phynexis-deps/lapack
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
3,428 Commits 10 Branches 16 Tags
9aed94de6d78beddb218baaebce809193b8160fc
Commit Graph
3428 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Nikola Forró 3c7d0f062d Fix macro used in LAPACKE_zgesvdq 
This looks like a copy-paste error. gcc emits a warning about it:

lapacke_zgesvdq.c: In function 'LAPACKE_zgesvdq':
../include/lapacke.h:50:42: warning: dereferencing type-punned pointer will break strict-aliasing rules [-Wstrict-aliasing]

Signed-off-by: Nikola Forró <nforro@redhat.com>
 2020-10-22 11:08:56 +02:00
langou c3a0b79af8 Merge pull request #455 from timleslie/fix-rotmg-bug 
[BLAS] Fix missing logic path in xrotmg.f
 2020-10-17 23:21:21 -06:00
Tim Leslie 6829744a23 Add inline comments with reference to literature 2020-10-18 15:51:43 +11:00
Ye Luo 3c7b0274fe Change the way of adding CMake_Fortran_FLAGS. 2020-10-17 22:20:50 -05:00
Tim Leslie 5f2b1aa019 [BLAS] Fix missing logic path in xrotmg.f 2020-10-17 16:58:00 +11:00
langou 4d1afcdec1 Merge pull request #454 from timleslie/rotg-opt 
[BLAS] Skip calculation of ROE when SCALE is zero
 2020-10-16 21:33:13 -06:00
Tim Leslie 187222c3f5 [BLAS] Skip calculation of ROE when SCALE is zero 2020-10-17 10:12:21 +11:00
langou ccbf01bdda Merge pull request #453 from timleslie/rotg-docs 
Mark xrotg() parameters as [in,out]
 2020-10-14 10:36:37 -05:00
langou c35b722cf2 Merge pull request #452 from timleslie/use-zx-in-zdrot 
Use ZX, ZY rather than CX, CY in ZDROT
 2020-10-14 10:34:59 -05:00
langou 16e1fb9961 Merge pull request #451 from timleslie/fix-dzasum-docs 
[DOCS] Fix typo in docs of dzasum.f
 2020-10-14 10:33:57 -05:00
Tim Leslie 9043ea2a84 Mark xrotg() parameters as [in,out] 2020-10-14 21:02:40 +11:00
Tim Leslie 74288e99fa Use ZX, ZY rather than CX, CY in ZDROT 2020-10-14 17:34:38 +11:00
Tim Leslie 7967e4f993 [DOCS] Fix typo in docs of dzasum.f 2020-10-14 17:21:44 +11:00
Martin Kroeker 122506cd8b Merge pull request #449 from timleslie/remove-ncola 
[BLAS] Remove unused NCOLA variable
 2020-09-29 22:30:42 +02:00
Martin Kroeker 4e71ce7d52 Merge pull request #445 from thijssteel/fix-dlanv2 
fix dlanv2 and slanv2
 2020-09-27 22:09:23 +02:00
Tim Leslie 50781b7649 Remove unused NCOLA variable 2020-09-27 09:42:32 +10:00
Martin Kroeker a0b221157a Merge pull request #447 from martin-frbg/issue430 
Add make.inc template for NVIDIA HPC compiler
 2020-09-22 23:39:17 +02:00
Martin Kroeker a0e1ccd9fc Merge pull request #448 from martin-frbg/issue416 
Improve prototype for LAPACKE_get_nancheck
 2020-09-22 23:38:25 +02:00
Martin Kroeker 80b0a2b5ab Improve prototype for LAPACKE_get_nancheck 2020-09-22 23:07:01 +02:00
Martin Kroeker 5a623e57b7 Comment PREFIX and LAPACKSOLIB entries to avoid surprises 2020-09-22 22:20:28 +02:00
Martin Kroeker 12a6a8afc2 Add make.inc template for NVIDIA HPC compiler 
from issue #430
 2020-09-22 21:56:01 +02:00
Martin Kroeker 2be5a09cbd Merge pull request #427 from martin-frbg/lintest-tau 
Fix out-of-bounds accesses in LIN test
 2020-09-21 13:37:11 +02:00
Martin Kroeker 5ade1b2fee Merge pull request #338 from martin-frbg/EIG-testing 
Avoid out-of-bounds accesses in complex EIG tests
 2020-09-21 13:35:43 +02:00
thijs d26c03955f fix-dlanv2 2020-09-18 11:22:57 +02:00
Martin Kroeker b4b56a5839 Merge pull request #444 from martin-frbg/travisosx 
Fix OSX builds on Travis CI
 2020-09-17 18:22:22 +02:00
Martin Kroeker 9861f27c0c Update .travis.yml 2020-09-17 11:59:07 +02:00
Martin Kroeker 692b7e5a8c Fix OSX build (gcc9 already installed now, so upgrade to gcc10) 2020-09-17 10:48:51 +02:00
Martin Kroeker 67ab142be0 Merge pull request #4 from Reference-LAPACK/master 
rebase
 2020-09-17 10:46:22 +02:00
Martin Kroeker 0148a3625c Merge pull request #443 from thijssteel/fix-xgetsls 
fix workspace query in xgelq and xgetsls
 2020-09-17 00:19:39 +02:00
Thijs Steel 6ffd9ec70d fix workspace query in dgelq 2020-09-16 12:14:21 +02:00
julielangou cc6e21e3c6 Merge pull request #438 from h-vetinari/backport_openblas_2778 
Fix some incorrect SLASET/DLASET calls (includes backport of openblas#2778)
 2020-09-07 18:21:04 -07:00
H. Vetinari bffac66e45 Fix bounds-violations in DLASET-calls 
The last parameter must satisfy `LDA >= max(1,M)`, where M is the
number of rows of the matrix, see
https://github.com/Reference-LAPACK/lapack/blob/v3.9.0/SRC/dlaset.f#L95

Suggested-By: Serguei Patchkovskii <Serguei.Patchkovskii@gmail.com>
 2020-09-06 11:29:13 +02:00
H. Vetinari 94994eb162 Fix more bounds-violations in SLASET-calls 
The last parameter must satisfy `LDA >= max(1,M)`, where M is the
number of rows of the matrix, see
https://github.com/Reference-LAPACK/lapack/blob/v3.9.0/SRC/slaset.f#L95

Suggested-By: Serguei Patchkovskii <Serguei.Patchkovskii@gmail.com>
 2020-09-06 11:28:44 +02:00
H. Vetinari a5bfdf80d6 Fix bounds-violation in SLASET 
These have been grandfathered in by compilers based on an ancient
version of the language definition, but would invoke undefined
behaviour per the standard. For more details, see the following:
https://github.com/Reference-LAPACK/lapack/pull/438#issuecomment-687560690

Suggested-By: Serguei Patchkovskii <Serguei.Patchkovskii@gmail.com>
 2020-09-06 10:53:52 +02:00
Martin Kroeker 7744f9e08a Fix incorrect argument to SLASET 
Reference-LAPACK issue 425 (and 318)

Suggested-By: Serguei Patchkovskii <Serguei.Patchkovskii@gmail.com>
 2020-09-04 18:31:03 +02:00
Martin Kroeker 7d8e2f0924 Fix incorrect argument to SLASET 
Reference-LAPACK issue 425 (and 318)

Suggested-By: Serguei Patchkovskii <Serguei.Patchkovskii@gmail.com>
 2020-09-04 18:30:56 +02:00
Martin Kroeker 0bba1319ed Fix incorrect calls to DLASET 
Reference-LAPACK issue 429

Suggested-By: Serguei Patchkovskii <Serguei.Patchkovskii@gmail.com>
 2020-09-04 18:29:52 +02:00
langou 9b136e860a Merge pull request #437 from h-vetinari/ggsv_followup 
?ggsv? followup to #434 & #409
 2020-09-04 07:57:44 -06:00
H. Vetinari 5ae1561e8e adapt ?ggsv?-functions to ambient code style in LAPACKE/include/lapack.h 2020-09-03 11:40:46 +02:00
H. Vetinari cc04b61950 Follow-up to #434 & #409: add missing 'const' in signatures 
Based on how the surrounding functions in lapack.h are handling the
parameters, particularly the ?ggsv?3-variants of the affected functions
 2020-09-03 11:37:37 +02:00
H. Vetinari d470f435d2 Follow-up to #434 & #409: fix signature mismatches 2020-09-02 22:42:23 +02:00
langou 0bc3070bac Merge pull request #436 from h-vetinari/fix 
Fix some minor inconsistencies in LAPACK(E)_[cz]gesvdq
 2020-09-02 12:50:08 -06:00
H. Vetinari d168b4d2ae Fix some minor inconsistencies in LAPACK(E)_[cz]gesvdq 
For LAPACKE_cgesvdq, align datatype of `rwork` with lapacke.h and callsite
in LAPACKE_cgesvdq_work; both of the latter use `float* rwork`, so there's
no need to calculate with doubles that only get downcast anyway.

Similarly, make the signature of lapacke_zgesvdq_work consistent across
lapack.h (modified here), the fortran side and lapacke_zgesvdq_work, see:
https://github.com/Reference-LAPACK/lapack/blob/v3.9.0/LAPACKE/src/lapacke_zgesvdq_work.c#L42
https://github.com/Reference-LAPACK/lapack/blob/v3.9.0/SRC/zgesvdq.f#L422
 2020-09-02 20:39:29 +02:00
julielangou fef53088d9 Merge pull request #434 from martin-frbg/lapacke_lapack_h 
LAPACKE/include/lapack.h: fix C++ compatibility and prototypes for xGGSVD/xGGSVP
 2020-09-01 12:08:37 -07:00
Martin Kroeker 14740d3aa3 Correct prototypes for LAPACK_?ggsvd and LAPACK_?ggsvp 2020-09-01 11:12:56 +02:00
Martin Kroeker a69969d5e1 Merge pull request #3 from Reference-LAPACK/master 
rebase
 2020-09-01 11:11:02 +02:00
langou 773eb18b31 Merge pull request #408 from ilayn/patch-1 
Fix the loop index scalar in ZHEEQUB.f
 2020-08-17 08:39:27 -06:00
langou 75611e79fe Merge pull request #433 from serguei-patchkovskii/issue428 
Update TWOPI and PIOVER2 to 39 significant digits (issue #428). Added make.inc for gfortran in quadruple precision
 2020-08-07 16:12:19 -06:00
Serguei Patchkovskii 9db9ff8556 Updated values of TWOPI and PIOVER2 to 39 significant digits, 
to prevent accuracy loss when REAL or DOUBLE PRECISION types
correspond to an extended precision real kind. These constants
should be sufficient to work with 128-bit extended-IEEE data
type.

Rebuilding the modified source using the default gfortran make.inc
(with gfortran 7.5.0) does not add any new test failures. The
only test failure, which is also present in the master baranch, is:

 ZDRGEV3: ZGGEV31 returned INFO=     7.
   N=    12, JTYPE=    16, ISEED=( 2855, 3551, 2406, 2393)

Additionally, included a sample build file for quadruple-precision
compilation using gfortran (version 4.7 or later). With this
config file, both REAL and DOUBLE PRECISION are mapped onto
a 128-bit REAL(16) type, with the corresponding change to the
COMPLEX and COMPLEX*16.

There are numerical failure 6 failures for each type, all in
*GBSVX type 6 matrix, with S/D routines losing 4 significant
digits, and C/Z routines 5 digits relative to the expected
result. I am not really qualified to debug this failure ...
 2020-08-07 16:05:22 +02:00
langou f53dd49bd3 Merge pull request #431 from martin-frbg/issue426 
LIN testsuite: Correct arguments used in workspace queries and (C,Z)SCAL call
 2020-08-06 11:55:09 -06:00