke_dgb_nancheck.c lapacke_sgb_nancheck.c lapacke_zge_nancheck.c
lapacke_cge_nancheck.c lapacke_dgb_trans.c lapacke_sgb_trans.c lapacke_zge_trans.c
lapacke_cge_trans.c lapacke_dge_nancheck.c lapacke_sge_nancheck.c lapacke_zgg_nancheck.c
lapacke_cgg_nancheck.c lapacke_dge_trans.c lapacke_sge_trans.c lapacke_zgg_trans.c
lapacke_cgg_trans.c lapacke_dgg_nancheck.c lapacke_sgg_nancheck.c lapacke_zgt_nancheck.c
lapacke_cgt_nancheck.c lapacke_dgg_trans.c lapacke_sgg_trans.c lapacke_zhb_nancheck.c
lapacke_chb_nancheck.c lapacke_dgt_nancheck.c lapacke_sgt_nancheck.c lapacke_zhb_trans.c
lapacke_chb_trans.c lapacke_dhs_nancheck.c lapacke_shs_nancheck.c lapacke_zhe_nancheck.c
lapacke_che_nancheck.c lapacke_dhs_trans.c lapacke_shs_trans.c lapacke_zhe_trans.c
lapacke_che_trans.c lapacke_dpb_nancheck.c lapacke_spb_nancheck.c lapacke_zhp_nancheck.c
lapacke_chp_nancheck.c lapacke_dpb_trans.c lapacke_spb_trans.c lapacke_zhp_trans.c
lapacke_chp_trans.c lapacke_dpf_nancheck.c lapacke_spf_nancheck.c lapacke_zhs_nancheck.c
lapacke_chs_nancheck.c lapacke_dpf_trans.c lapacke_spf_trans.c lapacke_zhs_trans.c
lapacke_chs_trans.c lapacke_dpo_nancheck.c lapacke_spo_nancheck.c lapacke_zpb_nancheck.c
lapacke_cpb_nancheck.c lapacke_dpo_trans.c lapacke_spo_trans.c lapacke_zpb_trans.c
lapacke_cpb_trans.c lapacke_dpp_nancheck.c lapacke_spp_nancheck.c lapacke_zpf_nancheck.c
lapacke_cpf_nancheck.c lapacke_dpp_trans.c lapacke_spp_trans.c lapacke_zpf_trans.c
lapacke_cpf_trans.c lapacke_dpt_nancheck.c lapacke_spt_nancheck.c lapacke_zpo_nancheck.c
lapacke_cpo_nancheck.c lapacke_dsb_nancheck.c lapacke_ssb_nancheck.c lapacke_zpo_trans.c
lapacke_cpo_trans.c lapacke_dsb_trans.c lapacke_ssb_trans.c lapacke_zpp_nancheck.c
lapacke_cpp_nancheck.c lapacke_dsp_nancheck.c lapacke_ssp_nancheck.c lapacke_zpp_trans.c
lapacke_cpp_trans.c lapacke_dsp_trans.c lapacke_ssp_trans.c lapacke_zpt_nancheck.c
lapacke_cpt_nancheck.c lapacke_dst_nancheck.c lapacke_sst_nancheck.c lapacke_zsp_nancheck.c
lapacke_csp_nancheck.c lapacke_dsy_nancheck.c lapacke_ssy_nancheck.c lapacke_zsp_trans.c
lapacke_csp_trans.c lapacke_dsy_trans.c lapacke_ssy_trans.c lapacke_zst_nancheck.c
lapacke_cst_nancheck.c lapacke_dtb_nancheck.c lapacke_stb_nancheck.c lapacke_zsy_nancheck.c
lapacke_csy_nancheck.c lapacke_dtb_trans.c lapacke_stb_trans.c lapacke_zsy_trans.c
lapacke_csy_trans.c lapacke_dtf_nancheck.c lapacke_stf_nancheck.c lapacke_ztb_nancheck.c
lapacke_ctb_nancheck.c lapacke_dtf_trans.c lapacke_stf_trans.c lapacke_ztb_trans.c
lapacke_ctb_trans.c lapacke_dtp_nancheck.c lapacke_stp_nancheck.c lapacke_ztf_nancheck.c
lapacke_ctf_nancheck.c lapacke_dtp_trans.c lapacke_stp_trans.c lapacke_ztf_trans.c
lapacke_ctf_trans.c lapacke_dtr_nancheck.c lapacke_str_nancheck.c lapacke_ztp_nancheck.c
lapacke_ctp_nancheck.c lapacke_dtr_trans.c lapacke_str_trans.c lapacke_ztp_trans.c
lapacke_ctp_trans.c lapacke_lsame.c lapacke_xerbla.c lapacke_ztr_nancheck.c
lapacke_ctr_nancheck.c lapacke_make_complex_double.c lapacke_z_nancheck.c lapacke_ztr_trans.c
)
set(LAPACKE_REL_SRC "")
@@ -2380,6 +2478,10 @@ if (BUILD_COMPLEX16)
list(APPEND LAPACKE_REL_SRC ${ZSRC})
endif ()
if (BUILD_MATGEN)
list(APPEND LAPACKE_REL_SRC ${MATGEN})
endif ()
# add lapack-netlib folder to the sources
set(LAPACKE_SOURCES "")
foreach (LAE_FILE ${LAPACKE_REL_SRC})
@@ -2391,6 +2493,6 @@ foreach (Utils_FILE ${Utils_SRC})
endforeach ()
set(lapacke_include_dir "${NETLIB_LAPACK_DIR}/LAPACKE/include")
execute_process(COMMAND ${CMAKE_COMMAND} -E copy "${lapacke_include_dir}/lapacke_mangling_with_flags.h.in" "${lapacke_include_dir}/lapacke_mangling.h")
configure_file("${lapacke_include_dir}/lapacke_mangling_with_flags.h.in" "${lapacke_include_dir}/lapacke_mangling.h" COPYONLY)
include_directories(${lapacke_include_dir})
set_source_files_properties(${LAPACKE_SOURCES} PROPERTIES COMPILE_FLAGS "${LAPACK_CFLAGS}")
+1
View File
@@ -1,6 +1,7 @@
libdir=@CMAKE_INSTALL_FULL_LIBDIR@
includedir=@CMAKE_INSTALL_FULL_INCLUDEDIR@
openblas_config=USE_64BITINT=@USE_64BITINT@ NO_CBLAS=@NO_CBLAS@ NO_LAPACK=@NO_LAPACK@ NO_LAPACKE=@NO_LAPACKE@ DYNAMIC_ARCH=@DYNAMIC_ARCH@ DYNAMIC_OLDER=@DYNAMIC_OLDER@ NO_AFFINITY=@NO_AFFINITY@ USE_OPENMP=@USE_OPENMP@ @CORE@ MAX_THREADS=@NUM_THREADS@
Name: OpenBLAS
Description: OpenBLAS is an optimized BLAS library based on GotoBLAS2 1.13 BSD version
Version: @OPENBLAS_VERSION@
+3 -16
View File
@@ -3,19 +3,6 @@
## Description: Ported from portion of OpenBLAS/Makefile.system
## Detects the OS and sets appropriate variables.
if (${CMAKE_SYSTEM_NAME} STREQUAL "Darwin")
set(ENV{MACOSX_DEPLOYMENT_TARGET} "10.2") # TODO: should be exported as an env var
set(MD5SUM "md5 -r")
endif ()
if (${CMAKE_SYSTEM_NAME} STREQUAL "FreeBSD")
set(MD5SUM "md5 -r")
endif ()
if (${CMAKE_SYSTEM_NAME} STREQUAL "NetBSD")
set(MD5SUM "md5 -n")
endif ()
if (${CMAKE_SYSTEM_NAME} STREQUAL "Linux")
set(EXTRALIB "${EXTRALIB} -lm")
set(NO_EXPRECISION 1)
@@ -56,7 +43,7 @@ if (${CMAKE_SYSTEM_NAME} STREQUAL "Windows")
# Ensure the correct stack alignment on Win32
# http://permalink.gmane.org/gmane.comp.lib.openblas.general/97
if (${ARCH} STREQUAL "x86")
if (X86)
if (NOT MSVC AND NOT ${CMAKE_C_COMPILER_ID} STREQUAL "Clang")
set(CCOMMON_OPT "${CCOMMON_OPT} -mincoming-stack-boundary=2")
endif ()
@@ -78,7 +65,7 @@ if (CYGWIN)
endif ()
if (NOT ${CMAKE_SYSTEM_NAME} STREQUAL "Windows" AND NOT ${CMAKE_SYSTEM_NAME} STREQUAL "Interix" AND NOT ${CMAKE_SYSTEM_NAME} STREQUAL "Android")
if (SMP)
if (USE_THREAD)
set(EXTRALIB "${EXTRALIB} -lpthread")
endif ()
endif ()
@@ -88,7 +75,7 @@ if (QUAD_PRECISION)
set(NO_EXPRECISION 1)
endif ()
if (${ARCH} STREQUAL "x86")
if (X86)
set(NO_EXPRECISION 1)
endif ()
+186 -90
View File
@@ -37,112 +37,208 @@
# CPUIDEMU = ../../cpuid/table.o
if (DEFINED CPUIDEMU)
set(EXFLAGS "-DCPUIDEMU -DVENDOR=99")
endif ()
if (DEFINED TARGET_CORE)
if (BUILD_KERNEL)
# set the C flags for just this file
set(GETARCH2_FLAGS "-DBUILD_KERNEL")
set(TARGET_MAKE "Makefile_kernel.conf")
set(TARGET_CONF "config_kernel.h")
set(TARGET_CONF_DIR ${PROJECT_BINARY_DIR}/kernel_config/${TARGET_CORE})
else()
set(TARGET_MAKE "Makefile.conf")
set(TARGET_CONF "config.h")
set(TARGET_CONF_DIR ${PROJECT_BINARY_DIR})
endif ()
include("${PROJECT_SOURCE_DIR}/cmake/c_check.cmake")
set(TARGET_CONF_TEMP "${PROJECT_BINARY_DIR}/${TARGET_CONF}.tmp")
# c_check
set(FU "")
if (APPLE OR (MSVC AND NOT ${CMAKE_C_COMPILER_ID} MATCHES "Clang"))
set(FU "_")
endif()
set(COMPILER_ID ${CMAKE_C_COMPILER_ID})
if (${COMPILER_ID} STREQUAL "GNU")
set(COMPILER_ID "GCC")
endif ()
string(TOUPPER ${ARCH} UC_ARCH)
file(WRITE ${TARGET_CONF_TEMP}
"#define OS_${HOST_OS}\t1\n"
"#define ARCH_${UC_ARCH}\t1\n"
"#define C_${COMPILER_ID}\t1\n"
"#define __${BINARY}BIT__\t1\n"
"#define FUNDERSCORE\t${FU}\n")
if (${HOST_OS} STREQUAL "WINDOWSSTORE")
file(APPEND ${TARGET_CONF_TEMP}
"#define OS_WINNT\t1\n")
endif ()
# f_check
if (NOT NOFORTRAN)
include("${PROJECT_SOURCE_DIR}/cmake/f_check.cmake")
endif ()
# compile getarch
set(GETARCH_SRC
${PROJECT_SOURCE_DIR}/getarch.c
${CPUIDEMO}
)
# Cannot run getarch on target if we are cross-compiling
if (DEFINED CORE AND CMAKE_CROSSCOMPILING)
# Write to config as getarch would
if (NOT MSVC)
list(APPEND GETARCH_SRC ${PROJECT_SOURCE_DIR}/cpuid.S)
endif ()
# TODO: Set up defines that getarch sets up based on every other target
# Perhaps this should be inside a different file as it grows larger
file(APPEND ${TARGET_CONF_TEMP}
"#define ${CORE}\n"
"#define CHAR_CORENAME \"${CORE}\"\n")
if ("${CORE}" STREQUAL "ARMV7")
file(APPEND ${TARGET_CONF_TEMP}
"#define L1_DATA_SIZE\t65536\n"
"#define L1_DATA_LINESIZE\t32\n"
"#define L2_SIZE\t512488\n"
"#define L2_LINESIZE\t32\n"
"#define DTB_DEFAULT_ENTRIES\t64\n"
"#define DTB_SIZE\t4096\n"
"#define L2_ASSOCIATIVE\t4\n"
"#define HAVE_VFPV3\n"
"#define HAVE_VFP\n")
set(SGEMM_UNROLL_M 4)
set(SGEMM_UNROLL_N 4)
set(DGEMM_UNROLL_M 4)
set(DGEMM_UNROLL_N 4)
elseif ("${CORE}" STREQUAL "ARMV8")
file(APPEND ${TARGET_CONF_TEMP}
"#define L1_DATA_SIZE\t32768\n"
"#define L1_DATA_LINESIZE\t64\n"
"#define L2_SIZE\t262144\n"
"#define L2_LINESIZE\t64\n"
"#define DTB_DEFAULT_ENTRIES\t64\n"
"#define DTB_SIZE\t4096\n"
"#define L2_ASSOCIATIVE\t32\n")
set(SGEMM_UNROLL_M 4)
set(SGEMM_UNROLL_N 4)
elseif ("${CORE}" STREQUAL "CORTEXA57")
file(APPEND ${TARGET_CONF_TEMP}
"#define L1_CODE_SIZE\t49152\n"
"#define L1_CODE_LINESIZE\t64\n"
"#define L1_CODE_ASSOCIATIVE\t3\n"
"#define L1_DATA_SIZE\t32768\n"
"#define L1_DATA_LINESIZE\t64\n"
"#define L1_DATA_ASSOCIATIVE\t2\n"
"#define L2_SIZE\t2097152\n"
"#define L2_LINESIZE\t64\n"
"#define L2_ASSOCIATIVE\t16\n"
"#define DTB_DEFAULT_ENTRIES\t64\n"
"#define DTB_SIZE\t4096\n"
"#define HAVE_VFPV4\n"
"#define HAVE_VFPV3\n"
"#define HAVE_VFP\n"
"#define HAVE_NEON\n")
set(SGEMM_UNROLL_M 16)
set(SGEMM_UNROLL_N 4)
set(DGEMM_UNROLL_M 8)
set(DGEMM_UNROLL_N 4)
set(CGEMM_UNROLL_M 8)
set(CGEMM_UNROLL_N 4)
set(ZGEMM_UNROLL_M 8)
set(ZGEMM_UNROLL_N 4)
endif()
# Or should this actually be NUM_CORES?
if (${NUM_THREADS} GREATER 0)
file(APPEND ${TARGET_CONF_TEMP} "#define NUM_CORES\t${NUM_THREADS}\n")
endif()
# GetArch_2nd
foreach(float_char S;D;Q;C;Z;X)
if (NOT DEFINED ${float_char}GEMM_UNROLL_M)
set(${float_char}GEMM_UNROLL_M 2)
endif()
if (NOT DEFINED ${float_char}GEMM_UNROLL_N)
set(${float_char}GEMM_UNROLL_N 2)
endif()
endforeach()
file(APPEND ${TARGET_CONF_TEMP}
"#define GEMM_MULTITHREAD_THRESHOLD\t${GEMM_MULTITHREAD_THRESHOLD}\n")
# Move to where gen_config_h would place it
file(RENAME ${TARGET_CONF_TEMP} "${TARGET_CONF_DIR}/${TARGET_CONF}")
else(NOT CMAKE_CROSSCOMPILING)
# compile getarch
set(GETARCH_SRC
${PROJECT_SOURCE_DIR}/getarch.c
${CPUIDEMU}
)
if ("${CMAKE_C_COMPILER_ID}" STREQUAL "MSVC")
#Use generic for MSVC now
message("MSVC")
set(GETARCH_FLAGS ${GETARCH_FLAGS} -DFORCE_GENERIC)
else()
list(APPEND GETARCH_SRC ${PROJECT_SOURCE_DIR}/cpuid.S)
endif ()
if ("${CMAKE_SYSTEM_NAME}" STREQUAL "WindowsStore")
# disable WindowsStore strict CRT checks
set(GETARCH_FLAGS ${GETARCH_FLAGS} -D_CRT_SECURE_NO_WARNINGS)
endif ()
set(GETARCH_DIR "${PROJECT_BINARY_DIR}/getarch_build")
set(GETARCH_BIN "getarch${CMAKE_EXECUTABLE_SUFFIX}")
file(MAKE_DIRECTORY ${GETARCH_DIR})
configure_file(${TARGET_CONF_TEMP} ${GETARCH_DIR}/${TARGET_CONF} COPYONLY)
if (NOT "${CMAKE_SYSTEM_NAME}" STREQUAL "WindowsStore")
try_compile(GETARCH_RESULT ${GETARCH_DIR}
SOURCES ${GETARCH_SRC}
COMPILE_DEFINITIONS ${EXFLAGS} ${GETARCH_FLAGS} -I${GETARCH_DIR} -I"${PROJECT_SOURCE_DIR}" -I"${PROJECT_BINARY_DIR}"
OUTPUT_VARIABLE GETARCH_LOG
COPY_FILE ${PROJECT_BINARY_DIR}/${GETARCH_BIN}
)
if (NOT ${GETARCH_RESULT})
MESSAGE(FATAL_ERROR "Compiling getarch failed ${GETARCH_LOG}")
endif ()
endif ()
message(STATUS "Running getarch")
# use the cmake binary w/ the -E param to run a shell command in a cross-platform way
execute_process(COMMAND "${PROJECT_BINARY_DIR}/${GETARCH_BIN}" 0 OUTPUT_VARIABLE GETARCH_MAKE_OUT)
execute_process(COMMAND "${PROJECT_BINARY_DIR}/${GETARCH_BIN}" 1 OUTPUT_VARIABLE GETARCH_CONF_OUT)
message(STATUS "GETARCH results:\n${GETARCH_MAKE_OUT}")
# append config data from getarch to the TARGET file and read in CMake vars
file(APPEND ${TARGET_CONF_TEMP} ${GETARCH_CONF_OUT})
ParseGetArchVars(${GETARCH_MAKE_OUT})
set(GETARCH2_DIR "${PROJECT_BINARY_DIR}/getarch2_build")
set(GETARCH2_BIN "getarch_2nd${CMAKE_EXECUTABLE_SUFFIX}")
file(MAKE_DIRECTORY ${GETARCH2_DIR})
configure_file(${TARGET_CONF_TEMP} ${GETARCH2_DIR}/${TARGET_CONF} COPYONLY)
if (NOT "${CMAKE_SYSTEM_NAME}" STREQUAL "WindowsStore")
try_compile(GETARCH2_RESULT ${GETARCH2_DIR}
SOURCES ${PROJECT_SOURCE_DIR}/getarch_2nd.c
COMPILE_DEFINITIONS ${EXFLAGS} ${GETARCH_FLAGS} ${GETARCH2_FLAGS} -I${GETARCH2_DIR} -I"${PROJECT_SOURCE_DIR}" -I"${PROJECT_BINARY_DIR}"
OUTPUT_VARIABLE GETARCH2_LOG
COPY_FILE ${PROJECT_BINARY_DIR}/${GETARCH2_BIN}
)
if (NOT ${GETARCH2_RESULT})
MESSAGE(FATAL_ERROR "Compiling getarch_2nd failed ${GETARCH2_LOG}")
endif ()
endif ()
# use the cmake binary w/ the -E param to run a shell command in a cross-platform way
execute_process(COMMAND "${PROJECT_BINARY_DIR}/${GETARCH2_BIN}" 0 OUTPUT_VARIABLE GETARCH2_MAKE_OUT)
execute_process(COMMAND "${PROJECT_BINARY_DIR}/${GETARCH2_BIN}" 1 OUTPUT_VARIABLE GETARCH2_CONF_OUT)
# append config data from getarch_2nd to the TARGET file and read in CMake vars
file(APPEND ${TARGET_CONF_TEMP} ${GETARCH2_CONF_OUT})
configure_file(${TARGET_CONF_TEMP} ${TARGET_CONF_DIR}/${TARGET_CONF} COPYONLY)
ParseGetArchVars(${GETARCH2_MAKE_OUT})
if (MSVC)
#Use generic for MSVC now
set(GETARCH_FLAGS ${GETARCH_FLAGS} -DFORCE_GENERIC)
endif()
if ("${CMAKE_SYSTEM_NAME}" STREQUAL "WindowsStore")
# disable WindowsStore strict CRT checks
set(GETARCH_FLAGS ${GETARCH_FLAGS} -D_CRT_SECURE_NO_WARNINGS)
endif ()
set(GETARCH_DIR "${PROJECT_BINARY_DIR}/getarch_build")
set(GETARCH_BIN "getarch${CMAKE_EXECUTABLE_SUFFIX}")
file(MAKE_DIRECTORY ${GETARCH_DIR})
if (NOT "${CMAKE_SYSTEM_NAME}" STREQUAL "WindowsStore")
try_compile(GETARCH_RESULT ${GETARCH_DIR}
SOURCES ${GETARCH_SRC}
COMPILE_DEFINITIONS ${EXFLAGS} ${GETARCH_FLAGS} -I${PROJECT_SOURCE_DIR}
OUTPUT_VARIABLE GETARCH_LOG
COPY_FILE ${PROJECT_BINARY_DIR}/${GETARCH_BIN}
)
if (NOT ${GETARCH_RESULT})
MESSAGE(FATAL_ERROR "Compiling getarch failed ${GETARCH_LOG}")
endif ()
endif ()
message(STATUS "Running getarch")
# use the cmake binary w/ the -E param to run a shell command in a cross-platform way
execute_process(COMMAND ${PROJECT_BINARY_DIR}/${GETARCH_BIN} 0 OUTPUT_VARIABLE GETARCH_MAKE_OUT)
execute_process(COMMAND ${PROJECT_BINARY_DIR}/${GETARCH_BIN} 1 OUTPUT_VARIABLE GETARCH_CONF_OUT)
message(STATUS "GETARCH results:\n${GETARCH_MAKE_OUT}")
# append config data from getarch to the TARGET file and read in CMake vars
file(APPEND ${TARGET_CONF} ${GETARCH_CONF_OUT})
ParseGetArchVars(${GETARCH_MAKE_OUT})
set(GETARCH2_DIR "${PROJECT_BINARY_DIR}/getarch2_build")
set(GETARCH2_BIN "getarch_2nd${CMAKE_EXECUTABLE_SUFFIX}")
file(MAKE_DIRECTORY ${GETARCH2_DIR})
if (NOT "${CMAKE_SYSTEM_NAME}" STREQUAL "WindowsStore")
try_compile(GETARCH2_RESULT ${GETARCH2_DIR}
SOURCES ${PROJECT_SOURCE_DIR}/getarch_2nd.c
COMPILE_DEFINITIONS ${EXFLAGS} ${GETARCH_FLAGS} ${GETARCH2_FLAGS} -I${PROJECT_SOURCE_DIR}
OUTPUT_VARIABLE GETARCH2_LOG
COPY_FILE ${PROJECT_BINARY_DIR}/${GETARCH2_BIN}
)
if (NOT ${GETARCH2_RESULT})
MESSAGE(FATAL_ERROR "Compiling getarch_2nd failed ${GETARCH2_LOG}")
endif ()
endif ()
# use the cmake binary w/ the -E param to run a shell command in a cross-platform way
execute_process(COMMAND ${PROJECT_BINARY_DIR}/${GETARCH2_BIN} 0 OUTPUT_VARIABLE GETARCH2_MAKE_OUT)
execute_process(COMMAND ${PROJECT_BINARY_DIR}/${GETARCH2_BIN} 1 OUTPUT_VARIABLE GETARCH2_CONF_OUT)
# append config data from getarch_2nd to the TARGET file and read in CMake vars
file(APPEND ${TARGET_CONF} ${GETARCH2_CONF_OUT})
ParseGetArchVars(${GETARCH2_MAKE_OUT})
# compile get_config_h
set(GEN_CONFIG_H_DIR "${PROJECT_BINARY_DIR}/genconfig_h_build")
set(GEN_CONFIG_H_BIN "gen_config_h${CMAKE_EXECUTABLE_SUFFIX}")
set(GEN_CONFIG_H_FLAGS "-DVERSION=\"${OpenBLAS_VERSION}\"")
file(MAKE_DIRECTORY ${GEN_CONFIG_H_DIR})
if (NOT "${CMAKE_SYSTEM_NAME}" STREQUAL "WindowsStore")
try_compile(GEN_CONFIG_H_RESULT ${GEN_CONFIG_H_DIR}
SOURCES ${PROJECT_SOURCE_DIR}/gen_config_h.c
COMPILE_DEFINITIONS ${EXFLAGS} ${GETARCH_FLAGS} ${GEN_CONFIG_H_FLAGS} -I${PROJECT_SOURCE_DIR}
OUTPUT_VARIABLE GEN_CONFIG_H_LOG
COPY_FILE ${PROJECT_BINARY_DIR}/${GEN_CONFIG_H_BIN}
)
if (NOT ${GEN_CONFIG_H_RESULT})
MESSAGE(FATAL_ERROR "Compiling gen_config_h failed ${GEN_CONFIG_H_LOG}")
endif ()
endif ()
+78 -136
View File
@@ -2,13 +2,27 @@
## Author: Hank Anderson <hank@statease.com>
## Description: Ported from OpenBLAS/Makefile.system
##
set(NETLIB_LAPACK_DIR "${PROJECT_SOURCE_DIR}/lapack-netlib")
# TODO: Makefile.system detects Darwin (mac) and switches to clang here -hpa
# http://stackoverflow.com/questions/714100/os-detecting-makefile
# System detection, via CMake.
include("${PROJECT_SOURCE_DIR}/cmake/system_check.cmake")
# TODO: Makefile.system sets HOSTCC = $(CC) here if not already set -hpa
if(CMAKE_CROSSCOMPILING AND NOT DEFINED TARGET)
# Detect target without running getarch
if (ARM64)
set(TARGET "ARMV8")
elseif(ARM)
set(TARGET "ARMV7") # TODO: Ask compiler which arch this is
else()
message(FATAL_ERROR "When cross compiling, a TARGET is required.")
endif()
endif()
# Other files expect CORE, which is actually TARGET and will become TARGET_CORE for kernel build. Confused yet?
# It seems we are meant to use TARGET as input and CORE internally as kernel.
if(NOT DEFINED CORE AND DEFINED TARGET)
set(CORE ${TARGET})
endif()
# TARGET_CORE will override TARGET which is used in DYNAMIC_ARCH=1.
if (DEFINED TARGET_CORE)
@@ -19,7 +33,7 @@ endif ()
if (DEFINED BINARY AND DEFINED TARGET AND BINARY EQUAL 32)
message(STATUS "Compiling a ${BINARY}-bit binary.")
set(NO_AVX 1)
if (${TARGET} STREQUAL "HASWELL" OR ${TARGET} STREQUAL "SANDYBRIDGE")
if (${TARGET} STREQUAL "HASWELL" OR ${TARGET} STREQUAL "SANDYBRIDGE" OR ${TARGET} STREQUAL "SKYLAKEX")
set(TARGET "NEHALEM")
endif ()
if (${TARGET} STREQUAL "BULLDOZER" OR ${TARGET} STREQUAL "PILEDRIVER" OR ${TARGET} STREQUAL "ZEN")
@@ -28,7 +42,7 @@ if (DEFINED BINARY AND DEFINED TARGET AND BINARY EQUAL 32)
endif ()
if (DEFINED TARGET)
message(STATUS "Targetting the ${TARGET} architecture.")
message(STATUS "Targeting the ${TARGET} architecture.")
set(GETARCH_FLAGS "-DFORCE_${TARGET}")
endif ()
@@ -53,21 +67,16 @@ if (NO_AVX2)
set(GETARCH_FLAGS "${GETARCH_FLAGS} -DNO_AVX2")
endif ()
if (CMAKE_BUILD_TYPE STREQUAL Debug)
set(GETARCH_FLAGS "${GETARCH_FLAGS} -g")
if (CMAKE_BUILD_TYPE STREQUAL "Debug")
set(GETARCH_FLAGS "${GETARCH_FLAGS} ${CMAKE_C_FLAGS_DEBUG}")
endif ()
# TODO: let CMake handle this? -hpa
#if (${QUIET_MAKE})
# set(MAKE "${MAKE} -s")
#endif()
if (NOT DEFINED NO_PARALLEL_MAKE)
set(NO_PARALLEL_MAKE 0)
endif ()
set(GETARCH_FLAGS "${GETARCH_FLAGS} -DNO_PARALLEL_MAKE=${NO_PARALLEL_MAKE}")
if (CMAKE_CXX_COMPILER STREQUAL loongcc)
if (CMAKE_C_COMPILER STREQUAL loongcc)
set(GETARCH_FLAGS "${GETARCH_FLAGS} -static")
endif ()
@@ -78,51 +87,44 @@ else ()
set(ONLY_CBLAS 0)
endif ()
include("${PROJECT_SOURCE_DIR}/cmake/prebuild.cmake")
# N.B. this is NUM_THREAD in Makefile.system which is probably a bug -hpa
if (NOT CMAKE_CROSSCOMPILING)
if (NOT DEFINED NUM_CORES)
include(ProcessorCount)
ProcessorCount(NUM_CORES)
endif()
endif()
if (NOT DEFINED NUM_PARALLEL)
set(NUM_PARALLEL 1)
endif()
if (NOT DEFINED NUM_THREADS)
set(NUM_THREADS ${NUM_CORES})
endif ()
if (DEFINED NUM_CORES AND NOT NUM_CORES EQUAL 0)
# HT?
set(NUM_THREADS ${NUM_CORES})
else ()
set(NUM_THREADS 0)
endif ()
endif()
if (${NUM_THREADS} EQUAL 1)
if (${NUM_THREADS} LESS 2)
set(USE_THREAD 0)
elseif(NOT DEFINED USE_THREAD)
set(USE_THREAD 1)
endif ()
if (DEFINED USE_THREAD)
if (NOT ${USE_THREAD})
unset(SMP)
else ()
set(SMP 1)
endif ()
else ()
# N.B. this is NUM_THREAD in Makefile.system which is probably a bug -hpa
if (${NUM_THREADS} EQUAL 1)
unset(SMP)
else ()
set(SMP 1)
endif ()
if (USE_THREAD)
message(STATUS "Multi-threading enabled with ${NUM_THREADS} threads.")
endif ()
if (${SMP})
message(STATUS "SMP enabled.")
endif ()
include("${PROJECT_SOURCE_DIR}/cmake/prebuild.cmake")
if (NOT DEFINED NEED_PIC)
set(NEED_PIC 1)
endif ()
# TODO: I think CMake should be handling all this stuff -hpa
unset(ARFLAGS)
set(CPP "${COMPILER} -E")
set(AR "${CROSS_SUFFIX}ar")
set(AS "${CROSS_SUFFIX}as")
set(LD "${CROSS_SUFFIX}ld")
set(RANLIB "${CROSS_SUFFIX}ranlib")
set(NM "${CROSS_SUFFIX}nm")
set(DLLWRAP "${CROSS_SUFFIX}dllwrap")
set(OBJCOPY "${CROSS_SUFFIX}objcopy")
set(OBJCONV "${CROSS_SUFFIX}objconv")
# OS dependent settings
include("${PROJECT_SOURCE_DIR}/cmake/os.cmake")
@@ -150,15 +152,20 @@ if (NEED_PIC)
set(CCOMMON_OPT "${CCOMMON_OPT} -fPIC")
endif ()
if (${F_COMPILER} STREQUAL "SUN")
set(FCOMMON_OPT "${FCOMMON_OPT} -pic")
else ()
set(FCOMMON_OPT "${FCOMMON_OPT} -fPIC")
endif ()
if (NOT NOFORTRAN)
if (${F_COMPILER} STREQUAL "SUN")
set(FCOMMON_OPT "${FCOMMON_OPT} -pic")
else ()
set(FCOMMON_OPT "${FCOMMON_OPT} -fPIC")
endif ()
endif()
endif ()
if (DYNAMIC_ARCH)
set(CCOMMON_OPT "${CCOMMON_OPT} -DDYNAMIC_ARCH")
if (DYNAMIC_OLDER)
set(CCOMMON_OPT "${CCOMMON_OPT} -DDYNAMIC_OLDER")
endif ()
endif ()
if (NO_LAPACK)
@@ -175,7 +182,7 @@ if (NO_AVX)
set(CCOMMON_OPT "${CCOMMON_OPT} -DNO_AVX")
endif ()
if (${ARCH} STREQUAL "x86")
if (X86)
set(CCOMMON_OPT "${CCOMMON_OPT} -DNO_AVX")
endif ()
@@ -183,25 +190,20 @@ if (NO_AVX2)
set(CCOMMON_OPT "${CCOMMON_OPT} -DNO_AVX2")
endif ()
if (SMP)
if (USE_THREAD)
# USE_SIMPLE_THREADED_LEVEL3 = 1
# NO_AFFINITY = 1
set(CCOMMON_OPT "${CCOMMON_OPT} -DSMP_SERVER")
if (${ARCH} STREQUAL "mips64")
if (MIPS64)
if (NOT ${CORE} STREQUAL "LOONGSON3B")
set(USE_SIMPLE_THREADED_LEVEL3 1)
endif ()
endif ()
if (USE_OPENMP)
# USE_SIMPLE_THREADED_LEVEL3 = 1
# NO_AFFINITY = 1
set(CCOMMON_OPT "${CCOMMON_OPT} -DUSE_OPENMP")
endif ()
if (BIGNUMA)
set(CCOMMON_OPT "${CCOMMON_OPT} -DBIGNUMA")
endif ()
endif ()
if (NO_WARMUP)
@@ -229,6 +231,8 @@ endif ()
set(CCOMMON_OPT "${CCOMMON_OPT} -DMAX_CPU_NUMBER=${NUM_THREADS}")
set(CCOMMON_OPT "${CCOMMON_OPT} -DMAX_PARALLEL_NUMBER=${NUM_PARALLEL}")
if (USE_SIMPLE_THREADED_LEVEL3)
set(CCOMMON_OPT "${CCOMMON_OPT} -DUSE_SIMPLE_THREADED_LEVEL3")
endif ()
@@ -264,7 +268,7 @@ if (NOT ${CMAKE_SYSTEM_NAME} STREQUAL "Linux")
set(NO_AFFINITY 1)
endif ()
if (NOT ${ARCH} STREQUAL "x86_64" AND NOT ${ARCH} STREQUAL "x86" AND NOT ${CORE} STREQUAL "LOONGSON3B")
if (NOT X86_64 AND NOT X86 AND NOT ${CORE} STREQUAL "LOONGSON3B")
set(NO_AFFINITY 1)
endif ()
@@ -296,52 +300,19 @@ if (MIXED_MEMORY_ALLOCATION)
set(CCOMMON_OPT "${CCOMMON_OPT} -DMIXED_MEMORY_ALLOCATION")
endif ()
if (${CMAKE_SYSTEM_NAME} STREQUAL "SunOS")
set(TAR gtar)
set(PATCH gpatch)
set(GREP ggrep)
else ()
set(TAR tar)
set(PATCH patch)
set(GREP grep)
endif ()
if (NOT DEFINED MD5SUM)
set(MD5SUM md5sum)
endif ()
set(AWK awk)
set(SED sed)
set(REVISION "-r${OpenBLAS_VERSION}")
set(MAJOR_VERSION ${OpenBLAS_MAJOR_VERSION})
if (DEBUG)
set(COMMON_OPT "${COMMON_OPT} -g")
endif ()
if (NOT DEFINED COMMON_OPT)
set(COMMON_OPT "-O2")
endif ()
#For x86 32-bit
if (DEFINED BINARY AND BINARY EQUAL 32)
if (NOT MSVC)
set(COMMON_OPT "${COMMON_OPT} -m32")
endif()
endif()
set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} ${COMMON_OPT} ${CCOMMON_OPT}")
set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} ${CCOMMON_OPT}")
if(NOT MSVC)
set(CMAKE_ASM_FLAGS "${CMAKE_ASM_FLAGS} ${COMMON_OPT} ${CCOMMON_OPT}")
set(CMAKE_ASM_FLAGS "${CMAKE_ASM_FLAGS} ${CCOMMON_OPT}")
endif()
# TODO: not sure what PFLAGS is -hpa
set(PFLAGS "${PFLAGS} ${COMMON_OPT} ${CCOMMON_OPT} -I${TOPDIR} -DPROFILE ${COMMON_PROF}")
set(PFLAGS "${PFLAGS} ${CCOMMON_OPT} -I${TOPDIR} -DPROFILE ${COMMON_PROF}")
set(CMAKE_Fortran_FLAGS "${CMAKE_Fortran_FLAGS} ${COMMON_OPT} ${FCOMMON_OPT}")
set(CMAKE_Fortran_FLAGS "${CMAKE_Fortran_FLAGS} ${FCOMMON_OPT}")
# TODO: not sure what FPFLAGS is -hpa
set(FPFLAGS "${FPFLAGS} ${COMMON_OPT} ${FCOMMON_OPT} ${COMMON_PROF}")
set(FPFLAGS "${FPFLAGS} ${FCOMMON_OPT} ${COMMON_PROF}")
#For LAPACK Fortran codes.
set(LAPACK_FFLAGS "${LAPACK_FFLAGS} ${CMAKE_Fortran_FLAGS}")
@@ -349,7 +320,7 @@ set(LAPACK_FPFLAGS "${LAPACK_FPFLAGS} ${FPFLAGS}")
#Disable -fopenmp for LAPACK Fortran codes on Windows.
if (${CMAKE_SYSTEM_NAME} STREQUAL "Windows")
set(FILTER_FLAGS "-fopenmp;-mp;-openmp;-xopenmp=parralel")
set(FILTER_FLAGS "-fopenmp;-mp;-openmp;-xopenmp=parallel")
foreach (FILTER_FLAG ${FILTER_FLAGS})
string(REPLACE ${FILTER_FLAG} "" LAPACK_FFLAGS ${LAPACK_FFLAGS})
string(REPLACE ${FILTER_FLAG} "" LAPACK_FPFLAGS ${LAPACK_FPFLAGS})
@@ -387,7 +358,7 @@ if (NOT DEFINED LIBSUFFIX)
endif ()
if (DYNAMIC_ARCH)
if (DEFINED SMP)
if (USE_THREAD)
set(LIBNAME "${LIBPREFIX}p${REVISION}.${LIBSUFFIX}")
set(LIBNAME_P "${LIBPREFIX}p${REVISION}_p.${LIBSUFFIX}")
else ()
@@ -395,7 +366,7 @@ if (DYNAMIC_ARCH)
set(LIBNAME_P "${LIBPREFIX}${REVISION}_p.${LIBSUFFIX}")
endif ()
else ()
if (DEFINED SMP)
if (USE_THREAD)
set(LIBNAME "${LIBPREFIX}_${LIBCORE}p${REVISION}.${LIBSUFFIX}")
set(LIBNAME_P "${LIBPREFIX}_${LIBCORE}p${REVISION}_p.${LIBSUFFIX}")
else ()
@@ -426,6 +397,9 @@ if (NOT NO_LAPACK)
if (NOT NO_LAPACKE)
set(LIB_COMPONENTS "${LIB_COMPONENTS} LAPACKE")
endif ()
if (BUILD_RELAPACK)
set(LIB_COMPONENTS "${LIB_COMPONENTS} ReLAPACK")
endif ()
endif ()
if (ONLY_CBLAS)
@@ -437,7 +411,7 @@ endif ()
set(USE_GEMM3M 0)
if (DEFINED ARCH)
if (${ARCH} STREQUAL "x86" OR ${ARCH} STREQUAL "x86_64" OR ${ARCH} STREQUAL "ia64" OR ${ARCH} STREQUAL "MIPS")
if (X86 OR X86_64 OR ${ARCH} STREQUAL "ia64" OR MIPS64)
set(USE_GEMM3M 1)
endif ()
@@ -520,35 +494,3 @@ endif ()
# export CUFLAGS
# export CULIB
#endif
#.SUFFIXES: .$(PSUFFIX) .$(SUFFIX) .f
#
#.f.$(SUFFIX):
# $(FC) $(FFLAGS) -c $< -o $(@F)
#
#.f.$(PSUFFIX):
# $(FC) $(FPFLAGS) -pg -c $< -o $(@F)
# these are not cross-platform
#ifdef BINARY64
#PATHSCALEPATH = /opt/pathscale/lib/3.1
#PGIPATH = /opt/pgi/linux86-64/7.1-5/lib
#else
#PATHSCALEPATH = /opt/pathscale/lib/3.1/32
#PGIPATH = /opt/pgi/linux86/7.1-5/lib
#endif
#ACMLPATH = /opt/acml/4.3.0
#ifneq ($(OSNAME), Darwin)
#MKLPATH = /opt/intel/mkl/10.2.2.025/lib
#else
#MKLPATH = /Library/Frameworks/Intel_MKL.framework/Versions/10.0.1.014/lib
#endif
#ATLASPATH = /opt/atlas/3.9.17/opteron
#FLAMEPATH = $(HOME)/flame/lib
#ifneq ($(OSNAME), SunOS)
#SUNPATH = /opt/sunstudio12.1
#else
#SUNPATH = /opt/SUNWspro
#endif
+77
View File
@@ -0,0 +1,77 @@
##
## Author: Hank Anderson <hank@statease.com>
## Description: Ported from the OpenBLAS/c_check perl script.
## This is triggered by prebuild.cmake and runs before any of the code is built.
## Creates config.h and Makefile.conf.
# Convert CMake vars into the format that OpenBLAS expects
string(TOUPPER ${CMAKE_SYSTEM_NAME} HOST_OS)
if (${HOST_OS} STREQUAL "WINDOWS")
set(HOST_OS WINNT)
endif ()
if(CMAKE_COMPILER_IS_GNUCC AND WIN32)
execute_process(COMMAND ${CMAKE_C_COMPILER} -dumpmachine
OUTPUT_VARIABLE OPENBLAS_GCC_TARGET_MACHINE
OUTPUT_STRIP_TRAILING_WHITESPACE)
if(OPENBLAS_GCC_TARGET_MACHINE MATCHES "amd64|x86_64|AMD64")
set(MINGW64 1)
endif()
endif()
# Pretty thorough determination of arch. Add more if needed
if(CMAKE_CL_64 OR MINGW64)
set(X86_64 1)
elseif(MINGW OR (MSVC AND NOT CMAKE_CROSSCOMPILING))
set(X86 1)
elseif(CMAKE_SYSTEM_PROCESSOR MATCHES "ppc.*|power.*|Power.*")
set(PPC 1)
elseif(CMAKE_SYSTEM_PROCESSOR MATCHES "mips64.*")
set(MIPS64 1)
elseif(CMAKE_SYSTEM_PROCESSOR MATCHES "amd64.*|x86_64.*|AMD64.*")
set(X86_64 1)
elseif(CMAKE_SYSTEM_PROCESSOR MATCHES "i686.*|i386.*|x86.*|amd64.*|AMD64.*")
set(X86 1)
elseif(CMAKE_SYSTEM_PROCESSOR MATCHES "^(arm.*|ARM.*)")
set(ARM 1)
elseif(CMAKE_SYSTEM_PROCESSOR MATCHES "^(aarch64.*|AARCH64.*)")
set(ARM64 1)
endif()
if (X86_64)
set(ARCH "x86_64")
elseif(X86)
set(ARCH "x86")
elseif(PPC)
set(ARCH "power")
elseif(ARM)
set(ARCH "arm")
elseif(ARM64)
set(ARCH "arm64")
else()
set(ARCH ${CMAKE_SYSTEM_PROCESSOR} CACHE STRING "Target Architecture")
endif ()
if (NOT BINARY)
if (X86_64 OR ARM64 OR PPC OR MIPS64)
set(BINARY 64)
else ()
set(BINARY 32)
endif ()
endif()
if(BINARY EQUAL 64)
set(BINARY64 1)
else()
set(BINARY32 1)
endif()
if (X86_64 OR X86)
file(WRITE ${PROJECT_BINARY_DIR}/avx512.tmp "int main(void){ __asm__ volatile(\"vbroadcastss -4 * 4(%rsi), %zmm2\"); }")
execute_process(COMMAND ${CMAKE_C_COMPILER} -march=skylake-avx512 -v -o ${PROJECT_BINARY_DIR}/avx512.o -x c ${PROJECT_BINARY_DIR}/avx512.tmp OUTPUT_QUIET ERROR_QUIET RESULT_VARIABLE NO_AVX512)
if (NO_AVX512 EQUAL 1)
set (CCOMMON_OPT "${CCOMMON_OPT} -DNO_AVX512")
endif()
file(REMOVE "avx512.tmp" "avx512.o")
endif()
+5 -1
View File
@@ -202,6 +202,8 @@ function(GenerateNamedObjects sources_in)
if (use_cblas)
set(obj_name "cblas_${obj_name}")
list(APPEND obj_defines "CBLAS")
elseif (NOT "${obj_name}" MATCHES "${ARCH_SUFFIX}")
set(obj_name "${obj_name}${ARCH_SUFFIX}")
endif ()
list(APPEND obj_defines "ASMNAME=${FU}${obj_name};ASMFNAME=${FU}${obj_name}${BU};NAME=${obj_name}${BU};CNAME=${obj_name};CHAR_NAME=\"${obj_name}${BU}\";CHAR_CNAME=\"${obj_name}\"")
@@ -234,7 +236,9 @@ function(GenerateNamedObjects sources_in)
string(REPLACE ";" "\n#define " define_source "${obj_defines}")
string(REPLACE "=" " " define_source "${define_source}")
file(WRITE ${new_source_file} "#define ${define_source}\n#include \"${old_source_file}\"")
file(WRITE ${new_source_file}.tmp "#define ${define_source}\n#include \"${old_source_file}\"")
configure_file(${new_source_file}.tmp ${new_source_file} COPYONLY)
file(REMOVE ${new_source_file}.tmp)
list(APPEND SRC_LIST_OUT ${new_source_file})
endforeach ()
+45 -27
View File
@@ -93,7 +93,7 @@ extern "C" {
#include <sched.h>
#endif
#if defined(OS_DARWIN) || defined(OS_FREEBSD) || defined(OS_NETBSD) || defined(OS_ANDROID)
#if defined(OS_DARWIN) || defined(OS_FREEBSD) || defined(OS_NETBSD) || defined(OS_OPENBSD) || defined(OS_DRAGONFLY) || defined(OS_ANDROID)
#include <sched.h>
#endif
@@ -179,7 +179,7 @@ extern "C" {
#define ALLOCA_ALIGN 63UL
#define NUM_BUFFERS (MAX_CPU_NUMBER * 2)
#define NUM_BUFFERS (MAX_CPU_NUMBER * 2 * MAX_PARALLEL_NUMBER)
#ifdef NEEDBUNDERSCORE
#define BLASFUNC(FUNC) FUNC##_
@@ -495,6 +495,33 @@ static void __inline blas_lock(volatile BLASULONG *address){
#define MMAP_POLICY (MAP_PRIVATE | MAP_ANONYMOUS)
#endif
#ifndef ASSEMBLER
/* C99 supports complex floating numbers natively, which GCC also offers as an
extension since version 3.0. If neither are available, use a compatible
structure as fallback (see Clause 6.2.5.13 of the C99 standard). */
#if ((defined(__STDC_IEC_559_COMPLEX__) || __STDC_VERSION__ >= 199901L || \
(__GNUC__ >= 3 && !defined(__cplusplus))) && !(defined(FORCE_OPENBLAS_COMPLEX_STRUCT))) && !defined(_MSC_VER)
#define OPENBLAS_COMPLEX_C99
#ifndef __cplusplus
#include <complex.h>
#endif
typedef float _Complex openblas_complex_float;
typedef double _Complex openblas_complex_double;
typedef xdouble _Complex openblas_complex_xdouble;
#define openblas_make_complex_float(real, imag) ((real) + ((imag) * _Complex_I))
#define openblas_make_complex_double(real, imag) ((real) + ((imag) * _Complex_I))
#define openblas_make_complex_xdouble(real, imag) ((real) + ((imag) * _Complex_I))
#else
#define OPENBLAS_COMPLEX_STRUCT
typedef struct { float real, imag; } openblas_complex_float;
typedef struct { double real, imag; } openblas_complex_double;
typedef struct { xdouble real, imag; } openblas_complex_xdouble;
#define openblas_make_complex_float(real, imag) {(real), (imag)}
#define openblas_make_complex_double(real, imag) {(real), (imag)}
#define openblas_make_complex_xdouble(real, imag) {(real), (imag)}
#endif
#endif
#include "param.h"
#include "common_param.h"
@@ -524,31 +551,6 @@ static void __inline blas_lock(volatile BLASULONG *address){
#include <stdio.h>
#endif // NOINCLUDE
/* C99 supports complex floating numbers natively, which GCC also offers as an
extension since version 3.0. If neither are available, use a compatible
structure as fallback (see Clause 6.2.5.13 of the C99 standard). */
#if ((defined(__STDC_IEC_559_COMPLEX__) || __STDC_VERSION__ >= 199901L || \
(__GNUC__ >= 3 && !defined(__cplusplus))) && !(defined(FORCE_OPENBLAS_COMPLEX_STRUCT)))
#define OPENBLAS_COMPLEX_C99
#ifndef __cplusplus
#include <complex.h>
#endif
typedef float _Complex openblas_complex_float;
typedef double _Complex openblas_complex_double;
typedef xdouble _Complex openblas_complex_xdouble;
#define openblas_make_complex_float(real, imag) ((real) + ((imag) * _Complex_I))
#define openblas_make_complex_double(real, imag) ((real) + ((imag) * _Complex_I))
#define openblas_make_complex_xdouble(real, imag) ((real) + ((imag) * _Complex_I))
#else
#define OPENBLAS_COMPLEX_STRUCT
typedef struct { float real, imag; } openblas_complex_float;
typedef struct { double real, imag; } openblas_complex_double;
typedef struct { xdouble real, imag; } openblas_complex_xdouble;
#define openblas_make_complex_float(real, imag) {(real), (imag)}
#define openblas_make_complex_double(real, imag) {(real), (imag)}
#define openblas_make_complex_xdouble(real, imag) {(real), (imag)}
#endif
#ifdef XDOUBLE
#define OPENBLAS_COMPLEX_FLOAT openblas_complex_xdouble
#define OPENBLAS_MAKE_COMPLEX_FLOAT(r,i) openblas_make_complex_xdouble(r,i)
@@ -640,6 +642,7 @@ void gotoblas_profile_init(void);
void gotoblas_profile_quit(void);
#ifdef USE_OPENMP
#ifndef C_MSVC
int omp_in_parallel(void);
int omp_get_num_procs(void);
@@ -647,6 +650,21 @@ int omp_get_num_procs(void);
__declspec(dllimport) int __cdecl omp_in_parallel(void);
__declspec(dllimport) int __cdecl omp_get_num_procs(void);
#endif
#if (__STDC_VERSION__ >= 201112L)
#if defined(C_GCC) && ( __GNUC__ < 7)
// workaround for GCC bug 65467
#ifndef _Atomic
#define _Atomic volatile
#endif
#endif
#include <stdatomic.h>
#else
#ifndef _Atomic
#define _Atomic volatile
#endif
#endif
#else
#ifdef __ELF__
int omp_in_parallel (void) __attribute__ ((weak));
+7 -7
View File
@@ -333,8 +333,8 @@ BLASLONG (*icamin_k)(BLASLONG, float *, BLASLONG);
float (*cnrm2_k) (BLASLONG, float *, BLASLONG);
float (*casum_k) (BLASLONG, float *, BLASLONG);
int (*ccopy_k) (BLASLONG, float *, BLASLONG, float *, BLASLONG);
float _Complex (*cdotu_k) (BLASLONG, float *, BLASLONG, float *, BLASLONG);
float _Complex (*cdotc_k) (BLASLONG, float *, BLASLONG, float *, BLASLONG);
openblas_complex_float (*cdotu_k) (BLASLONG, float *, BLASLONG, float *, BLASLONG);
openblas_complex_float (*cdotc_k) (BLASLONG, float *, BLASLONG, float *, BLASLONG);
int (*csrot_k) (BLASLONG, float *, BLASLONG, float *, BLASLONG, float, float);
int (*caxpy_k) (BLASLONG, BLASLONG, BLASLONG, float, float, float *, BLASLONG, float *, BLASLONG, float *, BLASLONG);
@@ -496,8 +496,8 @@ BLASLONG (*izamin_k)(BLASLONG, double *, BLASLONG);
double (*znrm2_k) (BLASLONG, double *, BLASLONG);
double (*zasum_k) (BLASLONG, double *, BLASLONG);
int (*zcopy_k) (BLASLONG, double *, BLASLONG, double *, BLASLONG);
double _Complex (*zdotu_k) (BLASLONG, double *, BLASLONG, double *, BLASLONG);
double _Complex (*zdotc_k) (BLASLONG, double *, BLASLONG, double *, BLASLONG);
openblas_complex_double (*zdotu_k) (BLASLONG, double *, BLASLONG, double *, BLASLONG);
openblas_complex_double (*zdotc_k) (BLASLONG, double *, BLASLONG, double *, BLASLONG);
int (*zdrot_k) (BLASLONG, double *, BLASLONG, double *, BLASLONG, double, double);
int (*zaxpy_k) (BLASLONG, BLASLONG, BLASLONG, double, double, double *, BLASLONG, double *, BLASLONG, double *, BLASLONG);
@@ -661,8 +661,8 @@ BLASLONG (*ixamin_k)(BLASLONG, xdouble *, BLASLONG);
xdouble (*xnrm2_k) (BLASLONG, xdouble *, BLASLONG);
xdouble (*xasum_k) (BLASLONG, xdouble *, BLASLONG);
int (*xcopy_k) (BLASLONG, xdouble *, BLASLONG, xdouble *, BLASLONG);
xdouble _Complex (*xdotu_k) (BLASLONG, xdouble *, BLASLONG, xdouble *, BLASLONG);
xdouble _Complex (*xdotc_k) (BLASLONG, xdouble *, BLASLONG, xdouble *, BLASLONG);
openblas_complex_xdouble (*xdotu_k) (BLASLONG, xdouble *, BLASLONG, xdouble *, BLASLONG);
openblas_complex_xdouble (*xdotc_k) (BLASLONG, xdouble *, BLASLONG, xdouble *, BLASLONG);
int (*xqrot_k) (BLASLONG, xdouble *, BLASLONG, xdouble *, BLASLONG, xdouble, xdouble);
int (*xaxpy_k) (BLASLONG, BLASLONG, BLASLONG, xdouble, xdouble, xdouble *, BLASLONG, xdouble *, BLASLONG, xdouble *, BLASLONG);
@@ -888,7 +888,7 @@ BLASLONG (*ixamin_k)(BLASLONG, xdouble *, BLASLONG);
int (*sgeadd_k) (BLASLONG, BLASLONG, float, float *, BLASLONG, float, float *, BLASLONG);
int (*dgeadd_k) (BLASLONG, BLASLONG, double, double *, BLASLONG, double, double *, BLASLONG);
int (*cgeadd_k) (BLASLONG, BLASLONG, float, float, float *, BLASLONG, float, float, float *, BLASLONG);
int (*zgeadd_k) (BLASLONG, BLASLONG, float, double, double *, BLASLONG, double, double, double *, BLASLONG);
int (*zgeadd_k) (BLASLONG, BLASLONG, double, double, double *, BLASLONG, double, double, double *, BLASLONG);
} gotoblas_t;
+9 -8
View File
@@ -47,14 +47,15 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
* - large enough to support all architectures and kernel
* Chosing a too small SIZE will lead to a stack smashing.
*/
#define STACK_ALLOC(SIZE, TYPE, BUFFER) \
/* make it volatile because some function (ex: dgemv_n.S) */ \
/* do not restore all register */ \
volatile int stack_alloc_size = SIZE; \
if(stack_alloc_size > MAX_STACK_ALLOC / sizeof(TYPE)) \
stack_alloc_size = 0; \
STACK_ALLOC_PROTECT_SET \
TYPE stack_buffer[stack_alloc_size] __attribute__((aligned(0x20))); \
#define STACK_ALLOC(SIZE, TYPE, BUFFER) \
/* make it volatile because some function (ex: dgemv_n.S) */ \
/* do not restore all register */ \
volatile int stack_alloc_size = SIZE; \
if (stack_alloc_size > MAX_STACK_ALLOC / sizeof(TYPE)) stack_alloc_size = 0; \
STACK_ALLOC_PROTECT_SET \
/* Avoid declaring an array of length 0 */ \
TYPE stack_buffer[stack_alloc_size ? stack_alloc_size : 1] \
__attribute__((aligned(0x20))); \
BUFFER = stack_alloc_size ? stack_buffer : (TYPE *)blas_memory_alloc(1);
#else
//Original OpenBLAS/GotoBLAS codes.
+8 -2
View File
@@ -178,7 +178,13 @@ static __inline int blas_quickdivide(unsigned int x, unsigned int y){
result = x/y;
return result;
#else
#if (MAX_CPU_NUMBER > 64)
if ( y > 64) {
result = x/y;
return result;
}
#endif
y = blas_quick_divide_table[y];
__asm__ __volatile__ ("mull %0" :"=d" (result) :"a"(x), "0" (y));
@@ -327,7 +333,7 @@ REALNAME:
#endif
#endif
#if defined(OS_LINUX) || defined(OS_FREEBSD) || defined(OS_NETBSD) || defined(__ELF__)
#if defined(OS_LINUX) || defined(OS_FREEBSD) || defined(OS_NETBSD) || defined(OS_OPENBSD) || defined(__ELF__)
#define PROLOGUE \
.text; \
.align 16; \
+15 -3
View File
@@ -60,8 +60,13 @@
#endif
*/
#define MB
#define WMB
#ifdef __GNUC__
#define MB do { __asm__ __volatile__("": : :"memory"); } while (0)
#define WMB do { __asm__ __volatile__("": : :"memory"); } while (0)
#else
#define MB do {} while (0)
#define WMB do {} while (0)
#endif
static void __inline blas_lock(volatile BLASULONG *address){
@@ -196,6 +201,13 @@ static __inline int blas_quickdivide(unsigned int x, unsigned int y){
if (y <= 1) return x;
#if (MAX_CPU_NUMBER > 64)
if (y > 64) {
result = x / y;
return result;
}
#endif
y = blas_quick_divide_table[y];
__asm__ __volatile__ ("mull %0" :"=d" (result) :"a"(x), "0" (y));
@@ -403,7 +415,7 @@ REALNAME:
#define EPILOGUE .end
#endif
#if defined(OS_LINUX) || defined(OS_FREEBSD) || defined(OS_NETBSD) || defined(__ELF__) || defined(C_PGI)
#if defined(OS_LINUX) || defined(OS_FREEBSD) || defined(OS_NETBSD) || defined(OS_OPENBSD) || defined(OS_DRAGONFLY) || defined(__ELF__) || defined(C_PGI)
#define PROLOGUE \
.text; \
.align 512; \
+3
View File
@@ -115,6 +115,7 @@
#define CORE_STEAMROLLER 25
#define CORE_EXCAVATOR 26
#define CORE_ZEN 27
#define CORE_SKYLAKEX 28
#define HAVE_SSE (1 << 0)
#define HAVE_SSE2 (1 << 1)
@@ -137,6 +138,7 @@
#define HAVE_AVX (1 << 18)
#define HAVE_FMA4 (1 << 19)
#define HAVE_FMA3 (1 << 20)
#define HAVE_AVX512VL (1 << 21)
#define CACHE_INFO_L1_I 1
#define CACHE_INFO_L1_D 2
@@ -211,5 +213,6 @@ typedef struct {
#define CPUTYPE_STEAMROLLER 49
#define CPUTYPE_EXCAVATOR 50
#define CPUTYPE_ZEN 51
#define CPUTYPE_SKYLAKEX 52
#endif
+5 -4
View File
@@ -114,13 +114,14 @@ int detect(void)
fclose(infile);
if(cpu_part != NULL && cpu_implementer != NULL) {
if (strstr(cpu_part, "0xd07") && strstr(cpu_implementer, "0x41"))
return CPU_CORTEXA57;
if (strstr(cpu_implementer, "0x41") &&
(strstr(cpu_part, "0xd07") || strstr(cpu_part,"0xd08") || strstr(cpu_part,"0xd03") ))
return CPU_CORTEXA57; //or compatible A53, A72
else if (strstr(cpu_part, "0x516") && strstr(cpu_implementer, "0x42"))
return CPU_VULCAN;
else if (strstr(cpu_part, "0x0a1") && strstr(cpu_implementer, "0x43"))
return CPU_THUNDERX;
else if (strstr(cpu_part, "0xFFF") && strstr(cpu_implementer, "0x43")) /* TODO */
else if (strstr(cpu_part, "0x0af") && strstr(cpu_implementer, "0x43"))
return CPU_THUNDERX2T99;
}
@@ -142,7 +143,7 @@ int detect(void)
if(p != NULL)
{
if (strstr(p, "AArch64"))
if ((strstr(p, "AArch64")) || (strstr(p, "8")))
{
return CPU_ARMV8;
+20 -38
View File
@@ -72,10 +72,12 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#define CPU_UNKNOWN 0
#define CPU_P5600 1
#define CPU_1004K 2
static char *cpuname[] = {
"UNKOWN",
"P5600"
"P5600",
"1004K"
};
int detect(void){
@@ -90,7 +92,7 @@ int detect(void){
if (!strncmp("cpu", buffer, 3)){
p = strchr(buffer, ':') + 2;
#if 0
fprintf(stderr, "%s\n", p);
fprintf(stderr, "%s \n", p);
#endif
break;
}
@@ -99,43 +101,13 @@ int detect(void){
fclose(infile);
if(p != NULL){
if (strstr(p, "Loongson-3A")){
return CPU_LOONGSON3A;
}else if(strstr(p, "Loongson-3B")){
return CPU_LOONGSON3B;
}else if (strstr(p, "Loongson-3")){
infile = fopen("/proc/cpuinfo", "r");
p = (char *)NULL;
while (fgets(buffer, sizeof(buffer), infile)){
if (!strncmp("system type", buffer, 11)){
p = strchr(buffer, ':') + 2;
break;
}
}
fclose(infile);
if (strstr(p, "loongson3a"))
return CPU_LOONGSON3A;
}else{
if (strstr(p, "5600")) {
return CPU_P5600;
} else if (strstr(p, "1004K")) {
return CPU_1004K;
} else
return CPU_UNKNOWN;
}
}
//Check model name for Loongson3
infile = fopen("/proc/cpuinfo", "r");
p = (char *)NULL;
while (fgets(buffer, sizeof(buffer), infile)){
if (!strncmp("model name", buffer, 10)){
p = strchr(buffer, ':') + 2;
break;
}
}
fclose(infile);
if(p != NULL){
if (strstr(p, "Loongson-3A")){
return CPU_LOONGSON3A;
}else if(strstr(p, "Loongson-3B")){
return CPU_LOONGSON3B;
}
}
#endif
return CPU_UNKNOWN;
}
@@ -149,7 +121,7 @@ void get_architecture(void){
}
void get_subarchitecture(void){
if(detect()==CPU_P5600){
if(detect()==CPU_P5600|| detect()==CPU_1004K){
printf("P5600");
}else{
printf("UNKNOWN");
@@ -170,6 +142,14 @@ void get_cpuconfig(void){
printf("#define DTB_DEFAULT_ENTRIES 64\n");
printf("#define DTB_SIZE 4096\n");
printf("#define L2_ASSOCIATIVE 8\n");
} else if (detect()==CPU_1004K) {
printf("#define MIPS1004K\n");
printf("#define L1_DATA_SIZE 32768\n");
printf("#define L1_DATA_LINESIZE 32\n");
printf("#define L2_SIZE 26144\n");
printf("#define DTB_DEFAULT_ENTRIES 8\n");
printf("#define DTB_SIZE 4096\n");
printf("#define L2_ASSOCIATIVE 4\n");
}else{
printf("#define UNKNOWN\n");
}
@@ -178,6 +158,8 @@ void get_cpuconfig(void){
void get_libname(void){
if(detect()==CPU_P5600) {
printf("p5600\n");
} else if (detect()==CPU_1004K) {
printf("1004K\n");
}else{
printf("mips\n");
}
+16 -1
View File
@@ -76,6 +76,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#define CPU_LOONGSON3B 3
#define CPU_I6400 4
#define CPU_P6600 5
#define CPU_I6500 6
static char *cpuname[] = {
"UNKOWN",
@@ -83,7 +84,8 @@ static char *cpuname[] = {
"LOONGSON3A",
"LOONGSON3B",
"I6400",
"P6600"
"P6600",
"I6500"
};
int detect(void){
@@ -165,6 +167,8 @@ void get_subarchitecture(void){
printf("I6400");
}else if(detect()==CPU_P6600){
printf("P6600");
}else if(detect()==CPU_I6500){
printf("I6500");
}else{
printf("SICORTEX");
}
@@ -211,6 +215,15 @@ void get_cpuconfig(void){
printf("#define DTB_DEFAULT_ENTRIES 64\n");
printf("#define DTB_SIZE 4096\n");
printf("#define L2_ASSOCIATIVE 8\n");
}else if(detect()==CPU_I6500){
printf("#define I6500\n");
printf("#define L1_DATA_SIZE 65536\n");
printf("#define L1_DATA_LINESIZE 32\n");
printf("#define L2_SIZE 1048576\n");
printf("#define L2_LINESIZE 32\n");
printf("#define DTB_DEFAULT_ENTRIES 64\n");
printf("#define DTB_SIZE 4096\n");
printf("#define L2_ASSOCIATIVE 8\n");
}else{
printf("#define SICORTEX\n");
printf("#define L1_DATA_SIZE 32768\n");
@@ -232,6 +245,8 @@ void get_libname(void){
printf("i6400\n");
}else if(detect()==CPU_P6600) {
printf("p6600\n");
}else if(detect()==CPU_I6500) {
printf("i6500\n");
}else{
printf("mips64\n");
}
+6
View File
@@ -49,6 +49,7 @@ void get_subdirname(void){
}
void get_cpuconfig(void){
printf("#define SPARC\n");
printf("#define V9\n");
printf("#define DTB_DEFAULT_ENTRIES 32\n");
}
@@ -56,3 +57,8 @@ void get_cpuconfig(void){
void get_libname(void){
printf("v9\n");
}
char *get_corename(void){
return "sparc";
}
+156 -15
View File
@@ -50,6 +50,8 @@
#ifdef NO_AVX
#define CPUTYPE_HASWELL CPUTYPE_NEHALEM
#define CORE_HASWELL CORE_NEHALEM
#define CPUTYPE_SKYLAKEX CPUTYPE_NEHALEM
#define CORE_SKYLAKEX CORE_NEHALEM
#define CPUTYPE_SANDYBRIDGE CPUTYPE_NEHALEM
#define CORE_SANDYBRIDGE CORE_NEHALEM
#define CPUTYPE_BULLDOZER CPUTYPE_BARCELONA
@@ -71,12 +73,23 @@ void cpuid(int op, int *eax, int *ebx, int *ecx, int *edx)
*edx = cpuInfo[3];
}
void cpuid_count(int op, int count, int *eax, int *ebx, int *ecx, int *edx)
{
int cpuInfo[4] = {-1};
__cpuidex(cpuInfo, op, count);
*eax = cpuInfo[0];
*ebx = cpuInfo[1];
*ecx = cpuInfo[2];
*edx = cpuInfo[3];
}
#else
#ifndef CPUIDEMU
#if defined(__APPLE__) && defined(__i386__)
void cpuid(int op, int *eax, int *ebx, int *ecx, int *edx);
void cpuid_count(int op, int count, int *eax, int *ebx, int *ecx, int *edx);
#else
static C_INLINE void cpuid(int op, int *eax, int *ebx, int *ecx, int *edx){
#if defined(__i386__) && defined(__PIC__)
@@ -90,6 +103,19 @@ static C_INLINE void cpuid(int op, int *eax, int *ebx, int *ecx, int *edx){
("cpuid": "=a" (*eax), "=b" (*ebx), "=c" (*ecx), "=d" (*edx) : "a" (op) : "cc");
#endif
}
static C_INLINE void cpuid_count(int op, int count ,int *eax, int *ebx, int *ecx, int *edx){
#if defined(__i386__) && defined(__PIC__)
__asm__ __volatile__
("mov %%ebx, %%edi;"
"cpuid;"
"xchgl %%ebx, %%edi;"
: "=a" (*eax), "=D" (*ebx), "=c" (*ecx), "=d" (*edx) : "0" (op), "2" (count) : "cc");
#else
__asm__ __volatile__
("cpuid": "=a" (*eax), "=b" (*ebx), "=c" (*ecx), "=d" (*edx) : "0" (op), "2" (count) : "cc");
#endif
}
#endif
#else
@@ -133,6 +159,10 @@ void cpuid(unsigned int op, unsigned int *eax, unsigned int *ebx, unsigned int *
*edx = idlist[current].d;
}
void cpuid_count (unsigned int op, unsigned int count, unsigned int *eax, unsigned int *ebx, unsigned int *ecx, unsigned int *edx) {
return cpuid (op, eax, ebx, ecx, edx);
}
#endif
#endif // _MSC_VER
@@ -312,9 +342,9 @@ int get_cacheinfo(int type, cache_info_t *cacheinfo){
cpuid(0, &cpuid_level, &ebx, &ecx, &edx);
if (cpuid_level > 1) {
int numcalls =0 ;
cpuid(2, &eax, &ebx, &ecx, &edx);
numcalls = BITMASK(eax, 0, 0xff); //FIXME some systems may require repeated calls to read all entries
info[ 0] = BITMASK(eax, 8, 0xff);
info[ 1] = BITMASK(eax, 16, 0xff);
info[ 2] = BITMASK(eax, 24, 0xff);
@@ -335,7 +365,6 @@ int get_cacheinfo(int type, cache_info_t *cacheinfo){
info[14] = BITMASK(edx, 24, 0xff);
for (i = 0; i < 15; i++){
switch (info[i]){
/* This table is from http://www.sandpile.org/ia32/cpuid.htm */
@@ -637,12 +666,13 @@ int get_cacheinfo(int type, cache_info_t *cacheinfo){
LD1.linesize = 64;
break;
case 0x63 :
DTB.size = 2048;
DTB.associative = 4;
DTB.linesize = 32;
LDTB.size = 4096;
LDTB.associative= 4;
LDTB.linesize = 32;
DTB.size = 2048;
DTB.associative = 4;
DTB.linesize = 32;
LDTB.size = 4096;
LDTB.associative= 4;
LDTB.linesize = 32;
break;
case 0x66 :
LD1.size = 8;
LD1.associative = 4;
@@ -675,12 +705,13 @@ int get_cacheinfo(int type, cache_info_t *cacheinfo){
LC1.associative = 8;
break;
case 0x76 :
ITB.size = 2048;
ITB.associative = 0;
ITB.linesize = 8;
LITB.size = 4096;
LITB.associative= 0;
LITB.linesize = 8;
ITB.size = 2048;
ITB.associative = 0;
ITB.linesize = 8;
LITB.size = 4096;
LITB.associative= 0;
LITB.linesize = 8;
break;
case 0x77 :
LC1.size = 16;
LC1.associative = 4;
@@ -891,6 +922,67 @@ int get_cacheinfo(int type, cache_info_t *cacheinfo){
}
if (get_vendor() == VENDOR_INTEL) {
if(LD1.size<=0 || LC1.size<=0){
//If we didn't detect L1 correctly before,
int count;
for (count=0;count <4;count++) {
cpuid_count(4, count, &eax, &ebx, &ecx, &edx);
switch (eax &0x1f) {
case 0:
continue;
case 1:
case 3:
{
switch ((eax >>5) &0x07)
{
case 1:
{
// fprintf(stderr,"L1 data cache...\n");
int sets = ecx+1;
int lines = (ebx & 0x0fff) +1;
ebx>>=12;
int part = (ebx&0x03ff)+1;
ebx >>=10;
int assoc = (ebx&0x03ff)+1;
LD1.size = (assoc*part*lines*sets)/1024;
LD1.associative = assoc;
LD1.linesize= lines;
break;
}
default:
break;
}
break;
}
case 2:
{
switch ((eax >>5) &0x07)
{
case 1:
{
// fprintf(stderr,"L1 instruction cache...\n");
int sets = ecx+1;
int lines = (ebx & 0x0fff) +1;
ebx>>=12;
int part = (ebx&0x03ff)+1;
ebx >>=10;
int assoc = (ebx&0x03ff)+1;
LC1.size = (assoc*part*lines*sets)/1024;
LC1.associative = assoc;
LC1.linesize= lines;
break;
}
default:
break;
}
break;
}
default:
break;
}
}
}
cpuid(0x80000000, &cpuid_level, &ebx, &ecx, &edx);
if (cpuid_level >= 0x80000006) {
if(L2.size<=0){
@@ -1209,6 +1301,19 @@ int get_cpuname(void){
else
return CPUTYPE_NEHALEM;
case 5:
// Skylake X
#ifndef NO_AVX512
return CPUTYPE_SKYLAKEX;
#else
if(support_avx())
#ifndef NO_AVX2
return CPUTYPE_HASWELL;
#else
return CPUTYPE_SANDYBRIDGE;
#endif
else
return CPUTYPE_NEHALEM;
#endif
case 14:
// Skylake
if(support_avx())
@@ -1234,6 +1339,23 @@ int get_cpuname(void){
return CPUTYPE_NEHALEM;
}
break;
case 6:
switch (model) {
case 6: // Cannon Lake
#ifndef NO_AVX512
return CPUTYPE_SKYLAKEX;
#else
if(support_avx())
#ifndef NO_AVX2
return CPUTYPE_HASWELL;
#else
return CPUTYPE_SANDYBRIDGE;
#endif
else
return CPUTYPE_NEHALEM;
#endif
}
break;
case 9:
case 8:
switch (model) {
@@ -1466,6 +1588,7 @@ static char *cpuname[] = {
"STEAMROLLER",
"EXCAVATOR",
"ZEN",
"SKYLAKEX"
};
static char *lowercpuname[] = {
@@ -1520,6 +1643,7 @@ static char *lowercpuname[] = {
"steamroller",
"excavator",
"zen",
"skylakex"
};
static char *corename[] = {
@@ -1551,6 +1675,7 @@ static char *corename[] = {
"STEAMROLLER",
"EXCAVATOR",
"ZEN",
"SKYLAKEX"
};
static char *corename_lower[] = {
@@ -1582,6 +1707,7 @@ static char *corename_lower[] = {
"steamroller",
"excavator",
"zen",
"skylakex"
};
@@ -1687,6 +1813,8 @@ int get_coretype(void){
break;
case 3:
switch (model) {
case 7:
return CORE_ATOM;
case 10:
case 14:
if(support_avx())
@@ -1768,6 +1896,19 @@ int get_coretype(void){
else
return CORE_NEHALEM;
case 5:
// Skylake X
#ifndef NO_AVX512
return CORE_SKYLAKEX;
#else
if(support_avx())
#ifndef NO_AVX2
return CORE_HASWELL;
#else
return CORE_SANDYBRIDGE;
#endif
else
return CORE_NEHALEM;
#endif
case 14:
// Skylake
if(support_avx())
+8
View File
@@ -60,6 +60,14 @@ OS_FREEBSD
OS_NETBSD
#endif
#if defined(__OpenBSD__)
OS_OPENBSD
#endif
#if defined(__DragonFly__)
OS_DRAGONFLY
#endif
#if defined(__sun)
OS_SUNOS
#endif
+4 -3
View File
@@ -1,4 +1,5 @@
include_directories(${PROJECT_SOURCE_DIR})
include_directories(${PROJECT_BINARY_DIR})
enable_language(Fortran)
@@ -15,7 +16,7 @@ foreach(float_type ${FLOAT_TYPES})
add_executable(x${float_char}cblat1
c_${float_char}blat1.f
c_${float_char}blas1.c)
target_link_libraries(x${float_char}cblat1 ${OpenBLAS_LIBNAME}_static)
target_link_libraries(x${float_char}cblat1 ${OpenBLAS_LIBNAME})
add_test(NAME "x${float_char}cblat1"
COMMAND "${CMAKE_CURRENT_BINARY_DIR}/x${float_char}cblat1")
@@ -27,7 +28,7 @@ foreach(float_type ${FLOAT_TYPES})
auxiliary.c
c_xerbla.c
constant.c)
target_link_libraries(x${float_char}cblat2 ${OpenBLAS_LIBNAME}_static)
target_link_libraries(x${float_char}cblat2 ${OpenBLAS_LIBNAME})
add_test(NAME "x${float_char}cblat2"
COMMAND sh "${CMAKE_CURRENT_BINARY_DIR}/test_cblas_helper.sh" "${CMAKE_CURRENT_BINARY_DIR}/x${float_char}cblat2" "${PROJECT_SOURCE_DIR}/ctest/${float_char}in2")
@@ -39,7 +40,7 @@ foreach(float_type ${FLOAT_TYPES})
auxiliary.c
c_xerbla.c
constant.c)
target_link_libraries(x${float_char}cblat3 ${OpenBLAS_LIBNAME}_static)
target_link_libraries(x${float_char}cblat3 ${OpenBLAS_LIBNAME})
add_test(NAME "x${float_char}cblat3"
COMMAND sh "${CMAKE_CURRENT_BINARY_DIR}/test_cblas_helper.sh" "${CMAKE_CURRENT_BINARY_DIR}/x${float_char}cblat3" "${PROJECT_SOURCE_DIR}/ctest/${float_char}in3")
+7 -1
View File
@@ -102,7 +102,13 @@ clean ::
rm -f x*
FLDFLAGS = $(FFLAGS:-fPIC=) $(LDFLAGS)
CEXTRALIB =
ifeq ($(USE_OPENMP), 1)
ifeq ($(F_COMPILER), GFORTRAN)
ifeq ($(C_COMPILER), CLANG)
CEXTRALIB = -lomp
endif
endif
endif
# Single real
xscblat1: $(stestl1o) c_sblat1.o $(TOPDIR)/$(LIBNAME)
+3 -3
View File
@@ -9,7 +9,7 @@
#include "common.h"
#include "cblas_test.h"
void F77_caxpy(const int *N, const void *alpha, void *X,
void F77_caxpy(const int *N, OPENBLAS_CONST void *alpha, void *X,
const int *incX, void *Y, const int *incY)
{
cblas_caxpy(*N, alpha, X, *incX, Y, *incY);
@@ -58,13 +58,13 @@ void F77_cswap( const int *N, void *X, const int *incX,
return;
}
int F77_icamax(const int *N, const void *X, const int *incX)
int F77_icamax(const int *N, OPENBLAS_CONST void *X, const int *incX)
{
if (*N < 1 || *incX < 1) return(0);
return (cblas_icamax(*N, X, *incX)+1);
}
float F77_scnrm2(const int *N, const void *X, const int *incX)
float F77_scnrm2(const int *N, OPENBLAS_CONST void *X, const int *incX)
{
return cblas_scnrm2(*N, X, *incX);
}
+3 -3
View File
@@ -9,9 +9,9 @@
#include "cblas_test.h"
void F77_cgemv(int *order, char *transp, int *m, int *n,
const void *alpha,
CBLAS_TEST_COMPLEX *a, int *lda, const void *x, int *incx,
const void *beta, void *y, int *incy) {
OPENBLAS_CONST void *alpha,
CBLAS_TEST_COMPLEX *a, int *lda, OPENBLAS_CONST void *x, int *incx,
OPENBLAS_CONST void *beta, void *y, int *incy) {
CBLAS_TEST_COMPLEX *A;
int i,j,LDA;
+3 -3
View File
@@ -349,13 +349,13 @@
CALL CCHK3( SNAMES( ISNUM ), EPS, THRESH, NOUT, NTRA, TRACE,
$ REWI, FATAL, NIDIM, IDIM, NKB, KB, NINC, INC,
$ NMAX, INCMAX, A, AA, AS, Y, YY, YS, YT, G, Z,
$ 0 )
$ 0 )
END IF
IF (RORDER) THEN
CALL CCHK3( SNAMES( ISNUM ), EPS, THRESH, NOUT, NTRA, TRACE,
$ REWI, FATAL, NIDIM, IDIM, NKB, KB, NINC, INC,
$ NMAX, INCMAX, A, AA, AS, Y, YY, YS, YT, G, Z,
$ 1 )
$ 1 )
END IF
GO TO 200
* Test CGERC, 12, CGERU, 13.
@@ -2660,7 +2660,7 @@
50 CONTINUE
END IF
*
60 CONTINUE
C 60 CONTINUE
LCERES = .TRUE.
GO TO 80
70 CONTINUE
+26 -26
View File
@@ -329,13 +329,13 @@
CALL CCHK3(SNAMES( ISNUM ), EPS, THRESH, NOUT, NTRA, TRACE,
$ REWI, FATAL, NIDIM, IDIM, NALF, ALF, NMAX, AB,
$ AA, AS, AB( 1, NMAX + 1 ), BB, BS, CT, G, C,
$ 0 )
$ 0 )
END IF
IF (RORDER) THEN
CALL CCHK3(SNAMES( ISNUM ), EPS, THRESH, NOUT, NTRA, TRACE,
$ REWI, FATAL, NIDIM, IDIM, NALF, ALF, NMAX, AB,
$ AA, AS, AB( 1, NMAX + 1 ), BB, BS, CT, G, C,
$ 1 )
$ 1 )
END IF
GO TO 190
* Test CHERK, 06, CSYRK, 07.
@@ -357,13 +357,13 @@
CALL CCHK5(SNAMES( ISNUM ), EPS, THRESH, NOUT, NTRA, TRACE,
$ REWI, FATAL, NIDIM, IDIM, NALF, ALF, NBET, BET,
$ NMAX, AB, AA, AS, BB, BS, C, CC, CS, CT, G, W,
$ 0 )
$ 0 )
END IF
IF (RORDER) THEN
CALL CCHK5(SNAMES( ISNUM ), EPS, THRESH, NOUT, NTRA, TRACE,
$ REWI, FATAL, NIDIM, IDIM, NALF, ALF, NBET, BET,
$ NMAX, AB, AA, AS, BB, BS, C, CC, CS, CT, G, W,
$ 1 )
$ 1 )
END IF
GO TO 190
*
@@ -707,9 +707,9 @@
9998 FORMAT(' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9996 FORMAT( ' ******* ', A12,' FAILED ON CALL NUMBER:' )
9995 FORMAT( 1X, I6, ': ', A12,'(''', A1, ''',''', A1, ''',',
$ 3( I3, ',' ), '(', F4.1, ',', F4.1, '), A,', I3, ', B,', I3,
$ ',(', F4.1, ',', F4.1, '), C,', I3, ').' )
C 9995 FORMAT( 1X, I6, ': ', A12,'(''', A1, ''',''', A1, ''',',
C $ 3( I3, ',' ), '(', F4.1, ',', F4.1, '), A,', I3, ', B,', I3,
C $ ',(', F4.1, ',', F4.1, '), C,', I3, ').' )
9994 FORMAT(' ******* FATAL ERROR - ERROR-EXIT TAKEN ON VALID CALL *',
$ '******' )
*
@@ -1033,9 +1033,9 @@
9998 FORMAT(' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9996 FORMAT( ' ******* ', A12,' FAILED ON CALL NUMBER:' )
9995 FORMAT(1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
$ '(', F4.1, ',', F4.1, '), A,', I3, ', B,', I3, ',(', F4.1,
$ ',', F4.1, '), C,', I3, ') .' )
C 9995 FORMAT(1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
C $ '(', F4.1, ',', F4.1, '), A,', I3, ', B,', I3, ',(', F4.1,
C $ ',', F4.1, '), C,', I3, ') .' )
9994 FORMAT(' ******* FATAL ERROR - ERROR-EXIT TAKEN ON VALID CALL *',
$ '******' )
*
@@ -1385,9 +1385,9 @@
9998 FORMAT(' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9996 FORMAT(' ******* ', A12,' FAILED ON CALL NUMBER:' )
9995 FORMAT(1X, I6, ': ', A12,'(', 4( '''', A1, ''',' ), 2( I3, ',' ),
$ '(', F4.1, ',', F4.1, '), A,', I3, ', B,', I3, ') ',
$ ' .' )
C 9995 FORMAT(1X, I6, ': ', A12,'(', 4( '''', A1, ''',' ), 2( I3, ',' ),
C $ '(', F4.1, ',', F4.1, '), A,', I3, ', B,', I3, ') ',
C $ ' .' )
9994 FORMAT(' ******* FATAL ERROR - ERROR-EXIT TAKEN ON VALID CALL *',
$ '******' )
*
@@ -1768,12 +1768,12 @@
$ 'ANGED INCORRECTLY *******' )
9996 FORMAT( ' ******* ', A12,' FAILED ON CALL NUMBER:' )
9995 FORMAT( ' THESE ARE THE RESULTS FOR COLUMN ', I3 )
9994 FORMAT(1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
$ F4.1, ', A,', I3, ',', F4.1, ', C,', I3, ') ',
$ ' .' )
9993 FORMAT(1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
$ '(', F4.1, ',', F4.1, ') , A,', I3, ',(', F4.1, ',', F4.1,
$ '), C,', I3, ') .' )
C 9994 FORMAT(1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
C $ F4.1, ', A,', I3, ',', F4.1, ', C,', I3, ') ',
C $ ' .' )
C 9993 FORMAT(1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
C $ '(', F4.1, ',', F4.1, ') , A,', I3, ',(', F4.1, ',', F4.1,
C $ '), C,', I3, ') .' )
9992 FORMAT(' ******* FATAL ERROR - ERROR-EXIT TAKEN ON VALID CALL *',
$ '******' )
*
@@ -2221,12 +2221,12 @@
$ 'ANGED INCORRECTLY *******' )
9996 FORMAT( ' ******* ', A12,' FAILED ON CALL NUMBER:' )
9995 FORMAT( ' THESE ARE THE RESULTS FOR COLUMN ', I3 )
9994 FORMAT(1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
$ '(', F4.1, ',', F4.1, '), A,', I3, ', B,', I3, ',', F4.1,
$ ', C,', I3, ') .' )
9993 FORMAT(1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
$ '(', F4.1, ',', F4.1, '), A,', I3, ', B,', I3, ',(', F4.1,
$ ',', F4.1, '), C,', I3, ') .' )
C 9994 FORMAT(1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
C $ '(', F4.1, ',', F4.1, '), A,', I3, ', B,', I3, ',', F4.1,
C $ ', C,', I3, ') .' )
C 9993 FORMAT(1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
C $ '(', F4.1, ',', F4.1, '), A,', I3, ', B,', I3, ',(', F4.1,
C $ ',', F4.1, '), C,', I3, ') .' )
9992 FORMAT(' ******* FATAL ERROR - ERROR-EXIT TAKEN ON VALID CALL *',
$ '******' )
*
@@ -2702,7 +2702,7 @@
50 CONTINUE
END IF
*
60 CONTINUE
C 60 CONTINUE
LCERES = .TRUE.
GO TO 80
70 CONTINUE
+6 -6
View File
@@ -14,7 +14,7 @@ double F77_dasum(const int *N, double *X, const int *incX)
return cblas_dasum(*N, X, *incX);
}
void F77_daxpy(const int *N, const double *alpha, const double *X,
void F77_daxpy(const int *N, const double *alpha, OPENBLAS_CONST double *X,
const int *incX, double *Y, const int *incY)
{
cblas_daxpy(*N, *alpha, X, *incX, Y, *incY);
@@ -28,13 +28,13 @@ void F77_dcopy(const int *N, double *X, const int *incX,
return;
}
double F77_ddot(const int *N, const double *X, const int *incX,
const double *Y, const int *incY)
double F77_ddot(const int *N, OPENBLAS_CONST double *X, const int *incX,
OPENBLAS_CONST double *Y, const int *incY)
{
return cblas_ddot(*N, X, *incX, Y, *incY);
}
double F77_dnrm2(const int *N, const double *X, const int *incX)
double F77_dnrm2(const int *N, OPENBLAS_CONST double *X, const int *incX)
{
return cblas_dnrm2(*N, X, *incX);
}
@@ -72,12 +72,12 @@ double F77_dzasum(const int *N, void *X, const int *incX)
return cblas_dzasum(*N, X, *incX);
}
double F77_dznrm2(const int *N, const void *X, const int *incX)
double F77_dznrm2(const int *N, OPENBLAS_CONST void *X, const int *incX)
{
return cblas_dznrm2(*N, X, *incX);
}
int F77_idamax(const int *N, const double *X, const int *incX)
int F77_idamax(const int *N, OPENBLAS_CONST double *X, const int *incX)
{
if (*N < 1 || *incX < 1) return(0);
return (cblas_idamax(*N, X, *incX)+1);
+2 -2
View File
@@ -211,11 +211,11 @@
IF (ICASE.EQ.7) THEN
* .. DNRM2TEST ..
STEMP(1) = DTRUE1(NP1)
CALL STEST1(DNRM2TEST(N,SX,INCX),STEMP,STEMP,SFAC)
CALL STEST1(DNRM2TEST(N,SX,INCX),STEMP(1),STEMP,SFAC)
ELSE IF (ICASE.EQ.8) THEN
* .. DASUMTEST ..
STEMP(1) = DTRUE3(NP1)
CALL STEST1(DASUMTEST(N,SX,INCX),STEMP,STEMP,SFAC)
CALL STEST1(DASUMTEST(N,SX,INCX),STEMP(1),STEMP,SFAC)
ELSE IF (ICASE.EQ.9) THEN
* .. DSCALTEST ..
CALL DSCALTEST(N,SA((INCX-1)*5+NP1),SX,INCX)
+28 -28
View File
@@ -345,13 +345,13 @@
CALL DCHK3( SNAMES( ISNUM ), EPS, THRESH, NOUT, NTRA, TRACE,
$ REWI, FATAL, NIDIM, IDIM, NKB, KB, NINC, INC,
$ NMAX, INCMAX, A, AA, AS, Y, YY, YS, YT, G, Z,
$ 0 )
$ 0 )
END IF
IF (RORDER) THEN
CALL DCHK3( SNAMES( ISNUM ), EPS, THRESH, NOUT, NTRA, TRACE,
$ REWI, FATAL, NIDIM, IDIM, NKB, KB, NINC, INC,
$ NMAX, INCMAX, A, AA, AS, Y, YY, YS, YT, G, Z,
$ 1 )
$ 1 )
END IF
GO TO 200
* Test DGER, 12.
@@ -797,9 +797,9 @@
$ ' (', I6, ' CALL', 'S)' )
9998 FORMAT( ' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
$ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
$ ' - SUSPECT *******' )
C 9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
C $ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
C $ ' - SUSPECT *******' )
9996 FORMAT( ' ******* ',A12, ' FAILED ON CALL NUMBER:' )
9995 FORMAT( 1X, I6, ': ',A12, '(', A14, ',', 4( I3, ',' ), F4.1,
$ ', A,', I3, ',',/ 10x,'X,', I2, ',', F4.1, ', Y,',
@@ -1004,7 +1004,7 @@
$ REWIND NTRA
CALL CDSBMV( IORDER, UPLO, N, K, ALPHA,
$ AA, LDA, XX, INCX, BETA, YY,
$ INCY )
$ INCY )
ELSE IF( PACKED )THEN
IF( TRACE )
$ WRITE( NTRA, FMT = 9995 )NC, SNAME,
@@ -1156,9 +1156,9 @@
$ ' (', I6, ' CALL', 'S)' )
9998 FORMAT( ' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
$ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
$ ' - SUSPECT *******' )
C 9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
C $ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
C $ ' - SUSPECT *******' )
9996 FORMAT( ' ******* ',A12, ' FAILED ON CALL NUMBER:' )
9995 FORMAT( 1X, I6, ': ',A12, '(', A14, ',', I3, ',', F4.1, ', AP',
$ ', X,', I2, ',', F4.1, ', Y,', I2, ') .' )
@@ -1191,7 +1191,7 @@
* .. Scalar Arguments ..
DOUBLE PRECISION EPS, THRESH
INTEGER INCMAX, NIDIM, NINC, NKB, NMAX, NOUT, NTRA,
$ IORDER
$ IORDER
LOGICAL FATAL, REWI, TRACE
CHARACTER*12 SNAME
* .. Array Arguments ..
@@ -1216,7 +1216,7 @@
EXTERNAL LDE, LDERES
* .. External Subroutines ..
EXTERNAL DMAKE, DMVCH, CDTBMV, CDTBSV, CDTPMV,
$ CDTPSV, CDTRMV, CDTRSV
$ CDTPSV, CDTRMV, CDTRSV
* .. Intrinsic Functions ..
INTRINSIC ABS, MAX
* .. Scalars in Common ..
@@ -1544,9 +1544,9 @@
$ ' (', I6, ' CALL', 'S)' )
9998 FORMAT( ' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
$ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
$ ' - SUSPECT *******' )
C 9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
C $ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
C $ ' - SUSPECT *******' )
9996 FORMAT( ' ******* ',A12, ' FAILED ON CALL NUMBER:' )
9995 FORMAT( 1X, I6, ': ',A12, '(', 3( A14,',' ),/ 10x, I3, ', AP, ',
$ 'X,', I2, ') .' )
@@ -1579,7 +1579,7 @@
* .. Scalar Arguments ..
DOUBLE PRECISION EPS, THRESH
INTEGER INCMAX, NALF, NIDIM, NINC, NMAX, NOUT, NTRA,
$ IORDER
$ IORDER
LOGICAL FATAL, REWI, TRACE
CHARACTER*12 SNAME
* .. Array Arguments ..
@@ -1819,9 +1819,9 @@
$ ' (', I6, ' CALL', 'S)' )
9998 FORMAT( ' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
$ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
$ ' - SUSPECT *******' )
C 9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
C $ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
C $ ' - SUSPECT *******' )
9996 FORMAT( ' ******* ',A12, ' FAILED ON CALL NUMBER:' )
9995 FORMAT( ' THESE ARE THE RESULTS FOR COLUMN ', I3 )
9994 FORMAT( 1X, I6, ': ',A12, '(', 2( I3, ',' ), F4.1, ', X,', I2,
@@ -1851,7 +1851,7 @@
* .. Scalar Arguments ..
DOUBLE PRECISION EPS, THRESH
INTEGER INCMAX, NALF, NIDIM, NINC, NMAX, NOUT, NTRA,
$ IORDER
$ IORDER
LOGICAL FATAL, REWI, TRACE
CHARACTER*12 SNAME
* .. Array Arguments ..
@@ -1973,7 +1973,7 @@
IF( REWI )
$ REWIND NTRA
CALL CDSYR( IORDER, UPLO, N, ALPHA, XX, INCX,
$ AA, LDA )
$ AA, LDA )
ELSE IF( PACKED )THEN
IF( TRACE )
$ WRITE( NTRA, FMT = 9994 )NC, SNAME, CUPLO, N,
@@ -2113,9 +2113,9 @@
$ ' (', I6, ' CALL', 'S)' )
9998 FORMAT( ' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
$ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
$ ' - SUSPECT *******' )
C 9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
C $ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
C $ ' - SUSPECT *******' )
9996 FORMAT( ' ******* ',A12, ' FAILED ON CALL NUMBER:' )
9995 FORMAT( ' THESE ARE THE RESULTS FOR COLUMN ', I3 )
9994 FORMAT( 1X, I6, ': ',A12, '(', A14, ',', I3, ',', F4.1, ', X,',
@@ -2147,7 +2147,7 @@
* .. Scalar Arguments ..
DOUBLE PRECISION EPS, THRESH
INTEGER INCMAX, NALF, NIDIM, NINC, NMAX, NOUT, NTRA,
$ IORDER
$ IORDER
LOGICAL FATAL, REWI, TRACE
CHARACTER*12 SNAME
* .. Array Arguments ..
@@ -2445,9 +2445,9 @@
$ ' (', I6, ' CALL', 'S)' )
9998 FORMAT( ' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
$ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
$ ' - SUSPECT *******' )
C 9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
C $ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
C $ ' - SUSPECT *******' )
9996 FORMAT( ' ******* ',A12, ' FAILED ON CALL NUMBER:' )
9995 FORMAT( ' THESE ARE THE RESULTS FOR COLUMN ', I3 )
9994 FORMAT( 1X, I6, ': ',A12, '(', A14, ',', I3, ',', F4.1, ', X,',
@@ -2833,7 +2833,7 @@
50 CONTINUE
END IF
*
60 CONTINUE
C 60 CONTINUE
LDERES = .TRUE.
GO TO 80
70 CONTINUE
+25 -25
View File
@@ -56,7 +56,7 @@
* .. Local Scalars ..
DOUBLE PRECISION EPS, ERR, THRESH
INTEGER I, ISNUM, J, N, NALF, NBET, NIDIM, NTRA,
$ LAYOUT
$ LAYOUT
LOGICAL FATAL, LTESTT, REWI, SAME, SFATAL, TRACE,
$ TSTERR, CORDER, RORDER
CHARACTER*1 TRANSA, TRANSB
@@ -78,7 +78,7 @@
EXTERNAL DDIFF, LDE
* .. External Subroutines ..
EXTERNAL DCHK1, DCHK2, DCHK3, DCHK4, DCHK5, CD3CHKE,
$ DMMCH
$ DMMCH
* .. Intrinsic Functions ..
INTRINSIC MAX, MIN
* .. Scalars in Common ..
@@ -323,13 +323,13 @@
CALL DCHK3( SNAMES( ISNUM ), EPS, THRESH, NOUT, NTRA, TRACE,
$ REWI, FATAL, NIDIM, IDIM, NALF, ALF, NMAX, AB,
$ AA, AS, AB( 1, NMAX + 1 ), BB, BS, CT, G, C,
$ 0 )
$ 0 )
END IF
IF (RORDER) THEN
CALL DCHK3( SNAMES( ISNUM ), EPS, THRESH, NOUT, NTRA, TRACE,
$ REWI, FATAL, NIDIM, IDIM, NALF, ALF, NMAX, AB,
$ AA, AS, AB( 1, NMAX + 1 ), BB, BS, CT, G, C,
$ 1 )
$ 1 )
END IF
GO TO 190
* Test DSYRK, 05.
@@ -351,13 +351,13 @@
CALL DCHK5( SNAMES( ISNUM ), EPS, THRESH, NOUT, NTRA, TRACE,
$ REWI, FATAL, NIDIM, IDIM, NALF, ALF, NBET, BET,
$ NMAX, AB, AA, AS, BB, BS, C, CC, CS, CT, G, W,
$ 0 )
$ 0 )
END IF
IF (RORDER) THEN
CALL DCHK5( SNAMES( ISNUM ), EPS, THRESH, NOUT, NTRA, TRACE,
$ REWI, FATAL, NIDIM, IDIM, NALF, ALF, NBET, BET,
$ NMAX, AB, AA, AS, BB, BS, C, CC, CS, CT, G, W,
$ 1 )
$ 1 )
END IF
GO TO 190
*
@@ -588,7 +588,7 @@
$ REWIND NTRA
CALL CDGEMM( IORDER, TRANSA, TRANSB, M, N,
$ K, ALPHA, AA, LDA, BB, LDB,
$ BETA, CC, LDC )
$ BETA, CC, LDC )
*
* Check if error-exit was taken incorrectly.
*
@@ -694,9 +694,9 @@
9998 FORMAT( ' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9996 FORMAT( ' ******* ', A12,' FAILED ON CALL NUMBER:' )
9995 FORMAT( 1X, I6, ': ', A12,'(''', A1, ''',''', A1, ''',',
$ 3( I3, ',' ), F4.1, ', A,', I3, ', B,', I3, ',', F4.1, ', ',
$ 'C,', I3, ').' )
C 9995 FORMAT( 1X, I6, ': ', A12,'(''', A1, ''',''', A1, ''',',
C $ 3( I3, ',' ), F4.1, ', A,', I3, ', B,', I3, ',', F4.1, ', ',
C $ 'C,', I3, ').' )
9994 FORMAT( ' ******* FATAL ERROR - ERROR-EXIT TAKEN ON VALID CALL *',
$ '******' )
*
@@ -1007,9 +1007,9 @@
9998 FORMAT( ' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9996 FORMAT( ' ******* ', A12,' FAILED ON CALL NUMBER:' )
9995 FORMAT( 1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
$ F4.1, ', A,', I3, ', B,', I3, ',', F4.1, ', C,', I3, ') ',
$ ' .' )
C 9995 FORMAT( 1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
C $ F4.1, ', A,', I3, ', B,', I3, ',', F4.1, ', C,', I3, ') ',
C $ ' .' )
9994 FORMAT( ' ******* FATAL ERROR - ERROR-EXIT TAKEN ON VALID CALL *',
$ '******' )
*
@@ -1201,7 +1201,7 @@
$ REWIND NTRA
CALL CDTRMM( IORDER, SIDE, UPLO, TRANSA,
$ DIAG, M, N, ALPHA, AA, LDA,
$ BB, LDB )
$ BB, LDB )
ELSE IF( SNAME( 10: 11 ).EQ.'sm' )THEN
IF( TRACE )
$ CALL DPRCN3( NTRA, NC, SNAME, IORDER,
@@ -1211,7 +1211,7 @@
$ REWIND NTRA
CALL CDTRSM( IORDER, SIDE, UPLO, TRANSA,
$ DIAG, M, N, ALPHA, AA, LDA,
$ BB, LDB )
$ BB, LDB )
END IF
*
* Check if error-exit was taken incorrectly.
@@ -1355,8 +1355,8 @@
9998 FORMAT( ' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9996 FORMAT( ' ******* ', A12,' FAILED ON CALL NUMBER:' )
9995 FORMAT( 1X, I6, ': ', A12,'(', 4( '''', A1, ''',' ), 2( I3, ',' ),
$ F4.1, ', A,', I3, ', B,', I3, ') .' )
C 9995 FORMAT( 1X, I6, ': ', A12,'(', 4( '''', A1, ''',' ), 2( I3, ',' ),
C $ F4.1, ', A,', I3, ', B,', I3, ') .' )
9994 FORMAT( ' ******* FATAL ERROR - ERROR-EXIT TAKEN ON VALID CALL *',
$ '******' )
*
@@ -1681,8 +1681,8 @@
$ 'ANGED INCORRECTLY *******' )
9996 FORMAT( ' ******* ', A12,' FAILED ON CALL NUMBER:' )
9995 FORMAT( ' THESE ARE THE RESULTS FOR COLUMN ', I3 )
9994 FORMAT( 1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
$ F4.1, ', A,', I3, ',', F4.1, ', C,', I3, ') .' )
C 9994 FORMAT( 1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
C $ F4.1, ', A,', I3, ',', F4.1, ', C,', I3, ') .' )
9993 FORMAT( ' ******* FATAL ERROR - ERROR-EXIT TAKEN ON VALID CALL *',
$ '******' )
*
@@ -1726,7 +1726,7 @@
SUBROUTINE DCHK5( SNAME, EPS, THRESH, NOUT, NTRA, TRACE, REWI,
$ FATAL, NIDIM, IDIM, NALF, ALF, NBET, BET, NMAX,
$ AB, AA, AS, BB, BS, C, CC, CS, CT, G, W,
$ IORDER )
$ IORDER )
*
* Tests DSYR2K.
*
@@ -1888,7 +1888,7 @@
$ REWIND NTRA
CALL CDSYR2K( IORDER, UPLO, TRANS, N, K,
$ ALPHA, AA, LDA, BB, LDB, BETA,
$ CC, LDC )
$ CC, LDC )
*
* Check if error-exit was taken incorrectly.
*
@@ -2037,9 +2037,9 @@
$ 'ANGED INCORRECTLY *******' )
9996 FORMAT( ' ******* ', A12,' FAILED ON CALL NUMBER:' )
9995 FORMAT( ' THESE ARE THE RESULTS FOR COLUMN ', I3 )
9994 FORMAT( 1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
$ F4.1, ', A,', I3, ', B,', I3, ',', F4.1, ', C,', I3, ') ',
$ ' .' )
C 9994 FORMAT( 1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
C $ F4.1, ', A,', I3, ', B,', I3, ',', F4.1, ', C,', I3, ') ',
C $ ' .' )
9993 FORMAT( ' ******* FATAL ERROR - ERROR-EXIT TAKEN ON VALID CALL *',
$ '******' )
*
@@ -2399,7 +2399,7 @@
50 CONTINUE
END IF
*
60 CONTINUE
C 60 CONTINUE
LDERES = .TRUE.
GO TO 80
70 CONTINUE
+7 -7
View File
@@ -14,7 +14,7 @@ float F77_sasum(blasint *N, float *X, blasint *incX)
return cblas_sasum(*N, X, *incX);
}
void F77_saxpy(blasint *N, const float *alpha, const float *X,
void F77_saxpy(blasint *N, const float *alpha, OPENBLAS_CONST float *X,
blasint *incX, float *Y, blasint *incY)
{
cblas_saxpy(*N, *alpha, X, *incX, Y, *incY);
@@ -26,25 +26,25 @@ float F77_scasum(blasint *N, float *X, blasint *incX)
return cblas_scasum(*N, X, *incX);
}
float F77_scnrm2(blasint *N, const float *X, blasint *incX)
float F77_scnrm2(blasint *N, OPENBLAS_CONST float *X, blasint *incX)
{
return cblas_scnrm2(*N, X, *incX);
}
void F77_scopy(blasint *N, const float *X, blasint *incX,
void F77_scopy(blasint *N, OPENBLAS_CONST float *X, blasint *incX,
float *Y, blasint *incY)
{
cblas_scopy(*N, X, *incX, Y, *incY);
return;
}
float F77_sdot(blasint *N, const float *X, blasint *incX,
const float *Y, blasint *incY)
float F77_sdot(blasint *N, OPENBLAS_CONST float *X, blasint *incX,
OPENBLAS_CONST float *Y, blasint *incY)
{
return cblas_sdot(*N, X, *incX, Y, *incY);
}
float F77_snrm2(blasint *N, const float *X, blasint *incX)
float F77_snrm2(blasint *N, OPENBLAS_CONST float *X, blasint *incX)
{
return cblas_snrm2(*N, X, *incX);
}
@@ -76,7 +76,7 @@ void F77_sswap( blasint *N, float *X, blasint *incX,
return;
}
int F77_isamax(blasint *N, const float *X, blasint *incX)
int F77_isamax(blasint *N, OPENBLAS_CONST float *X, blasint *incX)
{
if (*N < 1 || *incX < 1) return(0);
return (cblas_isamax(*N, X, *incX)+1);
+2 -2
View File
@@ -211,11 +211,11 @@
IF (ICASE.EQ.7) THEN
* .. SNRM2TEST ..
STEMP(1) = DTRUE1(NP1)
CALL STEST1(SNRM2TEST(N,SX,INCX),STEMP,STEMP,SFAC)
CALL STEST1(SNRM2TEST(N,SX,INCX),STEMP(1),STEMP,SFAC)
ELSE IF (ICASE.EQ.8) THEN
* .. SASUMTEST ..
STEMP(1) = DTRUE3(NP1)
CALL STEST1(SASUMTEST(N,SX,INCX),STEMP,STEMP,SFAC)
CALL STEST1(SASUMTEST(N,SX,INCX),STEMP(1),STEMP,SFAC)
ELSE IF (ICASE.EQ.9) THEN
* .. SSCALTEST ..
CALL SSCALTEST(N,SA((INCX-1)*5+NP1),SX,INCX)
+28 -28
View File
@@ -345,13 +345,13 @@
CALL SCHK3( SNAMES( ISNUM ), EPS, THRESH, NOUT, NTRA, TRACE,
$ REWI, FATAL, NIDIM, IDIM, NKB, KB, NINC, INC,
$ NMAX, INCMAX, A, AA, AS, Y, YY, YS, YT, G, Z,
$ 0 )
$ 0 )
END IF
IF (RORDER) THEN
CALL SCHK3( SNAMES( ISNUM ), EPS, THRESH, NOUT, NTRA, TRACE,
$ REWI, FATAL, NIDIM, IDIM, NKB, KB, NINC, INC,
$ NMAX, INCMAX, A, AA, AS, Y, YY, YS, YT, G, Z,
$ 1 )
$ 1 )
END IF
GO TO 200
* Test SGER, 12.
@@ -797,9 +797,9 @@
$ ' (', I6, ' CALL', 'S)' )
9998 FORMAT( ' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
$ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
$ ' - SUSPECT *******' )
C 9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
C $ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
C $ ' - SUSPECT *******' )
9996 FORMAT( ' ******* ',A12, ' FAILED ON CALL NUMBER:' )
9995 FORMAT( 1X, I6, ': ',A12, '(', A14, ',', 4( I3, ',' ), F4.1,
$ ', A,', I3, ',',/ 10x, 'X,', I2, ',', F4.1, ', Y,',
@@ -1004,7 +1004,7 @@
$ REWIND NTRA
CALL CSSBMV( IORDER, UPLO, N, K, ALPHA,
$ AA, LDA, XX, INCX, BETA, YY,
$ INCY )
$ INCY )
ELSE IF( PACKED )THEN
IF( TRACE )
$ WRITE( NTRA, FMT = 9995 )NC, SNAME,
@@ -1156,9 +1156,9 @@
$ ' (', I6, ' CALL', 'S)' )
9998 FORMAT( ' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
$ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
$ ' - SUSPECT *******' )
C 9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
C $ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
C $ ' - SUSPECT *******' )
9996 FORMAT( ' ******* ',A12, ' FAILED ON CALL NUMBER:' )
9995 FORMAT( 1X, I6, ': ',A12, '(', A14, ',', I3, ',', F4.1, ', AP',
$ ', X,', I2, ',', F4.1, ', Y,', I2, ') .' )
@@ -1191,7 +1191,7 @@
* .. Scalar Arguments ..
REAL EPS, THRESH
INTEGER INCMAX, NIDIM, NINC, NKB, NMAX, NOUT, NTRA,
$ IORDER
$ IORDER
LOGICAL FATAL, REWI, TRACE
CHARACTER*12 SNAME
* .. Array Arguments ..
@@ -1216,7 +1216,7 @@
EXTERNAL LSE, LSERES
* .. External Subroutines ..
EXTERNAL SMAKE, SMVCH, CSTBMV, CSTBSV, CSTPMV,
$ CSTPSV, CSTRMV, CSTRSV
$ CSTPSV, CSTRMV, CSTRSV
* .. Intrinsic Functions ..
INTRINSIC ABS, MAX
* .. Scalars in Common ..
@@ -1544,9 +1544,9 @@
$ ' (', I6, ' CALL', 'S)' )
9998 FORMAT( ' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
$ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
$ ' - SUSPECT *******' )
C 9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
C $ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
C $ ' - SUSPECT *******' )
9996 FORMAT( ' ******* ',A12, ' FAILED ON CALL NUMBER:' )
9995 FORMAT( 1X, I6, ': ',A12, '(', 3( A14,',' ),/ 10x, I3, ', AP, ',
$ 'X,', I2, ') .' )
@@ -1579,7 +1579,7 @@
* .. Scalar Arguments ..
REAL EPS, THRESH
INTEGER INCMAX, NALF, NIDIM, NINC, NMAX, NOUT, NTRA,
$ IORDER
$ IORDER
LOGICAL FATAL, REWI, TRACE
CHARACTER*12 SNAME
* .. Array Arguments ..
@@ -1819,9 +1819,9 @@
$ ' (', I6, ' CALL', 'S)' )
9998 FORMAT( ' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
$ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
$ ' - SUSPECT *******' )
C 9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
C $ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
C $ ' - SUSPECT *******' )
9996 FORMAT( ' ******* ',A12, ' FAILED ON CALL NUMBER:' )
9995 FORMAT( ' THESE ARE THE RESULTS FOR COLUMN ', I3 )
9994 FORMAT( 1X, I6, ': ',A12, '(', 2( I3, ',' ), F4.1, ', X,', I2,
@@ -1851,7 +1851,7 @@
* .. Scalar Arguments ..
REAL EPS, THRESH
INTEGER INCMAX, NALF, NIDIM, NINC, NMAX, NOUT, NTRA,
$ IORDER
$ IORDER
LOGICAL FATAL, REWI, TRACE
CHARACTER*12 SNAME
* .. Array Arguments ..
@@ -1973,7 +1973,7 @@
IF( REWI )
$ REWIND NTRA
CALL CSSYR( IORDER, UPLO, N, ALPHA, XX, INCX,
$ AA, LDA )
$ AA, LDA )
ELSE IF( PACKED )THEN
IF( TRACE )
$ WRITE( NTRA, FMT = 9994 )NC, SNAME, CUPLO, N,
@@ -2113,9 +2113,9 @@
$ ' (', I6, ' CALL', 'S)' )
9998 FORMAT( ' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
$ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
$ ' - SUSPECT *******' )
C 9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
C $ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
C $ ' - SUSPECT *******' )
9996 FORMAT( ' ******* ',A12, ' FAILED ON CALL NUMBER:' )
9995 FORMAT( ' THESE ARE THE RESULTS FOR COLUMN ', I3 )
9994 FORMAT( 1X, I6, ': ',A12, '(', A14, ',', I3, ',', F4.1, ', X,',
@@ -2147,7 +2147,7 @@
* .. Scalar Arguments ..
REAL EPS, THRESH
INTEGER INCMAX, NALF, NIDIM, NINC, NMAX, NOUT, NTRA,
$ IORDER
$ IORDER
LOGICAL FATAL, REWI, TRACE
CHARACTER*12 SNAME
* .. Array Arguments ..
@@ -2445,9 +2445,9 @@
$ ' (', I6, ' CALL', 'S)' )
9998 FORMAT( ' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
$ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
$ ' - SUSPECT *******' )
C 9997 FORMAT( ' ',A12, ' COMPLETED THE COMPUTATIONAL TESTS (', I6, ' C',
C $ 'ALLS)', /' ******* BUT WITH MAXIMUM TEST RATIO', F8.2,
C $ ' - SUSPECT *******' )
9996 FORMAT( ' ******* ',A12, ' FAILED ON CALL NUMBER:' )
9995 FORMAT( ' THESE ARE THE RESULTS FOR COLUMN ', I3 )
9994 FORMAT( 1X, I6, ': ',A12, '(', A14, ',', I3, ',', F4.1, ', X,',
@@ -2833,7 +2833,7 @@
50 CONTINUE
END IF
*
60 CONTINUE
C 60 CONTINUE
LSERES = .TRUE.
GO TO 80
70 CONTINUE
+14 -14
View File
@@ -694,9 +694,9 @@
9998 FORMAT( ' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9996 FORMAT( ' ******* ', A12,' FAILED ON CALL NUMBER:' )
9995 FORMAT( 1X, I6, ': ', A12,'(''', A1, ''',''', A1, ''',',
$ 3( I3, ',' ), F4.1, ', A,', I3, ', B,', I3, ',', F4.1, ', ',
$ 'C,', I3, ').' )
C 9995 FORMAT( 1X, I6, ': ', A12,'(''', A1, ''',''', A1, ''',',
C $ 3( I3, ',' ), F4.1, ', A,', I3, ', B,', I3, ',', F4.1, ', ',
C $ 'C,', I3, ').' )
9994 FORMAT( ' ******* FATAL ERROR - ERROR-EXIT TAKEN ON VALID CALL *',
$ '******' )
*
@@ -1011,9 +1011,9 @@
9998 FORMAT( ' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9996 FORMAT( ' ******* ', A12,' FAILED ON CALL NUMBER:' )
9995 FORMAT( 1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
$ F4.1, ', A,', I3, ', B,', I3, ',', F4.1, ', C,', I3, ') ',
$ ' .' )
C 9995 FORMAT( 1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
C $ F4.1, ', A,', I3, ', B,', I3, ',', F4.1, ', C,', I3, ') ',
C $ ' .' )
9994 FORMAT( ' ******* FATAL ERROR - ERROR-EXIT TAKEN ON VALID CALL *',
$ '******' )
*
@@ -1359,8 +1359,8 @@
9998 FORMAT( ' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9996 FORMAT( ' ******* ', A12,' FAILED ON CALL NUMBER:' )
9995 FORMAT( 1X, I6, ': ', A12,'(', 4( '''', A1, ''',' ), 2( I3, ',' ),
$ F4.1, ', A,', I3, ', B,', I3, ') .' )
C 9995 FORMAT( 1X, I6, ': ', A12,'(', 4( '''', A1, ''',' ), 2( I3, ',' ),
C $ F4.1, ', A,', I3, ', B,', I3, ') .' )
9994 FORMAT( ' ******* FATAL ERROR - ERROR-EXIT TAKEN ON VALID CALL *',
$ '******' )
*
@@ -1686,8 +1686,8 @@
$ 'ANGED INCORRECTLY *******' )
9996 FORMAT( ' ******* ', A12,' FAILED ON CALL NUMBER:' )
9995 FORMAT( ' THESE ARE THE RESULTS FOR COLUMN ', I3 )
9994 FORMAT( 1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
$ F4.1, ', A,', I3, ',', F4.1, ', C,', I3, ') .' )
C 9994 FORMAT( 1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
C $ F4.1, ', A,', I3, ',', F4.1, ', C,', I3, ') .' )
9993 FORMAT( ' ******* FATAL ERROR - ERROR-EXIT TAKEN ON VALID CALL *',
$ '******' )
*
@@ -2041,9 +2041,9 @@
$ 'ANGED INCORRECTLY *******' )
9996 FORMAT( ' ******* ', A12,' FAILED ON CALL NUMBER:' )
9995 FORMAT( ' THESE ARE THE RESULTS FOR COLUMN ', I3 )
9994 FORMAT( 1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
$ F4.1, ', A,', I3, ', B,', I3, ',', F4.1, ', C,', I3, ') ',
$ ' .' )
C 9994 FORMAT( 1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
C $ F4.1, ', A,', I3, ', B,', I3, ',', F4.1, ', C,', I3, ') ',
C $ ' .' )
9993 FORMAT( ' ******* FATAL ERROR - ERROR-EXIT TAKEN ON VALID CALL *',
$ '******' )
*
@@ -2403,7 +2403,7 @@
50 CONTINUE
END IF
*
60 CONTINUE
C 60 CONTINUE
LSERES = .TRUE.
GO TO 80
70 CONTINUE
+1 -1
View File
@@ -131,7 +131,7 @@ void F77_xerbla(char *srname, void *vinfo)
int BLASFUNC(xerbla)(char *name, blasint *info, blasint length) {
F77_xerbla(name, info);
return 0;
};
+5 -5
View File
@@ -9,7 +9,7 @@
#include "common.h"
#include "cblas_test.h"
void F77_zaxpy(const int *N, const void *alpha, void *X,
void F77_zaxpy(const int *N, OPENBLAS_CONST void *alpha, void *X,
const int *incX, void *Y, const int *incY)
{
cblas_zaxpy(*N, alpha, X, *incX, Y, *incY);
@@ -23,8 +23,8 @@ void F77_zcopy(const int *N, void *X, const int *incX,
return;
}
void F77_zdotc(const int *N, const void *X, const int *incX,
const void *Y, const int *incY,void *dotc)
void F77_zdotc(const int *N, OPENBLAS_CONST void *X, const int *incX,
OPENBLAS_CONST void *Y, const int *incY,void *dotc)
{
cblas_zdotc_sub(*N, X, *incX, Y, *incY, dotc);
return;
@@ -58,13 +58,13 @@ void F77_zswap( const int *N, void *X, const int *incX,
return;
}
int F77_izamax(const int *N, const void *X, const int *incX)
int F77_izamax(const int *N, OPENBLAS_CONST void *X, const int *incX)
{
if (*N < 1 || *incX < 1) return(0);
return(cblas_izamax(*N, X, *incX)+1);
}
double F77_dznrm2(const int *N, const void *X, const int *incX)
double F77_dznrm2(const int *N, OPENBLAS_CONST void *X, const int *incX)
{
return cblas_dznrm2(*N, X, *incX);
}
+3 -3
View File
@@ -9,9 +9,9 @@
#include "cblas_test.h"
void F77_zgemv(int *order, char *transp, int *m, int *n,
const void *alpha,
CBLAS_TEST_ZOMPLEX *a, int *lda, const void *x, int *incx,
const void *beta, void *y, int *incy) {
OPENBLAS_CONST void *alpha,
CBLAS_TEST_ZOMPLEX *a, int *lda, OPENBLAS_CONST void *x, int *incx,
OPENBLAS_CONST void *beta, void *y, int *incy) {
CBLAS_TEST_ZOMPLEX *A;
int i,j,LDA;
+2 -2
View File
@@ -349,13 +349,13 @@
CALL ZCHK3( SNAMES( ISNUM ), EPS, THRESH, NOUT, NTRA, TRACE,
$ REWI, FATAL, NIDIM, IDIM, NKB, KB, NINC, INC,
$ NMAX, INCMAX, A, AA, AS, Y, YY, YS, YT, G, Z,
$ 0 )
$ 0 )
END IF
IF (RORDER) THEN
CALL ZCHK3( SNAMES( ISNUM ), EPS, THRESH, NOUT, NTRA, TRACE,
$ REWI, FATAL, NIDIM, IDIM, NKB, KB, NINC, INC,
$ NMAX, INCMAX, A, AA, AS, Y, YY, YS, YT, G, Z,
$ 1 )
$ 1 )
END IF
GO TO 200
* Test ZGERC, 12, ZGERU, 13.
+26 -26
View File
@@ -330,13 +330,13 @@
CALL ZCHK3(SNAMES( ISNUM ), EPS, THRESH, NOUT, NTRA, TRACE,
$ REWI, FATAL, NIDIM, IDIM, NALF, ALF, NMAX, AB,
$ AA, AS, AB( 1, NMAX + 1 ), BB, BS, CT, G, C,
$ 0 )
$ 0 )
END IF
IF (RORDER) THEN
CALL ZCHK3(SNAMES( ISNUM ), EPS, THRESH, NOUT, NTRA, TRACE,
$ REWI, FATAL, NIDIM, IDIM, NALF, ALF, NMAX, AB,
$ AA, AS, AB( 1, NMAX + 1 ), BB, BS, CT, G, C,
$ 1 )
$ 1 )
END IF
GO TO 190
* Test ZHERK, 06, ZSYRK, 07.
@@ -358,13 +358,13 @@
CALL ZCHK5(SNAMES( ISNUM ), EPS, THRESH, NOUT, NTRA, TRACE,
$ REWI, FATAL, NIDIM, IDIM, NALF, ALF, NBET, BET,
$ NMAX, AB, AA, AS, BB, BS, C, CC, CS, CT, G, W,
$ 0 )
$ 0 )
END IF
IF (RORDER) THEN
CALL ZCHK5(SNAMES( ISNUM ), EPS, THRESH, NOUT, NTRA, TRACE,
$ REWI, FATAL, NIDIM, IDIM, NALF, ALF, NBET, BET,
$ NMAX, AB, AA, AS, BB, BS, C, CC, CS, CT, G, W,
$ 1 )
$ 1 )
END IF
GO TO 190
*
@@ -708,9 +708,9 @@
9998 FORMAT(' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9996 FORMAT( ' ******* ', A12,' FAILED ON CALL NUMBER:' )
9995 FORMAT( 1X, I6, ': ', A12,'(''', A1, ''',''', A1, ''',',
$ 3( I3, ',' ), '(', F4.1, ',', F4.1, '), A,', I3, ', B,', I3,
$ ',(', F4.1, ',', F4.1, '), C,', I3, ').' )
C 9995 FORMAT( 1X, I6, ': ', A12,'(''', A1, ''',''', A1, ''',',
C $ 3( I3, ',' ), '(', F4.1, ',', F4.1, '), A,', I3, ', B,', I3,
C $ ',(', F4.1, ',', F4.1, '), C,', I3, ').' )
9994 FORMAT(' ******* FATAL ERROR - ERROR-EXIT TAKEN ON VALID CALL *',
$ '******' )
*
@@ -1034,9 +1034,9 @@
9998 FORMAT(' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9996 FORMAT( ' ******* ', A12,' FAILED ON CALL NUMBER:' )
9995 FORMAT(1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
$ '(', F4.1, ',', F4.1, '), A,', I3, ', B,', I3, ',(', F4.1,
$ ',', F4.1, '), C,', I3, ') .' )
C 9995 FORMAT(1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
C $ '(', F4.1, ',', F4.1, '), A,', I3, ', B,', I3, ',(', F4.1,
C $ ',', F4.1, '), C,', I3, ') .' )
9994 FORMAT(' ******* FATAL ERROR - ERROR-EXIT TAKEN ON VALID CALL *',
$ '******' )
*
@@ -1386,9 +1386,9 @@
9998 FORMAT(' ******* FATAL ERROR - PARAMETER NUMBER ', I2, ' WAS CH',
$ 'ANGED INCORRECTLY *******' )
9996 FORMAT(' ******* ', A12,' FAILED ON CALL NUMBER:' )
9995 FORMAT(1X, I6, ': ', A12,'(', 4( '''', A1, ''',' ), 2( I3, ',' ),
$ '(', F4.1, ',', F4.1, '), A,', I3, ', B,', I3, ') ',
$ ' .' )
C 9995 FORMAT(1X, I6, ': ', A12,'(', 4( '''', A1, ''',' ), 2( I3, ',' ),
C $ '(', F4.1, ',', F4.1, '), A,', I3, ', B,', I3, ') ',
C $ ' .' )
9994 FORMAT(' ******* FATAL ERROR - ERROR-EXIT TAKEN ON VALID CALL *',
$ '******' )
*
@@ -1769,12 +1769,12 @@
$ 'ANGED INCORRECTLY *******' )
9996 FORMAT( ' ******* ', A12,' FAILED ON CALL NUMBER:' )
9995 FORMAT( ' THESE ARE THE RESULTS FOR COLUMN ', I3 )
9994 FORMAT(1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
$ F4.1, ', A,', I3, ',', F4.1, ', C,', I3, ') ',
$ ' .' )
9993 FORMAT(1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
$ '(', F4.1, ',', F4.1, ') , A,', I3, ',(', F4.1, ',', F4.1,
$ '), C,', I3, ') .' )
C 9994 FORMAT(1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
C $ F4.1, ', A,', I3, ',', F4.1, ', C,', I3, ') ',
C $ ' .' )
C 9993 FORMAT(1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
C $ '(', F4.1, ',', F4.1, ') , A,', I3, ',(', F4.1, ',', F4.1,
C $ '), C,', I3, ') .' )
9992 FORMAT(' ******* FATAL ERROR - ERROR-EXIT TAKEN ON VALID CALL *',
$ '******' )
*
@@ -2222,12 +2222,12 @@
$ 'ANGED INCORRECTLY *******' )
9996 FORMAT( ' ******* ', A12,' FAILED ON CALL NUMBER:' )
9995 FORMAT( ' THESE ARE THE RESULTS FOR COLUMN ', I3 )
9994 FORMAT(1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
$ '(', F4.1, ',', F4.1, '), A,', I3, ', B,', I3, ',', F4.1,
$ ', C,', I3, ') .' )
9993 FORMAT(1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
$ '(', F4.1, ',', F4.1, '), A,', I3, ', B,', I3, ',(', F4.1,
$ ',', F4.1, '), C,', I3, ') .' )
C 9994 FORMAT(1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
C $ '(', F4.1, ',', F4.1, '), A,', I3, ', B,', I3, ',', F4.1,
C $ ', C,', I3, ') .' )
C 9993 FORMAT(1X, I6, ': ', A12,'(', 2( '''', A1, ''',' ), 2( I3, ',' ),
C $ '(', F4.1, ',', F4.1, '), A,', I3, ', B,', I3, ',(', F4.1,
C $ ',', F4.1, '), C,', I3, ') .' )
9992 FORMAT(' ******* FATAL ERROR - ERROR-EXIT TAKEN ON VALID CALL *',
$ '******' )
*
@@ -2706,7 +2706,7 @@
50 CONTINUE
END IF
*
60 CONTINUE
C 60 CONTINUE
LZERES = .TRUE.
GO TO 80
70 CONTINUE
+5 -4
View File
@@ -1,5 +1,6 @@
include_directories(${PROJECT_SOURCE_DIR})
include_directories(${PROJECT_BINARY_DIR})
# sources that need to be compiled twice, once with no flags and once with LOWER
set(UL_SOURCES
@@ -72,7 +73,7 @@ GenerateNamedObjects("zgbmv_k.c" "TRANS;CONJ;XCONJ" "gbmv_d" false "" "" "" 2)
# special defines for complex
foreach (float_type ${FLOAT_TYPES})
if (SMP)
if (USE_THREAD)
GenerateNamedObjects("gemv_thread.c" "" "gemv_thread_n" false "" "" false ${float_type})
GenerateNamedObjects("gemv_thread.c" "TRANSA" "gemv_thread_t" false "" "" false ${float_type})
@@ -106,7 +107,7 @@ foreach (float_type ${FLOAT_TYPES})
GenerateNamedObjects("z${ulvm_source}" "LOWER;HEMVREV" "${op_name}_M" false "" "" false ${float_type})
endforeach()
if (SMP)
if (USE_THREAD)
GenerateNamedObjects("gemv_thread.c" "CONJ" "gemv_thread_r" false "" "" false ${float_type})
GenerateNamedObjects("gemv_thread.c" "CONJ;TRANSA" "gemv_thread_c" false "" "" false ${float_type})
@@ -185,7 +186,7 @@ foreach (float_type ${FLOAT_TYPES})
GenerateCombinationObjects("${l_source}" "UNIT" "N" "TRANSA" 0 "${op_name}_TU" false ${float_type})
endforeach ()
if (SMP)
if (USE_THREAD)
GenerateNamedObjects("ger_thread.c" "" "" false "" "" false ${float_type})
foreach(nu_smp_source ${NU_SMP_SOURCES})
string(REGEX MATCH "[a-z]+_[a-z]+" op_name ${nu_smp_source})
@@ -196,7 +197,7 @@ foreach (float_type ${FLOAT_TYPES})
endif ()
endforeach ()
if (SMP)
if (USE_THREAD)
GenerateCombinationObjects("${UL_SMP_SOURCES}" "LOWER" "U" "" 2)
endif ()
+2 -2
View File
@@ -62,13 +62,13 @@ void CNAME(BLASLONG m, BLASLONG n, BLASLONG ku, BLASLONG kl, FLOAT alpha,
if (incy != 1) {
Y = bufferY;
bufferX = (FLOAT *)(((BLASLONG)bufferY + M * sizeof(FLOAT) + 4095) & ~4095);
gemvbuffer = bufferX;
// gemvbuffer = bufferX;
COPY_K(M, y, incy, Y, 1);
}
if (incx != 1) {
X = bufferX;
gemvbuffer = (FLOAT *)(((BLASLONG)bufferX + N * sizeof(FLOAT) + 4095) & ~4095);
// gemvbuffer = (FLOAT *)(((BLASLONG)bufferX + N * sizeof(FLOAT) + 4095) & ~4095);
COPY_K(N, x, incx, X, 1);
}
+3 -1
View File
@@ -96,7 +96,7 @@ static int gbmv_kernel(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n, F
COPY_K(args -> m, x, incx, buffer, 1);
x = buffer;
buffer += ((COMPSIZE * args -> m + 1023) & ~1023);
// buffer += ((COMPSIZE * args -> m + 1023) & ~1023);
}
#endif
@@ -230,8 +230,10 @@ int CNAME(BLASLONG m, BLASLONG n, BLASLONG ku, BLASLONG kl, FLOAT *alpha, FLOAT
#ifndef TRANSA
range_m[num_cpu] = num_cpu * ((m + 15) & ~15);
if (range_m[num_cpu] > m * num_cpu) range_m[num_cpu] = m * num_cpu;
#else
range_m[num_cpu] = num_cpu * ((n + 15) & ~15);
if (range_m[num_cpu] > n * num_cpu) range_m[num_cpu] = n * num_cpu;
#endif
queue[num_cpu].mode = mode;
+2 -2
View File
@@ -55,13 +55,13 @@ int CNAME(BLASLONG n, BLASLONG k, FLOAT alpha,
if (incy != 1) {
Y = bufferY;
bufferX = (FLOAT *)(((BLASLONG)bufferY + n * sizeof(FLOAT) + 4095) & ~4095);
sbmvbuffer = bufferX;
// sbmvbuffer = bufferX;
COPY_K(n, y, incy, Y, 1);
}
if (incx != 1) {
X = bufferX;
sbmvbuffer = (FLOAT *)(((BLASLONG)bufferX + n * sizeof(FLOAT) + 4095) & ~4095);
// sbmvbuffer = (FLOAT *)(((BLASLONG)bufferX + n * sizeof(FLOAT) + 4095) & ~4095);
COPY_K(n, x, incx, X, 1);
}
+4 -1
View File
@@ -91,7 +91,7 @@ static int sbmv_kernel(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n, F
COPY_K(n, x, incx, buffer, 1);
x = buffer;
buffer += ((COMPSIZE * n + 1023) & ~1023);
// buffer += ((COMPSIZE * n + 1023) & ~1023);
}
SCAL_K(n, 0, 0, ZERO,
@@ -246,6 +246,7 @@ int CNAME(BLASLONG n, BLASLONG k, FLOAT *alpha, FLOAT *a, BLASLONG lda, FLOAT *x
range_m[MAX_CPU_NUMBER - num_cpu - 1] = range_m[MAX_CPU_NUMBER - num_cpu] - width;
range_n[num_cpu] = num_cpu * (((n + 15) & ~15) + 16);
if (range_n[num_cpu] > n * num_cpu) range_n[num_cpu] = n * num_cpu;
queue[num_cpu].mode = mode;
queue[num_cpu].routine = sbmv_kernel;
@@ -285,6 +286,7 @@ int CNAME(BLASLONG n, BLASLONG k, FLOAT *alpha, FLOAT *a, BLASLONG lda, FLOAT *x
range_m[num_cpu + 1] = range_m[num_cpu] + width;
range_n[num_cpu] = num_cpu * (((n + 15) & ~15) + 16);
if (range_n[num_cpu] > n * num_cpu) range_n[num_cpu] = n * num_cpu;
queue[num_cpu].mode = mode;
queue[num_cpu].routine = sbmv_kernel;
@@ -316,6 +318,7 @@ int CNAME(BLASLONG n, BLASLONG k, FLOAT *alpha, FLOAT *a, BLASLONG lda, FLOAT *x
range_m[num_cpu + 1] = range_m[num_cpu] + width;
range_n[num_cpu] = num_cpu * ((n + 15) & ~15);
if (range_n[num_cpu] > n * num_cpu) range_n[num_cpu] = n * num_cpu;
queue[num_cpu].mode = mode;
queue[num_cpu].routine = sbmv_kernel;
+2 -2
View File
@@ -53,13 +53,13 @@ int CNAME(BLASLONG m, FLOAT alpha, FLOAT *a,
if (incy != 1) {
Y = bufferY;
bufferX = (FLOAT *)(((BLASLONG)bufferY + m * sizeof(FLOAT) + 4095) & ~4095);
gemvbuffer = bufferX;
// gemvbuffer = bufferX;
COPY_K(m, y, incy, Y, 1);
}
if (incx != 1) {
X = bufferX;
gemvbuffer = (FLOAT *)(((BLASLONG)bufferX + m * sizeof(FLOAT) + 4095) & ~4095);
// gemvbuffer = (FLOAT *)(((BLASLONG)bufferX + m * sizeof(FLOAT) + 4095) & ~4095);
COPY_K(m, x, incx, X, 1);
}
+2
View File
@@ -246,6 +246,7 @@ int CNAME(BLASLONG m, FLOAT *alpha, FLOAT *a, FLOAT *x, BLASLONG incx, FLOAT *y,
range_m[MAX_CPU_NUMBER - num_cpu - 1] = range_m[MAX_CPU_NUMBER - num_cpu] - width;
range_n[num_cpu] = num_cpu * (((m + 15) & ~15) + 16);
if (range_n[num_cpu] > m * num_cpu) range_n[num_cpu] = m * num_cpu;
queue[num_cpu].mode = mode;
queue[num_cpu].routine = spmv_kernel;
@@ -285,6 +286,7 @@ int CNAME(BLASLONG m, FLOAT *alpha, FLOAT *a, FLOAT *x, BLASLONG incx, FLOAT *y,
range_m[num_cpu + 1] = range_m[num_cpu] + width;
range_n[num_cpu] = num_cpu * (((m + 15) & ~15) + 16);
if (range_n[num_cpu] > m * num_cpu) range_n[num_cpu] = m * num_cpu;
queue[num_cpu].mode = mode;
queue[num_cpu].routine = spmv_kernel;
+3 -1
View File
@@ -177,7 +177,8 @@ int CNAME(BLASLONG m, FLOAT *alpha, FLOAT *a, BLASLONG lda, FLOAT *x, BLASLONG i
range_m[num_cpu + 1] = range_m[num_cpu] + width;
range_n[num_cpu] = num_cpu * (((m + 15) & ~15) + 16);
if (range_n[num_cpu] > m * num_cpu) range_n[num_cpu] = m * num_cpu;
queue[MAX_CPU_NUMBER - num_cpu - 1].mode = mode;
queue[MAX_CPU_NUMBER - num_cpu - 1].routine = symv_kernel;
queue[MAX_CPU_NUMBER - num_cpu - 1].args = &args;
@@ -225,6 +226,7 @@ int CNAME(BLASLONG m, FLOAT *alpha, FLOAT *a, BLASLONG lda, FLOAT *x, BLASLONG i
range_m[num_cpu + 1] = range_m[num_cpu] + width;
range_n[num_cpu] = num_cpu * (((m + 15) & ~15) + 16);
if (range_n[num_cpu] > m * num_cpu) range_n[num_cpu] = m * num_cpu;
queue[num_cpu].mode = mode;
queue[num_cpu].routine = symv_kernel;
+1 -1
View File
@@ -40,7 +40,7 @@
#include <ctype.h>
#include "common.h"
const static FLOAT dp1 = 1.;
// const static FLOAT dp1 = 1.;
int CNAME(BLASLONG n, BLASLONG k, FLOAT *a, BLASLONG lda, FLOAT *b, BLASLONG incb, void *buffer){
+1 -1
View File
@@ -40,7 +40,7 @@
#include <ctype.h>
#include "common.h"
const static FLOAT dp1 = 1.;
// const static FLOAT dp1 = 1.;
int CNAME(BLASLONG n, BLASLONG k, FLOAT *a, BLASLONG lda, FLOAT *b, BLASLONG incb, void *buffer){
+4 -1
View File
@@ -107,7 +107,7 @@ static int trmv_kernel(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n, F
COPY_K(args -> n, x, incx, buffer, 1);
x = buffer;
buffer += ((args -> n * COMPSIZE + 1023) & ~1023);
// buffer += ((args -> n * COMPSIZE + 1023) & ~1023);
}
if (range_n) y += *range_n * COMPSIZE;
@@ -288,6 +288,7 @@ int CNAME(BLASLONG n, BLASLONG k, FLOAT *a, BLASLONG lda, FLOAT *x, BLASLONG inc
range_m[MAX_CPU_NUMBER - num_cpu - 1] = range_m[MAX_CPU_NUMBER - num_cpu] - width;
range_n[num_cpu] = num_cpu * (((n + 15) & ~15) + 16);
if (range_n[num_cpu] > n * num_cpu) range_n[num_cpu] = n * num_cpu;
queue[num_cpu].mode = mode;
queue[num_cpu].routine = trmv_kernel;
@@ -327,6 +328,7 @@ int CNAME(BLASLONG n, BLASLONG k, FLOAT *a, BLASLONG lda, FLOAT *x, BLASLONG inc
range_m[num_cpu + 1] = range_m[num_cpu] + width;
range_n[num_cpu] = num_cpu * (((n + 15) & ~15) + 16);
if (range_n[num_cpu] > n * num_cpu) range_n[num_cpu] = n * num_cpu;
queue[num_cpu].mode = mode;
queue[num_cpu].routine = trmv_kernel;
@@ -356,6 +358,7 @@ int CNAME(BLASLONG n, BLASLONG k, FLOAT *a, BLASLONG lda, FLOAT *x, BLASLONG inc
range_m[num_cpu + 1] = range_m[num_cpu] + width;
range_n[num_cpu] = num_cpu * (((n + 15) & ~15) + 16);
if (range_n[num_cpu] > n * num_cpu) range_n[num_cpu] = n * num_cpu;
queue[num_cpu].mode = mode;
queue[num_cpu].routine = trmv_kernel;
+1 -1
View File
@@ -40,7 +40,7 @@
#include <ctype.h>
#include "common.h"
const static FLOAT dp1 = 1.;
// const static FLOAT dp1 = 1.;
int CNAME(BLASLONG n, BLASLONG k, FLOAT *a, BLASLONG lda, FLOAT *b, BLASLONG incb, void *buffer){
+1 -1
View File
@@ -40,7 +40,7 @@
#include <ctype.h>
#include "common.h"
const static FLOAT dp1 = 1.;
// const static FLOAT dp1 = 1.;
int CNAME(BLASLONG n, BLASLONG k, FLOAT *a, BLASLONG lda, FLOAT *b, BLASLONG incb, void *buffer){
+1 -1
View File
@@ -40,7 +40,7 @@
#include <ctype.h>
#include "common.h"
const static FLOAT dp1 = 1.;
// const static FLOAT dp1 = 1.;
int CNAME(BLASLONG m, FLOAT *a, FLOAT *b, BLASLONG incb, void *buffer){
+1 -1
View File
@@ -40,7 +40,7 @@
#include <ctype.h>
#include "common.h"
const static FLOAT dp1 = 1.;
// const static FLOAT dp1 = 1.;
int CNAME(BLASLONG m, FLOAT *a, FLOAT *b, BLASLONG incb, void *buffer){
+4 -6
View File
@@ -112,7 +112,7 @@ static int tpmv_kernel(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n, F
#endif
x = buffer;
buffer += ((COMPSIZE * args -> m + 1023) & ~1023);
// buffer += ((COMPSIZE * args -> m + 1023) & ~1023);
}
#ifndef TRANS
@@ -234,11 +234,7 @@ static int tpmv_kernel(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n, F
return 0;
}
#ifndef COMPLEX
int CNAME(BLASLONG m, FLOAT *a, FLOAT *x, BLASLONG incx, FLOAT *buffer, int nthreads){
#else
int CNAME(BLASLONG m, FLOAT *a, FLOAT *x, BLASLONG incx, FLOAT *buffer, int nthreads){
#endif
blas_arg_t args;
blas_queue_t queue[MAX_CPU_NUMBER];
@@ -307,7 +303,8 @@ int CNAME(BLASLONG m, FLOAT *a, FLOAT *x, BLASLONG incx, FLOAT *buffer, int nthr
range_m[MAX_CPU_NUMBER - num_cpu - 1] = range_m[MAX_CPU_NUMBER - num_cpu] - width;
range_n[num_cpu] = num_cpu * (((m + 15) & ~15) + 16);
if (range_n[num_cpu] > m * num_cpu) range_n[num_cpu] = m * num_cpu;
queue[num_cpu].mode = mode;
queue[num_cpu].routine = tpmv_kernel;
queue[num_cpu].args = &args;
@@ -346,6 +343,7 @@ int CNAME(BLASLONG m, FLOAT *a, FLOAT *x, BLASLONG incx, FLOAT *buffer, int nthr
range_m[num_cpu + 1] = range_m[num_cpu] + width;
range_n[num_cpu] = num_cpu * (((m + 15) & ~15) + 16);
if (range_n[num_cpu] > m * num_cpu) range_n[num_cpu] = m * num_cpu;
queue[num_cpu].mode = mode;
queue[num_cpu].routine = tpmv_kernel;
+6 -2
View File
@@ -54,12 +54,16 @@ int CNAME(BLASLONG m, FLOAT *a, BLASLONG lda, FLOAT *b, BLASLONG incb, FLOAT *bu
COPY_K(m, b, incb, buffer, 1);
}
for (is = 0; is < m; is += DTB_ENTRIES){
/*FIXME the GEMV unrolling performed here was found to be broken, see issue 1332 */
/* Multiplying DTB size by 100 is just a quick-and-dirty hack to disable it for now[B */
min_i = MIN(m - is, DTB_ENTRIES);
for (is = 0; is < m; is += DTB_ENTRIES * 100){
min_i = MIN(m - is, DTB_ENTRIES * 100);
#ifndef TRANSA
if (is > 0){
fprintf(stderr,"WARNING unrolling of the trmv_U loop may give wrong results\n");
GEMV_N(is, min_i, 0, dp1,
a + is * lda, lda,
B + is, 1,
+2
View File
@@ -346,6 +346,7 @@ int CNAME(BLASLONG m, FLOAT *a, BLASLONG lda, FLOAT *x, BLASLONG incx, FLOAT *bu
range_m[MAX_CPU_NUMBER - num_cpu - 1] = range_m[MAX_CPU_NUMBER - num_cpu] - width;
range_n[num_cpu] = num_cpu * (((m + 15) & ~15) + 16);
if (range_n[num_cpu] > m) range_n[num_cpu] = m;
queue[num_cpu].mode = mode;
queue[num_cpu].routine = trmv_kernel;
@@ -385,6 +386,7 @@ int CNAME(BLASLONG m, FLOAT *a, BLASLONG lda, FLOAT *x, BLASLONG incx, FLOAT *bu
range_m[num_cpu + 1] = range_m[num_cpu] + width;
range_n[num_cpu] = num_cpu * (((m + 15) & ~15) + 16);
if (range_n[num_cpu] > m) range_n[num_cpu] = m;
queue[num_cpu].mode = mode;
queue[num_cpu].routine = trmv_kernel;
+2 -2
View File
@@ -83,13 +83,13 @@ void CNAME(BLASLONG m, BLASLONG n, BLASLONG ku, BLASLONG kl, FLOAT alpha_r, FLOA
if (incy != 1) {
Y = bufferY;
bufferX = (FLOAT *)(((BLASLONG)bufferY + M * sizeof(FLOAT) * 2 + 4095) & ~4095);
gemvbuffer = bufferX;
// gemvbuffer = bufferX;
COPY_K(M, y, incy, Y, 1);
}
if (incx != 1) {
X = bufferX;
gemvbuffer = (FLOAT *)(((BLASLONG)bufferX + N * sizeof(FLOAT) * 2 + 4095) & ~4095);
// gemvbuffer = (FLOAT *)(((BLASLONG)bufferX + N * sizeof(FLOAT) * 2 + 4095) & ~4095);
COPY_K(N, x, incx, X, 1);
}
+2 -2
View File
@@ -61,13 +61,13 @@ int CNAME(BLASLONG n, BLASLONG k, FLOAT alpha_r, FLOAT alpha_i,
if (incy != 1) {
Y = bufferY;
bufferX = (FLOAT *)(((BLASLONG)bufferY + n * sizeof(FLOAT) * COMPSIZE + 4095) & ~4095);
sbmvbuffer = bufferX;
// sbmvbuffer = bufferX;
COPY_K(n, y, incy, Y, 1);
}
if (incx != 1) {
X = bufferX;
sbmvbuffer = (FLOAT *)(((BLASLONG)bufferX + n * sizeof(FLOAT) * COMPSIZE + 4095) & ~4095);
// sbmvbuffer = (FLOAT *)(((BLASLONG)bufferX + n * sizeof(FLOAT) * COMPSIZE + 4095) & ~4095);
COPY_K(n, x, incx, X, 1);
}
+2 -2
View File
@@ -56,13 +56,13 @@ int CNAME(BLASLONG m, FLOAT alpha_r, FLOAT alpha_i,
if (incy != 1) {
Y = bufferY;
bufferX = (FLOAT *)(((BLASLONG)bufferY + m * sizeof(FLOAT) * 2 + 4095) & ~4095);
gemvbuffer = bufferX;
// gemvbuffer = bufferX;
COPY_K(m, y, incy, Y, 1);
}
if (incx != 1) {
X = bufferX;
gemvbuffer = (FLOAT *)(((BLASLONG)bufferX + m * sizeof(FLOAT) * 2 + 4095) & ~4095);
// gemvbuffer = (FLOAT *)(((BLASLONG)bufferX + m * sizeof(FLOAT) * 2 + 4095) & ~4095);
COPY_K(m, x, incx, X, 1);
}
+2 -2
View File
@@ -60,13 +60,13 @@ int CNAME(BLASLONG n, BLASLONG k, FLOAT alpha_r, FLOAT alpha_i,
if (incy != 1) {
Y = bufferY;
bufferX = (FLOAT *)(((BLASLONG)bufferY + n * sizeof(FLOAT) * COMPSIZE + 4095) & ~4095);
sbmvbuffer = bufferX;
// sbmvbuffer = bufferX;
COPY_K(n, y, incy, Y, 1);
}
if (incx != 1) {
X = bufferX;
sbmvbuffer = (FLOAT *)(((BLASLONG)bufferX + n * sizeof(FLOAT) * COMPSIZE + 4095) & ~4095);
// sbmvbuffer = (FLOAT *)(((BLASLONG)bufferX + n * sizeof(FLOAT) * COMPSIZE + 4095) & ~4095);
COPY_K(n, x, incx, X, 1);
}
+2 -2
View File
@@ -55,13 +55,13 @@ int CNAME(BLASLONG m, FLOAT alpha_r, FLOAT alpha_i,
if (incy != 1) {
Y = bufferY;
bufferX = (FLOAT *)(((BLASLONG)bufferY + m * sizeof(FLOAT) * 2 + 4095) & ~4095);
gemvbuffer = bufferX;
// gemvbuffer = bufferX;
COPY_K(m, y, incy, Y, 1);
}
if (incx != 1) {
X = bufferX;
gemvbuffer = (FLOAT *)(((BLASLONG)bufferX + m * sizeof(FLOAT) * 2 + 4095) & ~4095);
// gemvbuffer = (FLOAT *)(((BLASLONG)bufferX + m * sizeof(FLOAT) * 2 + 4095) & ~4095);
COPY_K(m, x, incx, X, 1);
}
+1 -1
View File
@@ -40,7 +40,7 @@
#include <ctype.h>
#include "common.h"
const static FLOAT dp1 = 1.;
// const static FLOAT dp1 = 1.;
int CNAME(BLASLONG n, BLASLONG k, FLOAT *a, BLASLONG lda, FLOAT *b, BLASLONG incb, void *buffer){
+1 -1
View File
@@ -40,7 +40,7 @@
#include <ctype.h>
#include "common.h"
const static FLOAT dp1 = 1.;
// const static FLOAT dp1 = 1.;
int CNAME(BLASLONG n, BLASLONG k, FLOAT *a, BLASLONG lda, FLOAT *b, BLASLONG incb, void *buffer){
+1 -1
View File
@@ -40,7 +40,7 @@
#include <ctype.h>
#include "common.h"
const static FLOAT dp1 = 1.;
// const static FLOAT dp1 = 1.;
int CNAME(BLASLONG n, BLASLONG k, FLOAT *a, BLASLONG lda, FLOAT *b, BLASLONG incb, void *buffer){
+1 -1
View File
@@ -40,7 +40,7 @@
#include <ctype.h>
#include "common.h"
const static FLOAT dp1 = 1.;
// const static FLOAT dp1 = 1.;
int CNAME(BLASLONG n, BLASLONG k, FLOAT *a, BLASLONG lda, FLOAT *b, BLASLONG incb, void *buffer){
+1 -1
View File
@@ -40,7 +40,7 @@
#include <ctype.h>
#include "common.h"
const static FLOAT dm1 = -1.;
// const static FLOAT dm1 = -1.;
int CNAME(BLASLONG m, FLOAT *a, FLOAT *b, BLASLONG incb, void *buffer){
+6 -5
View File
@@ -1,4 +1,5 @@
include_directories(${PROJECT_SOURCE_DIR})
include_directories(${PROJECT_BINARY_DIR})
# N.B. In the original makefile there was a BLOCKS define used in the compilation of these files but I don't see any evidence of it being set anywhere. -hpa
@@ -8,7 +9,7 @@ set(GEMM_COMPLEX_DEFINES RN CN RT CT NR TR RR CR NC TC RC CC)
foreach (GEMM_DEFINE ${GEMM_DEFINES})
string(TOLOWER ${GEMM_DEFINE} GEMM_DEFINE_LC)
GenerateNamedObjects("gemm.c" "${GEMM_DEFINE}" "gemm_${GEMM_DEFINE_LC}" 0)
if (SMP AND NOT USE_SIMPLE_THREADED_LEVEL3)
if (USE_THREAD AND NOT USE_SIMPLE_THREADED_LEVEL3)
GenerateNamedObjects("gemm.c" "${GEMM_DEFINE};THREADED_LEVEL3" "gemm_thread_${GEMM_DEFINE_LC}" 0)
endif ()
endforeach ()
@@ -31,7 +32,7 @@ GenerateCombinationObjects("syrk_k.c" "LOWER;TRANS" "U;N" "" 1)
GenerateCombinationObjects("syr2k_k.c" "LOWER;TRANS" "U;N" "" 1)
GenerateCombinationObjects("syrk_kernel.c" "LOWER" "U" "" 2)
GenerateCombinationObjects("syr2k_kernel.c" "LOWER" "U" "" 2)
if (SMP)
if (USE_THREAD)
# N.B. these do NOT have a float type (e.g. DOUBLE) defined!
GenerateNamedObjects("gemm_thread_m.c;gemm_thread_n.c;gemm_thread_mn.c;gemm_thread_variable.c;syrk_thread.c" "" "" 0 "" "" 1)
@@ -70,7 +71,7 @@ foreach (float_type ${FLOAT_TYPES})
GenerateNamedObjects("zher2k_k.c" "HER2K;LOWER" "her2k_LN" false "" "" false ${float_type})
GenerateNamedObjects("zher2k_k.c" "HER2K;LOWER;TRANS;CONJ" "her2k_LC" false "" "" false ${float_type})
if (SMP AND NOT USE_SIMPLE_THREADED_LEVEL3)
if (USE_THREAD AND NOT USE_SIMPLE_THREADED_LEVEL3)
#herk
GenerateCombinationObjects("zherk_k.c" "LOWER" "U" "HERK;THREADED_LEVEL3" 3 "herk_thread_N" false ${float_type})
GenerateCombinationObjects("zherk_k.c" "LOWER" "U" "HERK;THREADED_LEVEL3;TRANS;CONJ" 3 "herk_thread_C" false ${float_type})
@@ -92,7 +93,7 @@ foreach (float_type ${FLOAT_TYPES})
if(USE_GEMM3M)
GenerateNamedObjects("gemm3m.c" "${gemm_define}" "gemm3m_${gemm_define_LC}" false "" "" false ${float_type})
endif()
if (SMP AND NOT USE_SIMPLE_THREADED_LEVEL3)
if (USE_THREAD AND NOT USE_SIMPLE_THREADED_LEVEL3)
GenerateNamedObjects("gemm.c" "${gemm_define};THREADED_LEVEL3" "gemm_thread_${gemm_define_LC}" false "" "" false ${float_type})
if(USE_GEMM3M)
GenerateNamedObjects("gemm3m.c" "${gemm_define};THREADED_LEVEL3" "gemm3m_thread_${gemm_define_LC}" false "" "" false ${float_type})
@@ -105,7 +106,7 @@ foreach (float_type ${FLOAT_TYPES})
foreach (GEMM_DEFINE ${GEMM_DEFINES})
string(TOLOWER ${GEMM_DEFINE} GEMM_DEFINE_LC)
GenerateNamedObjects("gemm3m.c" "${GEMM_DEFINE}" "gemm3m_${GEMM_DEFINE_LC}" false "" "" false ${float_type})
if (SMP AND NOT USE_SIMPLE_THREADED_LEVEL3)
if (USE_THREAD AND NOT USE_SIMPLE_THREADED_LEVEL3)
GenerateNamedObjects("gemm3m.c" "${GEMM_DEFINE};THREADED_LEVEL3" "gemm3m_thread_${GEMM_DEFINE_LC}" false "" "" false ${float_type})
endif ()
endforeach ()
+24 -24
View File
@@ -362,7 +362,7 @@ cgemm_ct.$(SUFFIX) : gemm.c level3.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCT $< -o $(@F)
cgemm_cr.$(SUFFIX) : gemm.c level3.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(CFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
cgemm_cc.$(SUFFIX) : gemm.c level3.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -410,7 +410,7 @@ zgemm_ct.$(SUFFIX) : gemm.c level3.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCT $< -o $(@F)
zgemm_cr.$(SUFFIX) : gemm.c level3.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(CFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
zgemm_cc.$(SUFFIX) : gemm.c level3.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -458,7 +458,7 @@ xgemm_ct.$(SUFFIX) : gemm.c level3.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCT $< -o $(@F)
xgemm_cr.$(SUFFIX) : gemm.c level3.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(CFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
xgemm_cc.$(SUFFIX) : gemm.c level3.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -558,7 +558,7 @@ cgemm_thread_ct.$(SUFFIX) : gemm.c level3_thread.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCT $< -o $(@F)
cgemm_thread_cr.$(SUFFIX) : gemm.c level3_thread.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
cgemm_thread_cc.$(SUFFIX) : gemm.c level3_thread.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -606,7 +606,7 @@ zgemm_thread_ct.$(SUFFIX) : gemm.c level3_thread.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCT $< -o $(@F)
zgemm_thread_cr.$(SUFFIX) : gemm.c level3_thread.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
zgemm_thread_cc.$(SUFFIX) : gemm.c level3_thread.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -654,7 +654,7 @@ xgemm_thread_ct.$(SUFFIX) : gemm.c level3_thread.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCT $< -o $(@F)
xgemm_thread_cr.$(SUFFIX) : gemm.c level3_thread.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
xgemm_thread_cc.$(SUFFIX) : gemm.c level3_thread.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -1821,7 +1821,7 @@ cgemm3m_ct.$(SUFFIX) : gemm3m.c gemm3m_level3.c
$(CC) $(CFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCT $< -o $(@F)
cgemm3m_cr.$(SUFFIX) : gemm3m.c gemm3m_level3.c
$(CC) $(CFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(CFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
cgemm3m_cc.$(SUFFIX) : gemm3m.c gemm3m_level3.c
$(CC) $(CFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -1869,7 +1869,7 @@ zgemm3m_ct.$(SUFFIX) : gemm3m.c gemm3m_level3.c
$(CC) $(CFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCT $< -o $(@F)
zgemm3m_cr.$(SUFFIX) : gemm3m.c gemm3m_level3.c
$(CC) $(CFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(CFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
zgemm3m_cc.$(SUFFIX) : gemm3m.c gemm3m_level3.c
$(CC) $(CFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -1917,7 +1917,7 @@ xgemm3m_ct.$(SUFFIX) : gemm3m.c gemm3m_level3.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCT $< -o $(@F)
xgemm3m_cr.$(SUFFIX) : gemm3m.c gemm3m_level3.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(CFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
xgemm3m_cc.$(SUFFIX) : gemm3m.c gemm3m_level3.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -1974,7 +1974,7 @@ cgemm3m_thread_ct.$(SUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCT $< -o $(@F)
cgemm3m_thread_cr.$(SUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
cgemm3m_thread_cc.$(SUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -2022,7 +2022,7 @@ zgemm3m_thread_ct.$(SUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCT $< -o $(@F)
zgemm3m_thread_cr.$(SUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
zgemm3m_thread_cc.$(SUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -2070,7 +2070,7 @@ xgemm3m_thread_ct.$(SUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCT $< -o $(@F)
xgemm3m_thread_cr.$(SUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
xgemm3m_thread_cc.$(SUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
$(CC) $(CFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -2731,7 +2731,7 @@ cgemm_ct.$(PSUFFIX) : gemm.c level3.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCT $< -o $(@F)
cgemm_cr.$(PSUFFIX) : gemm.c level3.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(PFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
cgemm_cc.$(PSUFFIX) : gemm.c level3.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -2779,7 +2779,7 @@ zgemm_ct.$(PSUFFIX) : gemm.c level3.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCT $< -o $(@F)
zgemm_cr.$(PSUFFIX) : gemm.c level3.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(PFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
zgemm_cc.$(PSUFFIX) : gemm.c level3.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -2827,7 +2827,7 @@ xgemm_ct.$(PSUFFIX) : gemm.c level3.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCT $< -o $(@F)
xgemm_cr.$(PSUFFIX) : gemm.c level3.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(PFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
xgemm_cc.$(PSUFFIX) : gemm.c level3.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -2927,7 +2927,7 @@ cgemm_thread_ct.$(PSUFFIX) : gemm.c level3_thread.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCT $< -o $(@F)
cgemm_thread_cr.$(PSUFFIX) : gemm.c level3_thread.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
cgemm_thread_cc.$(PSUFFIX) : gemm.c level3_thread.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -2975,7 +2975,7 @@ zgemm_thread_ct.$(PSUFFIX) : gemm.c level3_thread.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCT $< -o $(@F)
zgemm_thread_cr.$(PSUFFIX) : gemm.c level3_thread.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
zgemm_thread_cc.$(PSUFFIX) : gemm.c level3_thread.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -3023,7 +3023,7 @@ xgemm_thread_ct.$(PSUFFIX) : gemm.c level3_thread.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCT $< -o $(@F)
xgemm_thread_cr.$(PSUFFIX) : gemm.c level3_thread.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
xgemm_thread_cc.$(PSUFFIX) : gemm.c level3_thread.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -4190,7 +4190,7 @@ cgemm3m_ct.$(PSUFFIX) : gemm3m.c gemm3m_level3.c
$(CC) $(PFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCT $< -o $(@F)
cgemm3m_cr.$(PSUFFIX) : gemm3m.c gemm3m_level3.c
$(CC) $(PFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(PFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
cgemm3m_cc.$(PSUFFIX) : gemm3m.c gemm3m_level3.c
$(CC) $(PFLAGS) $(BLOCKS) -c -UDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -4238,7 +4238,7 @@ zgemm3m_ct.$(PSUFFIX) : gemm3m.c gemm3m_level3.c
$(CC) $(PFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCT $< -o $(@F)
zgemm3m_cr.$(PSUFFIX) : gemm3m.c gemm3m_level3.c
$(CC) $(PFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(PFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
zgemm3m_cc.$(PSUFFIX) : gemm3m.c gemm3m_level3.c
$(CC) $(PFLAGS) $(BLOCKS) -c -DDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -4286,7 +4286,7 @@ xgemm3m_ct.$(PSUFFIX) : gemm3m.c gemm3m_level3.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCT $< -o $(@F)
xgemm3m_cr.$(PSUFFIX) : gemm3m.c gemm3m_level3.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(PFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
xgemm3m_cc.$(PSUFFIX) : gemm3m.c gemm3m_level3.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DXDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -4343,7 +4343,7 @@ cgemm3m_thread_ct.$(PSUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCT $< -o $(@F)
cgemm3m_thread_cr.$(PSUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
cgemm3m_thread_cc.$(PSUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -UDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -4391,7 +4391,7 @@ zgemm3m_thread_ct.$(PSUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCT $< -o $(@F)
zgemm3m_thread_cr.$(PSUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
zgemm3m_thread_cc.$(PSUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DDOUBLE -DCOMPLEX -DCC $< -o $(@F)
@@ -4439,7 +4439,7 @@ xgemm3m_thread_ct.$(PSUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCT $< -o $(@F)
xgemm3m_thread_cr.$(PSUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCR $< -o $(@F)
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCR=CR $< -o $(@F)
xgemm3m_thread_cc.$(PSUFFIX) : gemm3m.c level3_gemm3m_thread.c ../../param.h
$(CC) $(PFLAGS) $(BLOCKS) -c -DTHREADED_LEVEL3 -DXDOUBLE -DCOMPLEX -DCC $< -o $(@F)
+4 -4
View File
@@ -251,11 +251,11 @@ int CNAME(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n,
if ((k == 0) || (alpha == NULL)) return 0;
#if !defined(XDOUBLE) || !defined(QUAD_PRECISION)
if ((alpha[0] == ZERO)
if ( alpha[0] == ZERO
#ifdef COMPLEX
&& (alpha[1] == ZERO)
&& alpha[1] == ZERO
#endif
) return 0;
) return 0;
#else
if (((alpha[0].x[0] | alpha[0].x[1]
#ifdef COMPLEX
@@ -293,7 +293,7 @@ int CNAME(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n,
min_l = k - ls;
if (min_l >= GEMM_Q * 2) {
gemm_p = GEMM_P;
// gemm_p = GEMM_P;
min_l = GEMM_Q;
} else {
if (min_l > GEMM_Q) {
+9 -3
View File
@@ -91,7 +91,12 @@
#endif
typedef struct {
volatile BLASLONG working[MAX_CPU_NUMBER][CACHE_LINE_SIZE * DIVIDE_RATE];
#if __STDC_VERSION__ >= 201112L
_Atomic
#else
volatile
#endif
BLASLONG working[MAX_CPU_NUMBER][CACHE_LINE_SIZE * DIVIDE_RATE];
} job_t;
@@ -974,7 +979,7 @@ static int gemm_driver(blas_arg_t *args, BLASLONG *range_m, BLASLONG
int CNAME(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n, FLOAT *sa, FLOAT *sb, BLASLONG mypos){
BLASLONG m = args -> m;
BLASLONG n = args -> n;
// BLASLONG n = args -> n;
BLASLONG nthreads = args -> nthreads;
BLASLONG divN, divT;
int mode;
@@ -985,13 +990,14 @@ int CNAME(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n, FLOAT *sa, FLO
m = m_to - m_from;
}
/*
if (range_n) {
BLASLONG n_from = *(((BLASLONG *)range_n) + 0);
BLASLONG n_to = *(((BLASLONG *)range_n) + 1);
n = n_to - n_from;
}
*/
if ((args -> m < nthreads * SWITCH_RATIO) || (args -> n < nthreads * SWITCH_RATIO)) {
GEMM3M_LOCAL(args, range_m, range_n, sa, sb, 0);
+2 -2
View File
@@ -154,9 +154,9 @@ int CNAME(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n, FLOAT *sa, FLO
if ((k == 0) || (alpha == NULL)) return 0;
if ((alpha[0] == ZERO)
if (alpha[0] == ZERO
#ifdef COMPLEX
&& (alpha[1] == ZERO)
&& alpha[1] == ZERO
#endif
) return 0;
+2 -2
View File
@@ -158,9 +158,9 @@ int CNAME(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n, FLOAT *sa, FLO
if ((k == 0) || (alpha == NULL)) return 0;
if ((alpha[0] == ZERO)
if (alpha[0] == ZERO
#if defined(COMPLEX) && !defined(HERK)
&& (alpha[1] == ZERO)
&& alpha[1] == ZERO
#endif
) return 0;
+8 -3
View File
@@ -67,7 +67,12 @@
#endif
typedef struct {
volatile BLASLONG working[MAX_CPU_NUMBER][CACHE_LINE_SIZE * DIVIDE_RATE];
#if __STDC_VERSION__ >= 201112L
_Atomic
#else
volatile
#endif
BLASLONG working[MAX_CPU_NUMBER][CACHE_LINE_SIZE * DIVIDE_RATE];
} job_t;
@@ -200,9 +205,9 @@ static int inner_thread(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n,
if ((k == 0) || (alpha == NULL)) return 0;
if ((alpha[0] == ZERO)
if (alpha[0] == ZERO
#if defined(COMPLEX) && !defined(HERK)
&& (alpha[1] == ZERO)
&& alpha[1] == ZERO
#endif
) return 0;
+201 -252
View File
@@ -91,27 +91,28 @@
#endif
typedef struct {
volatile BLASLONG working[MAX_CPU_NUMBER][CACHE_LINE_SIZE * DIVIDE_RATE];
volatile
BLASLONG working[MAX_CPU_NUMBER][CACHE_LINE_SIZE * DIVIDE_RATE];
} job_t;
#ifndef BETA_OPERATION
#ifndef COMPLEX
#define BETA_OPERATION(M_FROM, M_TO, N_FROM, N_TO, BETA, C, LDC) \
GEMM_BETA((M_TO) - (M_FROM), (N_TO - N_FROM), 0, \
BETA[0], NULL, 0, NULL, 0, \
(FLOAT *)(C) + ((M_FROM) + (N_FROM) * (LDC)) * COMPSIZE, LDC)
#define BETA_OPERATION(M_FROM, M_TO, N_FROM, N_TO, BETA, C, LDC) \
GEMM_BETA((M_TO) - (M_FROM), (N_TO - N_FROM), 0, \
BETA[0], NULL, 0, NULL, 0, \
(FLOAT *)(C) + ((M_FROM) + (N_FROM) * (LDC)) * COMPSIZE, LDC)
#else
#define BETA_OPERATION(M_FROM, M_TO, N_FROM, N_TO, BETA, C, LDC) \
GEMM_BETA((M_TO) - (M_FROM), (N_TO - N_FROM), 0, \
BETA[0], BETA[1], NULL, 0, NULL, 0, \
(FLOAT *)(C) + ((M_FROM) + (N_FROM) * (LDC)) * COMPSIZE, LDC)
#define BETA_OPERATION(M_FROM, M_TO, N_FROM, N_TO, BETA, C, LDC) \
GEMM_BETA((M_TO) - (M_FROM), (N_TO - N_FROM), 0, \
BETA[0], BETA[1], NULL, 0, NULL, 0, \
(FLOAT *)(C) + ((M_FROM) + (N_FROM) * (LDC)) * COMPSIZE, LDC)
#endif
#endif
#ifndef ICOPY_OPERATION
#if defined(NN) || defined(NT) || defined(NC) || defined(NR) || \
defined(RN) || defined(RT) || defined(RC) || defined(RR)
defined(RN) || defined(RT) || defined(RC) || defined(RR)
#define ICOPY_OPERATION(M, N, A, LDA, X, Y, BUFFER) GEMM_ITCOPY(M, N, (FLOAT *)(A) + ((Y) + (X) * (LDA)) * COMPSIZE, LDA, BUFFER);
#else
#define ICOPY_OPERATION(M, N, A, LDA, X, Y, BUFFER) GEMM_INCOPY(M, N, (FLOAT *)(A) + ((X) + (Y) * (LDA)) * COMPSIZE, LDA, BUFFER);
@@ -120,7 +121,7 @@ typedef struct {
#ifndef OCOPY_OPERATION
#if defined(NN) || defined(TN) || defined(CN) || defined(RN) || \
defined(NR) || defined(TR) || defined(CR) || defined(RR)
defined(NR) || defined(TR) || defined(CR) || defined(RR)
#define OCOPY_OPERATION(M, N, A, LDA, X, Y, BUFFER) GEMM_ONCOPY(M, N, (FLOAT *)(A) + ((X) + (Y) * (LDA)) * COMPSIZE, LDA, BUFFER);
#else
#define OCOPY_OPERATION(M, N, A, LDA, X, Y, BUFFER) GEMM_OTCOPY(M, N, (FLOAT *)(A) + ((Y) + (X) * (LDA)) * COMPSIZE, LDA, BUFFER);
@@ -144,36 +145,36 @@ typedef struct {
#ifndef KERNEL_OPERATION
#ifndef COMPLEX
#define KERNEL_OPERATION(M, N, K, ALPHA, SA, SB, C, LDC, X, Y) \
KERNEL_FUNC(M, N, K, ALPHA[0], SA, SB, (FLOAT *)(C) + ((X) + (Y) * LDC) * COMPSIZE, LDC)
#define KERNEL_OPERATION(M, N, K, ALPHA, SA, SB, C, LDC, X, Y) \
KERNEL_FUNC(M, N, K, ALPHA[0], SA, SB, (FLOAT *)(C) + ((X) + (Y) * LDC) * COMPSIZE, LDC)
#else
#define KERNEL_OPERATION(M, N, K, ALPHA, SA, SB, C, LDC, X, Y) \
KERNEL_FUNC(M, N, K, ALPHA[0], ALPHA[1], SA, SB, (FLOAT *)(C) + ((X) + (Y) * LDC) * COMPSIZE, LDC)
#define KERNEL_OPERATION(M, N, K, ALPHA, SA, SB, C, LDC, X, Y) \
KERNEL_FUNC(M, N, K, ALPHA[0], ALPHA[1], SA, SB, (FLOAT *)(C) + ((X) + (Y) * LDC) * COMPSIZE, LDC)
#endif
#endif
#ifndef FUSED_KERNEL_OPERATION
#if defined(NN) || defined(TN) || defined(CN) || defined(RN) || \
defined(NR) || defined(TR) || defined(CR) || defined(RR)
defined(NR) || defined(TR) || defined(CR) || defined(RR)
#ifndef COMPLEX
#define FUSED_KERNEL_OPERATION(M, N, K, ALPHA, SA, SB, B, LDB, C, LDC, I, J, L) \
FUSED_GEMM_KERNEL_N(M, N, K, ALPHA[0], SA, SB, \
(FLOAT *)(B) + ((L) + (J) * LDB) * COMPSIZE, LDB, (FLOAT *)(C) + ((I) + (J) * LDC) * COMPSIZE, LDC)
FUSED_GEMM_KERNEL_N(M, N, K, ALPHA[0], SA, SB, \
(FLOAT *)(B) + ((L) + (J) * LDB) * COMPSIZE, LDB, (FLOAT *)(C) + ((I) + (J) * LDC) * COMPSIZE, LDC)
#else
#define FUSED_KERNEL_OPERATION(M, N, K, ALPHA, SA, SB, B, LDB, C, LDC, I, J, L) \
FUSED_GEMM_KERNEL_N(M, N, K, ALPHA[0], ALPHA[1], SA, SB, \
(FLOAT *)(B) + ((L) + (J) * LDB) * COMPSIZE, LDB, (FLOAT *)(C) + ((I) + (J) * LDC) * COMPSIZE, LDC)
FUSED_GEMM_KERNEL_N(M, N, K, ALPHA[0], ALPHA[1], SA, SB, \
(FLOAT *)(B) + ((L) + (J) * LDB) * COMPSIZE, LDB, (FLOAT *)(C) + ((I) + (J) * LDC) * COMPSIZE, LDC)
#endif
#else
#ifndef COMPLEX
#define FUSED_KERNEL_OPERATION(M, N, K, ALPHA, SA, SB, B, LDB, C, LDC, I, J, L) \
FUSED_GEMM_KERNEL_T(M, N, K, ALPHA[0], SA, SB, \
(FLOAT *)(B) + ((J) + (L) * LDB) * COMPSIZE, LDB, (FLOAT *)(C) + ((I) + (J) * LDC) * COMPSIZE, LDC)
FUSED_GEMM_KERNEL_T(M, N, K, ALPHA[0], SA, SB, \
(FLOAT *)(B) + ((J) + (L) * LDB) * COMPSIZE, LDB, (FLOAT *)(C) + ((I) + (J) * LDC) * COMPSIZE, LDC)
#else
#define FUSED_KERNEL_OPERATION(M, N, K, ALPHA, SA, SB, B, LDB, C, LDC, I, J, L) \
FUSED_GEMM_KERNEL_T(M, N, K, ALPHA[0], ALPHA[1], SA, SB, \
(FLOAT *)(B) + ((J) + (L) * LDB) * COMPSIZE, LDB, (FLOAT *)(C) + ((I) + (J) * LDC) * COMPSIZE, LDC)
FUSED_GEMM_KERNEL_T(M, N, K, ALPHA[0], ALPHA[1], SA, SB, \
(FLOAT *)(B) + ((J) + (L) * LDB) * COMPSIZE, LDB, (FLOAT *)(C) + ((I) + (J) * LDC) * COMPSIZE, LDC)
#endif
#endif
#endif
@@ -219,15 +220,17 @@ static int inner_thread(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n,
FLOAT *buffer[DIVIDE_RATE];
BLASLONG k, lda, ldb, ldc;
BLASLONG m_from, m_to, n_from, n_to, N_from, N_to;
BLASLONG m_from, m_to, n_from, n_to;
FLOAT *alpha, *beta;
FLOAT *a, *b, *c;
job_t *job = (job_t *)args -> common;
BLASLONG xxx, bufferside;
BLASLONG ls, min_l, jjs, min_jj;
BLASLONG is, min_i, div_n;
BLASLONG nthreads_m;
BLASLONG mypos_m, mypos_n;
BLASLONG is, js, ls, bufferside, jjs;
BLASLONG min_i, min_l, div_n, min_jj;
BLASLONG i, current;
BLASLONG l1stride;
@@ -259,74 +262,69 @@ static int inner_thread(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n,
alpha = (FLOAT *)args -> alpha;
beta = (FLOAT *)args -> beta;
/* Initialize 2D CPU distribution */
nthreads_m = args -> nthreads;
if (range_m) {
nthreads_m = range_m[-1];
}
mypos_n = blas_quickdivide(mypos, nthreads_m); /* mypos_n = mypos / nthreads_m */
mypos_m = mypos - mypos_n * nthreads_m; /* mypos_m = mypos % nthreads_m */
/* Initialize m and n */
m_from = 0;
m_to = M;
if (range_m) {
m_from = range_m[0];
m_to = range_m[1];
m_from = range_m[mypos_m + 0];
m_to = range_m[mypos_m + 1];
}
n_from = 0;
n_to = N;
N_from = 0;
N_to = N;
if (range_n) {
n_from = range_n[mypos + 0];
n_to = range_n[mypos + 1];
N_from = range_n[0];
N_to = range_n[args -> nthreads];
}
/* Multiply C by beta if needed */
if (beta) {
#ifndef COMPLEX
if (beta[0] != ONE)
#else
if ((beta[0] != ONE) || (beta[1] != ZERO))
#endif
BETA_OPERATION(m_from, m_to, N_from, N_to, beta, c, ldc);
BETA_OPERATION(m_from, m_to, range_n[mypos_n * nthreads_m], range_n[(mypos_n + 1) * nthreads_m], beta, c, ldc);
}
/* Return early if no more computation is needed */
if ((k == 0) || (alpha == NULL)) return 0;
if ((alpha[0] == ZERO)
if (alpha[0] == ZERO
#ifdef COMPLEX
&& (alpha[1] == ZERO)
&& alpha[1] == ZERO
#endif
) return 0;
#if 0
fprintf(stderr, "Thread[%ld] m_from : %ld m_to : %ld n_from : %ld n_to : %ld N_from : %ld N_to : %ld\n",
mypos, m_from, m_to, n_from, n_to, N_from, N_to);
fprintf(stderr, "GEMM: P = %4ld Q = %4ld R = %4ld\n", (BLASLONG)GEMM_P, (BLASLONG)GEMM_Q, (BLASLONG)GEMM_R);
#endif
/* Initialize workspace for local region of B */
div_n = (n_to - n_from + DIVIDE_RATE - 1) / DIVIDE_RATE;
buffer[0] = sb;
for (i = 1; i < DIVIDE_RATE; i++) {
buffer[i] = buffer[i - 1] + GEMM_Q * ((div_n + GEMM_UNROLL_N - 1)/GEMM_UNROLL_N) * GEMM_UNROLL_N * COMPSIZE;
}
/* Iterate through steps of k */
for(ls = 0; ls < k; ls += min_l){
/* Determine step size in k */
min_l = k - ls;
if (min_l >= GEMM_Q * 2) {
min_l = GEMM_Q;
} else {
if (min_l > GEMM_Q) min_l = (min_l + 1) / 2;
}
/* Determine step size in m
* Note: We are currently on the first step in m
*/
l1stride = 1;
min_i = m_to - m_from;
if (min_i >= GEMM_P * 2) {
min_i = GEMM_P;
} else {
@@ -337,108 +335,107 @@ static int inner_thread(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n,
}
}
/* Copy local region of A into workspace */
START_RPCC();
ICOPY_OPERATION(min_l, min_i, a, lda, ls, m_from, sa);
STOP_RPCC(copy_A);
/* Copy local region of B into workspace and apply kernel */
div_n = (n_to - n_from + DIVIDE_RATE - 1) / DIVIDE_RATE;
for (js = n_from, bufferside = 0; js < n_to; js += div_n, bufferside ++) {
for (xxx = n_from, bufferside = 0; xxx < n_to; xxx += div_n, bufferside ++) {
/* Make sure if no one is using workspace */
START_RPCC();
/* Make sure if no one is using buffer */
for (i = 0; i < args -> nthreads; i++)
while (job[mypos].working[i][CACHE_LINE_SIZE * bufferside]) {YIELDING;};
while (job[mypos].working[i][CACHE_LINE_SIZE * bufferside]) {YIELDING;MB;};
STOP_RPCC(waiting1);
#if defined(FUSED_GEMM) && !defined(TIMING)
FUSED_KERNEL_OPERATION(min_i, MIN(n_to, xxx + div_n) - xxx, min_l, alpha,
sa, buffer[bufferside], b, ldb, c, ldc, m_from, xxx, ls);
/* Fused operation to copy region of B into workspace and apply kernel */
FUSED_KERNEL_OPERATION(min_i, MIN(n_to, js + div_n) - js, min_l, alpha,
sa, buffer[bufferside], b, ldb, c, ldc, m_from, js, ls);
#else
for(jjs = xxx; jjs < MIN(n_to, xxx + div_n); jjs += min_jj){
min_jj = MIN(n_to, xxx + div_n) - jjs;
/* Split local region of B into parts */
for(jjs = js; jjs < MIN(n_to, js + div_n); jjs += min_jj){
min_jj = MIN(n_to, js + div_n) - jjs;
if (min_jj >= 3*GEMM_UNROLL_N) min_jj = 3*GEMM_UNROLL_N;
else
if (min_jj >= 2*GEMM_UNROLL_N) min_jj = 2*GEMM_UNROLL_N;
else
if (min_jj > GEMM_UNROLL_N) min_jj = GEMM_UNROLL_N;
if (min_jj >= 2*GEMM_UNROLL_N) min_jj = 2*GEMM_UNROLL_N;
else
if (min_jj > GEMM_UNROLL_N) min_jj = GEMM_UNROLL_N;
/* Copy part of local region of B into workspace */
START_RPCC();
OCOPY_OPERATION(min_l, min_jj, b, ldb, ls, jjs,
buffer[bufferside] + min_l * (jjs - xxx) * COMPSIZE * l1stride);
buffer[bufferside] + min_l * (jjs - js) * COMPSIZE * l1stride);
STOP_RPCC(copy_B);
/* Apply kernel with local region of A and part of local region of B */
START_RPCC();
KERNEL_OPERATION(min_i, min_jj, min_l, alpha,
sa, buffer[bufferside] + min_l * (jjs - xxx) * COMPSIZE * l1stride,
sa, buffer[bufferside] + min_l * (jjs - js) * COMPSIZE * l1stride,
c, ldc, m_from, jjs);
STOP_RPCC(kernel);
#ifdef TIMING
ops += 2 * min_i * min_jj * min_l;
ops += 2 * min_i * min_jj * min_l;
#endif
}
#endif
for (i = 0; i < args -> nthreads; i++) job[mypos].working[i][CACHE_LINE_SIZE * bufferside] = (BLASLONG)buffer[bufferside];
/* Set flag so other threads can access local region of B */
for (i = mypos_n * nthreads_m; i < (mypos_n + 1) * nthreads_m; i++)
job[mypos].working[i][CACHE_LINE_SIZE * bufferside] = (BLASLONG)buffer[bufferside];
WMB;
}
/* Get regions of B from other threads and apply kernel */
current = mypos;
do {
/* This thread accesses regions of B from threads in the range
* [ mypos_n * nthreads_m, (mypos_n+1) * nthreads_m ) */
current ++;
if (current >= args -> nthreads) current = 0;
if (current >= (mypos_n + 1) * nthreads_m) current = mypos_n * nthreads_m;
/* Split other region of B into parts */
div_n = (range_n[current + 1] - range_n[current] + DIVIDE_RATE - 1) / DIVIDE_RATE;
for (js = range_n[current], bufferside = 0; js < range_n[current + 1]; js += div_n, bufferside ++) {
if (current != mypos) {
for (xxx = range_n[current], bufferside = 0; xxx < range_n[current + 1]; xxx += div_n, bufferside ++) {
if (current != mypos) {
/* Wait until other region of B is initialized */
START_RPCC();
/* thread has to wait */
while(job[current].working[mypos][CACHE_LINE_SIZE * bufferside] == 0) {YIELDING;};
while(job[current].working[mypos][CACHE_LINE_SIZE * bufferside] == 0) {YIELDING;MB;};
STOP_RPCC(waiting2);
/* Apply kernel with local region of A and part of other region of B */
START_RPCC();
KERNEL_OPERATION(min_i, MIN(range_n[current + 1] - xxx, div_n), min_l, alpha,
KERNEL_OPERATION(min_i, MIN(range_n[current + 1] - js, div_n), min_l, alpha,
sa, (FLOAT *)job[current].working[mypos][CACHE_LINE_SIZE * bufferside],
c, ldc, m_from, xxx);
c, ldc, m_from, js);
STOP_RPCC(kernel);
STOP_RPCC(kernel);
#ifdef TIMING
ops += 2 * min_i * MIN(range_n[current + 1] - xxx, div_n) * min_l;
ops += 2 * min_i * MIN(range_n[current + 1] - js, div_n) * min_l;
#endif
}
/* Clear synchronization flag if this thread is done with other region of B */
if (m_to - m_from == min_i) {
job[current].working[mypos][CACHE_LINE_SIZE * bufferside] &= 0;
WMB;
}
}
} while (current != mypos);
/* Iterate through steps of m
* Note: First step has already been finished */
for(is = m_from + min_i; is < m_to; is += min_i){
min_i = m_to - is;
if (min_i >= GEMM_P * 2) {
min_i = GEMM_P;
} else
@@ -446,40 +443,41 @@ static int inner_thread(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n,
min_i = (((min_i + 1) / 2 + GEMM_UNROLL_M - 1)/GEMM_UNROLL_M) * GEMM_UNROLL_M;
}
/* Copy local region of A into workspace */
START_RPCC();
ICOPY_OPERATION(min_l, min_i, a, lda, ls, is, sa);
STOP_RPCC(copy_A);
/* Get regions of B and apply kernel */
current = mypos;
do {
/* Split region of B into parts and apply kernel */
div_n = (range_n[current + 1] - range_n[current] + DIVIDE_RATE - 1) / DIVIDE_RATE;
for (js = range_n[current], bufferside = 0; js < range_n[current + 1]; js += div_n, bufferside ++) {
for (xxx = range_n[current], bufferside = 0; xxx < range_n[current + 1]; xxx += div_n, bufferside ++) {
/* Apply kernel with local region of A and part of region of B */
START_RPCC();
KERNEL_OPERATION(min_i, MIN(range_n[current + 1] - xxx, div_n), min_l, alpha,
KERNEL_OPERATION(min_i, MIN(range_n[current + 1] - js, div_n), min_l, alpha,
sa, (FLOAT *)job[current].working[mypos][CACHE_LINE_SIZE * bufferside],
c, ldc, is, xxx);
STOP_RPCC(kernel);
c, ldc, is, js);
STOP_RPCC(kernel);
#ifdef TIMING
ops += 2 * min_i * MIN(range_n[current + 1] - xxx, div_n) * min_l;
ops += 2 * min_i * MIN(range_n[current + 1] - js, div_n) * min_l;
#endif
if (is + min_i >= m_to) {
/* Thread doesn't need this buffer any more */
job[current].working[mypos][CACHE_LINE_SIZE * bufferside] &= 0;
WMB;
}
/* Clear synchronization flag if this thread is done with region of B */
if (is + min_i >= m_to) {
job[current].working[mypos][CACHE_LINE_SIZE * bufferside] &= 0;
WMB;
}
}
/* This thread accesses regions of B from threads in the range
* [ mypos_n * nthreads_m, (mypos_n+1) * nthreads_m ) */
current ++;
if (current >= args -> nthreads) current = 0;
if (current >= (mypos_n + 1) * nthreads_m) current = mypos_n * nthreads_m;
} while (current != mypos);
@@ -487,14 +485,13 @@ static int inner_thread(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n,
}
/* Wait until all other threads are done with local region of B */
START_RPCC();
for (i = 0; i < args -> nthreads; i++) {
for (xxx = 0; xxx < DIVIDE_RATE; xxx++) {
while (job[mypos].working[i][CACHE_LINE_SIZE * xxx] ) {YIELDING;};
for (js = 0; js < DIVIDE_RATE; js++) {
while (job[mypos].working[i][CACHE_LINE_SIZE * js] ) {YIELDING;MB;};
}
}
STOP_RPCC(waiting3);
#ifdef TIMING
@@ -507,17 +504,6 @@ static int inner_thread(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n,
(double)waiting2 /(double)total * 100.,
(double)waiting3 /(double)total * 100.,
(double)ops/(double)kernel / 4. * 100.);
#if 0
fprintf(stderr, "GEMM [%2ld] Copy_A : %6.2ld Copy_B : %6.2ld Wait : %6.2ld\n",
mypos, copy_A, copy_B, waiting);
fprintf(stderr, "Waiting[%2ld] %6.2f %6.2f %6.2f\n",
mypos,
(double)waiting1/(double)waiting * 100.,
(double)waiting2/(double)waiting * 100.,
(double)waiting3/(double)waiting * 100.);
#endif
fprintf(stderr, "\n");
#endif
@@ -525,7 +511,8 @@ static int inner_thread(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n,
}
static int gemm_driver(blas_arg_t *args, BLASLONG *range_m, BLASLONG
*range_n, FLOAT *sa, FLOAT *sb, BLASLONG mypos){
*range_n, FLOAT *sa, FLOAT *sb,
BLASLONG nthreads_m, BLASLONG nthreads_n) {
blas_arg_t newarg;
@@ -537,17 +524,18 @@ static int gemm_driver(blas_arg_t *args, BLASLONG *range_m, BLASLONG
blas_queue_t queue[MAX_CPU_NUMBER];
BLASLONG range_M[MAX_CPU_NUMBER + 1];
BLASLONG range_N[MAX_CPU_NUMBER + 1];
BLASLONG num_cpu_m, num_cpu_n;
BLASLONG range_M_buffer[MAX_CPU_NUMBER + 2];
BLASLONG range_N_buffer[MAX_CPU_NUMBER + 2];
BLASLONG *range_M, *range_N;
BLASLONG num_parts;
BLASLONG nthreads = args -> nthreads;
BLASLONG width, i, j, k, js;
BLASLONG m, n, n_from, n_to;
int mode;
int mode;
/* Get execution mode */
#ifndef COMPLEX
#ifdef XDOUBLE
mode = BLAS_XDOUBLE | BLAS_REAL | BLAS_NODE;
@@ -566,6 +554,16 @@ static int gemm_driver(blas_arg_t *args, BLASLONG *range_m, BLASLONG
#endif
#endif
#ifdef USE_ALLOC_HEAP
/* Dynamically allocate workspace */
job = (job_t*)malloc(MAX_CPU_NUMBER * sizeof(job_t));
if(job==NULL){
fprintf(stderr, "OpenBLAS: malloc failed in %s\n", __func__);
exit(1);
}
#endif
/* Initialize struct for arguments */
newarg.m = args -> m;
newarg.n = args -> n;
newarg.k = args -> k;
@@ -578,23 +576,19 @@ static int gemm_driver(blas_arg_t *args, BLASLONG *range_m, BLASLONG
newarg.alpha = args -> alpha;
newarg.beta = args -> beta;
newarg.nthreads = args -> nthreads;
#ifdef USE_ALLOC_HEAP
job = (job_t*)malloc(MAX_CPU_NUMBER * sizeof(job_t));
if(job==NULL){
fprintf(stderr, "OpenBLAS: malloc failed in %s\n", __func__);
exit(1);
}
#endif
newarg.common = (void *)job;
#ifdef PARAMTEST
newarg.gemm_p = args -> gemm_p;
newarg.gemm_q = args -> gemm_q;
newarg.gemm_r = args -> gemm_r;
newarg.gemm_p = args -> gemm_p;
newarg.gemm_q = args -> gemm_q;
newarg.gemm_r = args -> gemm_r;
#endif
/* Initialize partitions in m and n
* Note: The number of CPU partitions is stored in the -1 entry */
range_M = &range_M_buffer[1];
range_N = &range_N_buffer[1];
range_M[-1] = nthreads_m;
range_N[-1] = nthreads_n;
if (!range_m) {
range_M[0] = 0;
m = args -> m;
@@ -603,34 +597,35 @@ static int gemm_driver(blas_arg_t *args, BLASLONG *range_m, BLASLONG
m = range_m[1] - range_m[0];
}
num_cpu_m = 0;
/* Partition m into nthreads_m regions */
num_parts = 0;
while (m > 0){
width = blas_quickdivide(m + nthreads - num_cpu_m - 1, nthreads - num_cpu_m);
width = blas_quickdivide(m + nthreads_m - num_parts - 1, nthreads_m - num_parts);
m -= width;
if (m < 0) width = width + m;
range_M[num_cpu_m + 1] = range_M[num_cpu_m] + width;
num_cpu_m ++;
range_M[num_parts + 1] = range_M[num_parts] + width;
num_parts ++;
}
for (i = num_parts; i < MAX_CPU_NUMBER; i++) {
range_M[i + 1] = range_M[num_parts];
}
for (i = 0; i < num_cpu_m; i++) {
/* Initialize parameters for parallel execution */
for (i = 0; i < nthreads; i++) {
queue[i].mode = mode;
queue[i].routine = inner_thread;
queue[i].args = &newarg;
queue[i].range_m = &range_M[i];
queue[i].range_n = &range_N[0];
queue[i].range_m = range_M;
queue[i].range_n = range_N;
queue[i].sa = NULL;
queue[i].sb = NULL;
queue[i].next = &queue[i + 1];
}
queue[0].sa = sa;
queue[0].sb = sb;
queue[nthreads - 1].next = NULL;
/* Iterate through steps of n */
if (!range_n) {
n_from = 0;
n_to = args -> n;
@@ -638,38 +633,38 @@ static int gemm_driver(blas_arg_t *args, BLASLONG *range_m, BLASLONG
n_from = range_n[0];
n_to = range_n[1];
}
for(js = n_from; js < n_to; js += GEMM_R * nthreads){
n = n_to - js;
if (n > GEMM_R * nthreads) n = GEMM_R * nthreads;
/* Partition (a step of) n into nthreads regions */
range_N[0] = js;
num_cpu_n = 0;
num_parts = 0;
while (n > 0){
width = blas_quickdivide(n + nthreads - num_cpu_n - 1, nthreads - num_cpu_n);
width = blas_quickdivide(n + nthreads - num_parts - 1, nthreads - num_parts);
if (width < SWITCH_RATIO) {
width = SWITCH_RATIO;
}
n -= width;
if (n < 0) width = width + n;
range_N[num_cpu_n + 1] = range_N[num_cpu_n] + width;
num_cpu_n ++;
range_N[num_parts + 1] = range_N[num_parts] + width;
num_parts ++;
}
for (j = num_parts; j < MAX_CPU_NUMBER; j++) {
range_N[j + 1] = range_N[num_parts];
}
for (j = 0; j < num_cpu_m; j++) {
for (i = 0; i < num_cpu_m; i++) {
/* Clear synchronization flags */
for (i = 0; i < nthreads; i++) {
for (j = 0; j < nthreads; j++) {
for (k = 0; k < DIVIDE_RATE; k++) {
job[j].working[i][CACHE_LINE_SIZE * k] = 0;
job[i].working[j][CACHE_LINE_SIZE * k] = 0;
}
}
}
queue[num_cpu_m - 1].next = NULL;
exec_blas(num_cpu_m, queue);
/* Execute parallel computation */
exec_blas(nthreads, queue);
}
#ifdef USE_ALLOC_HEAP
@@ -683,88 +678,42 @@ int CNAME(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n, FLOAT *sa, FLO
BLASLONG m = args -> m;
BLASLONG n = args -> n;
BLASLONG nthreads = args -> nthreads;
BLASLONG divN, divT;
int mode;
if (nthreads == 1) {
GEMM_LOCAL(args, range_m, range_n, sa, sb, 0);
return 0;
}
BLASLONG nthreads_m, nthreads_n;
/* Get dimensions from index ranges if available */
if (range_m) {
BLASLONG m_from = *(((BLASLONG *)range_m) + 0);
BLASLONG m_to = *(((BLASLONG *)range_m) + 1);
m = m_to - m_from;
m = range_m[1] - range_m[0];
}
if (range_n) {
BLASLONG n_from = *(((BLASLONG *)range_n) + 0);
BLASLONG n_to = *(((BLASLONG *)range_n) + 1);
n = n_to - n_from;
n = range_n[1] - range_n[0];
}
if ((m < nthreads * SWITCH_RATIO) || (n < nthreads * SWITCH_RATIO)) {
GEMM_LOCAL(args, range_m, range_n, sa, sb, 0);
return 0;
}
divT = nthreads;
divN = 1;
#if 0
while ((GEMM_P * divT > m * SWITCH_RATIO) && (divT > 1)) {
do {
divT --;
divN = 1;
while (divT * divN < nthreads) divN ++;
} while ((divT * divN != nthreads) && (divT > 1));
}
#endif
// fprintf(stderr, "divN = %4ld divT = %4ld\n", divN, divT);
args -> nthreads = divT;
if (divN == 1){
gemm_driver(args, range_m, range_n, sa, sb, 0);
/* Partitions in m should have at least SWITCH_RATIO rows */
if (m < 2 * SWITCH_RATIO) {
nthreads_m = 1;
} else {
#ifndef COMPLEX
#ifdef XDOUBLE
mode = BLAS_XDOUBLE | BLAS_REAL;
#elif defined(DOUBLE)
mode = BLAS_DOUBLE | BLAS_REAL;
#else
mode = BLAS_SINGLE | BLAS_REAL;
#endif
#else
#ifdef XDOUBLE
mode = BLAS_XDOUBLE | BLAS_COMPLEX;
#elif defined(DOUBLE)
mode = BLAS_DOUBLE | BLAS_COMPLEX;
#else
mode = BLAS_SINGLE | BLAS_COMPLEX;
#endif
#endif
nthreads_m = args -> nthreads;
while (m < nthreads_m * SWITCH_RATIO) {
nthreads_m = nthreads_m / 2;
}
}
#if defined(TN) || defined(TT) || defined(TR) || defined(TC) || \
defined(CN) || defined(CT) || defined(CR) || defined(CC)
mode |= (BLAS_TRANSA_T);
#endif
#if defined(NT) || defined(TT) || defined(RT) || defined(CT) || \
defined(NC) || defined(TC) || defined(RC) || defined(CC)
mode |= (BLAS_TRANSB_T);
#endif
#ifdef OS_WINDOWS
gemm_thread_n(mode, args, range_m, range_n, GEMM_LOCAL, sa, sb, divN);
#else
gemm_thread_n(mode, args, range_m, range_n, gemm_driver, sa, sb, divN);
#endif
/* Partitions in n should have at most SWITCH_RATIO * nthreads_m columns */
if (n < SWITCH_RATIO * nthreads_m) {
nthreads_n = 1;
} else {
nthreads_n = (n + SWITCH_RATIO * nthreads_m - 1) / (SWITCH_RATIO * nthreads_m);
if (nthreads_m * nthreads_n > args -> nthreads) {
nthreads_n = blas_quickdivide(args -> nthreads, nthreads_m);
}
}
/* Execute serial or parallel computation */
if (nthreads_m * nthreads_n <= 1) {
GEMM_LOCAL(args, range_m, range_n, sa, sb, 0);
} else {
args -> nthreads = nthreads_m * nthreads_n;
gemm_driver(args, range_m, range_n, sa, sb, nthreads_m, nthreads_n);
}
return 0;
+2 -1
View File
@@ -1,4 +1,5 @@
include_directories(${PROJECT_SOURCE_DIR})
include_directories(${PROJECT_BINARY_DIR})
if (${CORE} STREQUAL "PPC440")
set(MEMORY memory_qalloc.c)
@@ -6,7 +7,7 @@ else ()
set(MEMORY memory.c)
endif ()
if (SMP)
if (USE_THREAD)
if (USE_OPENMP)
set(BLAS_SERVER blas_server_omp.c)

Some files were not shown because too many files have changed in this diff Show More