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
107 Commits
Author SHA1 Message Date
Mark Gates d0b1591e28 fix PROGRAM line in INSTALL 2025-08-26 23:27:58 -04:00
Mark Gates 594dc253ae implicit none in INSTALL 2025-08-26 23:27:58 -04:00
langou cf2c4402e9 Merge pull request #1095 from julielangou/master 
Updating LAPACK version to 3.12.1
 2025-01-13 01:08:10 -07:00
Julie 3e242fc8d5 Updating LAPACK version to 3.12.1 2025-01-12 14:42:24 -08:00
Simon Maertens 0af92d8230 Bump minimum CMake version to 3.13 2024-06-05 13:13:38 +02:00
Simon Maertens 231db955f2 Use add_compile_options(...) and add_link_options(...) instead of appending flags to CACHE variables. 2024-06-05 13:12:09 +02:00
Maria Kraynyuk 40ca7a2713 Include lapack_64.h only during the build 2024-02-05 10:54:04 -08:00
Maria Kraynyuk 3ccdccea9b LAPACK SRC long lines updated manually 2024-02-05 10:28:26 -08:00
Maria Kraynyuk b891145899 Add extended API with _64 suffix to LAPACK with lapack_64.h 2024-02-05 10:28:08 -08:00
Julie d7be813203 LAPACK 3.12.0 release 
Update documentation for coming release
 2023-11-24 12:36:54 -08:00
scivision 0a52d71f09 LAPACK CMakeLists.txt was using features from CMake 3.6 already 
CMake 3.27 warns for cmake_minimum_required < 3.5
 2023-10-12 15:42:13 -04:00
Mark Gates 62b19990a2 doxygen: new group structure 2023-07-04 01:29:12 -04:00
Weslley S Pereira 35b7a4b5d3 Use a better (more informative) message when we say the compiler intrinsics fail 2023-05-30 11:53:40 -06:00
Weslley da Silva Pereira 57f15d84c1 Change ILAVER and Doxyfile accordingly 2022-11-10 10:43:17 -07:00
langou def1271eda Merge pull request #623 from weslleyspereira/try-test-fotrancompiler_abs_div_operations 
Test the Fortran intrinsic ABS and complex divisions and report fails during build
 2022-04-25 21:25:00 -06:00
Weslley da Silva Pereira e7ad0370d4 Preparing software to release 3.10.1 2022-04-12 14:11:40 -06:00
Weslley S. Pereira 1dd9548c4d Adding tests with intrinsic MIN, MAX and complex multiplication 2022-01-03 16:09:55 -07:00
Weslley S. Pereira f461d4f03b Fixes bugs. Updates documentation. Tests for ABS and complex division run on the Makefile build 2021-09-22 14:55:27 -06:00
Weslley S. Pereira 19fafd4fd2 Test the Fortran intrinsic ABS and complex divisions and report during compile time. Currently only working with the CMake build! 2021-09-22 14:20:50 -06:00
Quellyn Snead d2781152e1 Address issues with IBM XL builds (#606). 
Be more explicit about selecting recursive fortran flags, as not
every compiler will correctly reject an incorrect option (e.g., XL).
 2021-09-15 11:14:42 -06:00
Sébastien Villemot c657043a2b make.inc.gfortran: define a default value for DOCSDIR 2021-09-01 14:15:24 +02:00
weslley.spereira c7400e5f9b Improve documentation of the new routines. Force roundup only if it is necessary 2021-08-04 18:43:22 -03:00
weslley.spereira cea95686f5 Fix bug: ROUNDUP_LWORK must use (1+EPSILON(0)) instead of (1+DLAMCH(EPS)) 
Test results:

 SROUNDUP_LWORK(0)        =    0.00000000
 SROUNDUP_LWORK(1)        =    1.00000000
 SROUNDUP_LWORK(317)      =    317.000000
 DROUNDUP_LWORK(0)        =    0.0000000000000000
 DROUNDUP_LWORK(1)        =    1.0000000000000000
 DROUNDUP_LWORK(317)      =    317.00000000000000

 Tests with X = 16777216 = 2**24
X-2 = 16777214; SROUNDUP_LWORK(X-2) = 16777214.0
X-1 = 16777215; SROUNDUP_LWORK(X-1) = 16777215.0
X   = 16777216; SROUNDUP_LWORK(X  ) = 16777218.0
X+1 = 16777217; SROUNDUP_LWORK(X+1) = 16777218.0
X+2 = 16777218; SROUNDUP_LWORK(X+2) = 16777220.0

 Tests with X = 9007199254740992 = 2**53
X-2 = 9007199254740990; SROUNDUP_LWORK(X-2) = 9007200328482816.0
X-1 = 9007199254740991; SROUNDUP_LWORK(X-1) = 9007200328482816.0
X   = 9007199254740992; SROUNDUP_LWORK(X  ) = 9007200328482816.0
X+1 = 9007199254740993; SROUNDUP_LWORK(X+1) = 9007200328482816.0
X+2 = 9007199254740994; SROUNDUP_LWORK(X+2) = 9007200328482816.0

 Tests with X = 9007199254740992 = 2**53
X-2 = 9007199254740990; DROUNDUP_LWORK(X-2) = 9007199254740990.0
X-1 = 9007199254740991; DROUNDUP_LWORK(X-1) = 9007199254740991.0
X   = 9007199254740992; DROUNDUP_LWORK(X  ) = 9007199254740994.0
X+1 = 9007199254740993; DROUNDUP_LWORK(X+1) = 9007199254740994.0
X+2 = 9007199254740994; DROUNDUP_LWORK(X+2) = 9007199254740996.0
 2021-07-27 20:05:07 -03:00
weslley.spereira f12515a491 RoundUp functions now always returns an integer 2021-07-27 16:56:21 -03:00
weslley.spereira a620672e29 Add (s,d)ROUNDUP_LWORK functions to fix the workspace computation 2021-07-26 19:30:53 -03:00
julielangou 3aa00e63fc Updating LAPACK version number before 3.10.0 release 
Updating LAPACK version number before 3.10.0 release
 2021-06-28 09:13:34 -07:00
Julie e18d437924 Making LAPACK's code eternal... no more version and date in source files. 
GitHub is now enabling us to track accurately version and date.
No need for this anymore.
 2021-03-25 10:16:58 -07:00
Julie b84fbaf156 Merge remote-tracking branch 'upstream/master' 2021-03-23 10:35:58 -07:00
Julie f67034373e Update LAPACK version number to 3.9.1 
... waiting on decision for modified source files
 2021-03-18 09:54:00 -07:00
Martin Kroeker 4ba73aa2aa Add sample configuration file for the NAG Fortran compiler 2021-03-14 11:27:21 +01:00
weslley.spereira 96880da294 Fix typo and use o instead of .f in the INSTALL/Makefile 2021-03-01 12:04:06 -03:00
Julien Langou 786127da71 fix issues #435 and #479 
Thanks to @weslleyspereira
 2021-02-26 11:26:04 -07:00
Julien Schueller 8f004b353a CMake: Bump minimum version to 3.2 
Newer cmake issues a warning when required version is < 3.
Version 3.2 should be old enough (eg ubuntu 16.04 has version 3.5).
 2021-01-08 17:50:49 +01:00
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
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
mvdb e2700c72c4 Added -Mrecursive option to pgf95 configuration 
This is done for similar reasons as the '-frecursive'
option in the gfortran configuration file (see e.g. bug112)
 2020-03-31 15:53:21 -04:00
mvdb 0ec7f11be1 Added -recursive option to ifort configuration 
This is done for similar reasons as the '-frecursive'
option in the gfortran configuration file (see e.g. bug112)
 2020-03-30 22:15:59 +02:00
Julie 6acc99d5f3 Updating Version Number to 3.9.0 2019-11-21 08:57:43 +01:00
Kyle Guinn dd2f632f2e Add PHONY targets 2019-07-27 23:07:37 -05:00
Kyle Guinn be23965929 Handle relative paths consistently 
Place built libraries in $(TOPSRCDIR) and refer to their location by
using $(TOPSRCDIR).
 2019-07-27 18:48:52 -05:00
Kyle Guinn 33258cb4b2 Switch to POSIX make variable names 
Allows us to use the built-in suffix rules where appropriate.

  FORTRAN   -> FC            (POSIX)
  OPTS      -> FFLAGS        (POSIX)
  DRVOPTS   -> FFLAGS_DRV
  NOOPT     -> FFLAGS_NOOPT

  LOADER    -> FC            (use the compiler driver for linking)
  LOADOPTS  -> LDFLAGS       (POSIX)

  ARCH      -> AR            (POSIX)
  ARCHFLAGS -> ARFLAGS       (POSIX)
 2019-07-27 18:48:52 -05:00
Julien Langou 8a2b4dce36 changing ILAVER to 3.9.0 in view of imminent release 2019-06-26 07:09:11 +02:00
Mark Gates 550c290bb9 fix docs: add types in ilaver, lamch 2018-11-03 22:09:57 -04:00
David Wells 85747adb25 fix docs: missing character argument 2018-05-29 13:32:28 -04:00
Julie d97a30482e Updating Version number on files for 3.8.0 2017-11-12 20:15:12 -08:00
julielangou 6bcfa0d632 Merge pull request #188 from zerothi/recursive 
Reinserted frecursive due to prior bug-reports
 2017-11-05 10:22:36 -08:00
Nick Papior dc0e59aff5 Fixed (alot) missing external subroutines and removed "dead" variables 
This commit is purely of cleaning up missing externals. Currently there
is no check on EXTERNAL's not used.

Signed-off-by: Nick Papior <nickpapior@gmail.com>
 2017-11-05 07:47:18 +01:00
Nick Papior 5afcdaf000 Reinserted frecursive due to prior bug-reports. 
See 3a87b2df45 for details.

Signed-off-by: Nick Papior <nickpapior@gmail.com>
 2017-10-12 20:38:59 +02:00
Nick Papior 7b651690c7 Added ilaenv2stage -> circumvents requirement of recursive compilation 
Added ilaenv2stage.f to enable the extraction of 2stage machine dependent
variables from the 2stage solver without having to rely on compiling the
entire library with the recursive flag.

Secondly, all calls to ILAENV( 17 <= 21, ...) have been fixed to
use the corresponding ILAENV2STAGE routine by translating ISPEC by -16
I.e.

  ILAENV( 17, ...) == ILAENV2STAGE( 1, ...)

Also fixed the documentation of ilaenv.f to specify the cases of ISPEC > 16.

Signed-off-by: Nick Papior <nickpapior@gmail.com>
 2017-07-20 13:26:56 +02:00