platform/upstream/openblas.git
10 years agomodified common.h
wernsaar [Sun, 1 Dec 2013 15:11:13 +0000 (16:11 +0100)]
modified common.h

10 years agochanged level3.c
wernsaar [Sun, 1 Dec 2013 12:49:59 +0000 (13:49 +0100)]
changed level3.c

10 years agoRefs #283. Fixed the incorrect usage of long data type for Windows 64.
Zhang Xianyi [Thu, 14 Nov 2013 05:46:42 +0000 (13:46 +0800)]
Refs #283. Fixed the incorrect usage of long data type for Windows 64.

10 years agoRefs #307. Fixed the hang bug when free OpenBLAS dll in Windows.
Zhang Xianyi [Wed, 13 Nov 2013 02:00:18 +0000 (10:00 +0800)]
Refs #307. Fixed the hang bug when free OpenBLAS dll in Windows.

10 years agoAvoid argument list too long issue in make clean.
Zhang Xianyi [Thu, 7 Nov 2013 05:06:42 +0000 (13:06 +0800)]
Avoid argument list too long issue in make clean.

10 years agoRefs #314. Fixed clang compiling bug on OSX.
Zhang Xianyi [Thu, 7 Nov 2013 00:12:03 +0000 (08:12 +0800)]
Refs #314. Fixed clang compiling bug on OSX.

10 years agoRefs #309. Fixed trtri_U single thread computational bug.
Zhang Xianyi [Wed, 6 Nov 2013 17:08:39 +0000 (01:08 +0800)]
Refs #309. Fixed trtri_U single thread computational bug.

10 years agoFixed #315. Added OPENBLAS_ prefix to openblas_config.h.
Zhang Xianyi [Sat, 2 Nov 2013 07:59:00 +0000 (15:59 +0800)]
Fixed #315. Added OPENBLAS_ prefix to openblas_config.h.

10 years agorefs #287. Don't enable OpenMP for netlib LAPACK sequential Fortran codes.
Zhang Xianyi [Sat, 2 Nov 2013 07:09:33 +0000 (15:09 +0800)]
refs #287. Don't enable OpenMP for netlib LAPACK sequential Fortran codes.

10 years agoMerge pull request #311 from loladiro/patch-1
Zhang Xianyi [Fri, 25 Oct 2013 06:41:22 +0000 (23:41 -0700)]
Merge pull request #311 from loladiro/patch-1

Use FC instead of CC to link the dynamic library on OS X

10 years agoUse FC instead of CC to link the dynamic library on OS X
Keno Fischer [Wed, 23 Oct 2013 22:43:00 +0000 (18:43 -0400)]
Use FC instead of CC to link the dynamic library on OS X

Avoids problems of libgfortran not being found.

10 years agoAdded gfortran dependency for LSB/lsbcc.
Zhang Xianyi [Tue, 22 Oct 2013 05:24:47 +0000 (13:24 +0800)]
Added gfortran dependency for LSB/lsbcc.

10 years agoFixed buffer overflow bug in kernel/x86_64/dgemv_t.S file.
wangqian [Thu, 10 Oct 2013 19:20:20 +0000 (03:20 +0800)]
Fixed buffer overflow bug in kernel/x86_64/dgemv_t.S file.

10 years agoMerge pull request #301 from yieldthought/develop
Zhang Xianyi [Wed, 9 Oct 2013 07:46:49 +0000 (00:46 -0700)]
Merge pull request #301 from yieldthought/develop

Remove -Wl,--retain-symbols-file from dynamic link line to fix tool support

10 years agoRemove -Wl,--retain-symbols-file from dynamic library linking to fix tool support
yieldthought [Tue, 8 Oct 2013 14:37:17 +0000 (16:37 +0200)]
Remove -Wl,--retain-symbols-file from dynamic library linking to fix tool support

The aim is to restrict the symbols exported from openBLAS, but for dynamic libraries --retain-symbols-file has the opposite effect intended. It removes symbols from the .symtab section of the library, which is used by tools such as debuggers, profilers, objdump etc. but does not remove them from .dynsym, which is the section used by the runtime when loading the binary.

In short, using --retain-symbols-file like this doesn't prevent symbol leakage but does prevent tools from analyzing applications linked with openBLAS.

This patch improves tool support for openBLAS applications but does not address symbol leakage through .dynsym.

10 years agoRefs #262. Fixed compatibility issues of GNU stack markings with PathScale EKOPath...
Zhang Xianyi [Sun, 22 Sep 2013 01:37:59 +0000 (09:37 +0800)]
Refs #262. Fixed compatibility issues of GNU stack markings with PathScale EKOPath(tm) Compiler Suite: Version 4.0.12.1

11 years agoAdded backers.
Zhang Xianyi [Thu, 5 Sep 2013 07:39:45 +0000 (15:39 +0800)]
Added backers.

11 years agoMerge pull request #290 from larsmans/missing-threshold
Lars Buitinck [Wed, 28 Aug 2013 15:20:16 +0000 (17:20 +0200)]
Merge pull request #290 from larsmans/missing-threshold

check if GEMM_MULTITHREAD_THRESHOLD defined in gemm.c
Set a fallback value.

11 years agoMerge pull request #291 from larsmans/fix-makefile-prefix
Zhang Xianyi [Wed, 28 Aug 2013 16:26:16 +0000 (09:26 -0700)]
Merge pull request #291 from larsmans/fix-makefile-prefix

fix default prefix handling in makefiles

11 years agoMerge pull request #289 from larsmans/no-noconst
Zhang Xianyi [Wed, 28 Aug 2013 16:25:23 +0000 (09:25 -0700)]
Merge pull request #289 from larsmans/no-noconst

get rid of the generated cblas_noconst.h file

11 years agofix default prefix handling in makefiles
Lars Buitinck [Wed, 28 Aug 2013 15:39:54 +0000 (17:39 +0200)]
fix default prefix handling in makefiles

PREFIX wasn't communicated to Makefile.install (where it matters)
by Makefile. The result is that the default PREFIX is empty and
OpenBLAS was being installed in /lib.

11 years agoget rid of the generated cblas_noconst.h file
Lars Buitinck [Wed, 28 Aug 2013 14:52:24 +0000 (16:52 +0200)]
get rid of the generated cblas_noconst.h file

11 years agoMerge pull request #288 from sebastien-villemot/develop
Zhang Xianyi [Wed, 28 Aug 2013 13:26:37 +0000 (06:26 -0700)]
Merge pull request #288 from sebastien-villemot/develop

Avoid failure on qemu guests declaring an Athlon CPU without 3dnow!

11 years agoAvoid failure on qemu guests declaring an Athlon CPU without 3dnow!
Sébastien Villemot [Wed, 28 Aug 2013 12:27:59 +0000 (14:27 +0200)]
Avoid failure on qemu guests declaring an Athlon CPU without 3dnow!

The present patch verifies that, on machines declaring an Athlon CPU model and
family, the 3dnow and 3dnowext feature flags are indeed present. If they are
not, it fallbacks on the most generic x86 kernel. This prevents crashes due to
illegal instruction on qemu guests with a weird configuration.

Closes #272

11 years agoMerge branch 'bulldozer' into develop
Zhang Xianyi [Sat, 24 Aug 2013 14:46:18 +0000 (11:46 -0300)]
Merge branch 'bulldozer' into develop

11 years agoRefs #281. Detect __CYGWIN__ macro for Cygwin x86_64.
Zhang Xianyi [Sat, 24 Aug 2013 05:09:49 +0000 (13:09 +0800)]
Refs #281. Detect __CYGWIN__ macro for Cygwin x86_64.

Signed-off-by: Zhang Xianyi <traits.zhang@gmail.com>
11 years agoRefs #281. Detect _WIN32 macro for Windows API.
Zhang Xianyi [Fri, 23 Aug 2013 17:10:02 +0000 (01:10 +0800)]
Refs #281. Detect _WIN32 macro for Windows API.

http://www.mail-archive.com/bug-gnulib@gnu.org/msg05722.html

11 years agoremoved unnessesary instructions from zgemm_kernel_2x2_bulldozer.S
wernsaar [Sat, 17 Aug 2013 04:46:17 +0000 (06:46 +0200)]
removed unnessesary instructions from zgemm_kernel_2x2_bulldozer.S

11 years agoremoved unnessesary instructions
wernsaar [Fri, 16 Aug 2013 18:23:34 +0000 (20:23 +0200)]
removed unnessesary instructions

11 years agoRefs #282. Fixed zgemv_n typo bug on Win64.
Zhang Xianyi [Fri, 23 Aug 2013 08:27:17 +0000 (16:27 +0800)]
Refs #282. Fixed zgemv_n typo bug on Win64.

11 years agoMerge pull request #280 from ViralBShah/develop
Zhang Xianyi [Wed, 21 Aug 2013 15:21:51 +0000 (08:21 -0700)]
Merge pull request #280 from ViralBShah/develop

Patch LAPACK XLASD4.f as discussed in JuliaLang/julia#2340

11 years agoPatch LAPACK XLASD4.f as discussed in JuliaLang/julia#2340
Viral B. Shah [Wed, 21 Aug 2013 13:44:07 +0000 (19:14 +0530)]
Patch LAPACK XLASD4.f as discussed in JuliaLang/julia#2340

11 years agoRefs #279. Provide ONLY_CBLAS flag. If you only need CBLAS without
Zhang Xianyi [Tue, 20 Aug 2013 16:03:25 +0000 (00:03 +0800)]
Refs #279. Provide ONLY_CBLAS flag. If you only need CBLAS without
a fortran compiler, please try make ONLY_CBLAS=1.

This mode only compiler CBLAS without BLAS fortran interface and LAPACK.

11 years agoMerge branch 'bulldozer' into develop
Zhang Xianyi [Mon, 12 Aug 2013 15:22:10 +0000 (23:22 +0800)]
Merge branch 'bulldozer' into develop

11 years agoFixed #276. Merge branch 'wernsaar-develop' into bulldozer
Zhang Xianyi [Fri, 9 Aug 2013 02:49:44 +0000 (10:49 +0800)]
Fixed #276. Merge branch 'wernsaar-develop' into bulldozer

11 years agoMerge branch 'develop' of https://github.com/wernsaar/OpenBLAS into wernsaar-develop
Zhang Xianyi [Fri, 9 Aug 2013 02:48:46 +0000 (10:48 +0800)]
Merge branch 'develop' of https://github.com/wernsaar/OpenBLAS into wernsaar-develop

11 years agomodified KERNEL.BULLDOZER
wernsaar [Thu, 8 Aug 2013 15:49:30 +0000 (17:49 +0200)]
modified KERNEL.BULLDOZER

11 years agoadded dtrsm_kernel_RN_8x2_bulldozer.S
wernsaar [Thu, 8 Aug 2013 05:14:08 +0000 (07:14 +0200)]
added dtrsm_kernel_RN_8x2_bulldozer.S

11 years agodtrsm_kernel_LT_8x2_bulldozer.S performance optimization
wernsaar [Mon, 5 Aug 2013 09:27:16 +0000 (11:27 +0200)]
dtrsm_kernel_LT_8x2_bulldozer.S performance optimization

11 years agoRefs #270 #268. Merge branch 'wernsaar-develop' into bulldozer
Zhang Xianyi [Mon, 5 Aug 2013 08:17:15 +0000 (16:17 +0800)]
Refs #270 #268. Merge branch 'wernsaar-develop' into bulldozer

11 years agoMerge branch 'develop' of https://github.com/wernsaar/OpenBLAS into wernsaar-develop
Zhang Xianyi [Mon, 5 Aug 2013 08:09:47 +0000 (16:09 +0800)]
Merge branch 'develop' of https://github.com/wernsaar/OpenBLAS into wernsaar-develop

11 years agoEnable bulldozer kernels.
Zhang Xianyi [Mon, 5 Aug 2013 08:07:54 +0000 (16:07 +0800)]
Enable bulldozer kernels.

11 years agoMerge branch 'develop' into bulldozer
Zhang Xianyi [Mon, 5 Aug 2013 07:51:53 +0000 (15:51 +0800)]
Merge branch 'develop' into bulldozer

11 years agomodified dtrsm_kernel_LT_8x2_bulldozer.S
wernsaar [Sun, 4 Aug 2013 10:16:12 +0000 (12:16 +0200)]
modified dtrsm_kernel_LT_8x2_bulldozer.S

11 years agomodified dtrsm_kernel_LT_8x2_bulldozer.S
wernsaar [Sun, 4 Aug 2013 08:15:33 +0000 (10:15 +0200)]
modified dtrsm_kernel_LT_8x2_bulldozer.S

11 years agoadded dtrsm_kernel_LT_8x2_bulldozer.S
wernsaar [Sun, 4 Aug 2013 07:54:40 +0000 (09:54 +0200)]
added dtrsm_kernel_LT_8x2_bulldozer.S

11 years agoremoved dtrsm_kernel_LT_8x2_bulldozer.S
wernsaar [Sat, 3 Aug 2013 13:40:51 +0000 (15:40 +0200)]
removed dtrsm_kernel_LT_8x2_bulldozer.S

11 years agofixed bug in dgemv_t_bulldozer.S
wernsaar [Sat, 3 Aug 2013 10:19:29 +0000 (12:19 +0200)]
fixed bug in dgemv_t_bulldozer.S

11 years agorepaired trmm bug in sgemm_kernel_16x2_bulldozer.S
wernsaar [Sat, 3 Aug 2013 09:43:25 +0000 (11:43 +0200)]
repaired trmm bug in sgemm_kernel_16x2_bulldozer.S

11 years agorepaired trmm bug in cgemm_kernel_4x2_bulldozer.S
wernsaar [Sat, 3 Aug 2013 08:32:51 +0000 (10:32 +0200)]
repaired trmm bug in cgemm_kernel_4x2_bulldozer.S

11 years agorepaired trmm bug in zgemm_kernel_2x2_bulldozer.S
wernsaar [Sat, 3 Aug 2013 08:17:08 +0000 (10:17 +0200)]
repaired trmm bug in zgemm_kernel_2x2_bulldozer.S

11 years agorepaired trmm bug in dgemm_kernel_8x2_bulldozer.S
wernsaar [Sat, 3 Aug 2013 07:35:39 +0000 (09:35 +0200)]
repaired trmm bug in dgemm_kernel_8x2_bulldozer.S

11 years agoMerge branch 'hotfix-v0.2.8' into develop
Zhang Xianyi [Thu, 1 Aug 2013 15:57:19 +0000 (23:57 +0800)]
Merge branch 'hotfix-v0.2.8' into develop

11 years agoUpdate the doc for 0.2.8 version.
Zhang Xianyi [Thu, 1 Aug 2013 15:52:43 +0000 (23:52 +0800)]
Update the doc for 0.2.8 version.

11 years agoOpenBLAS 0.2.8 rc1.
Zhang Xianyi [Wed, 31 Jul 2013 06:49:16 +0000 (14:49 +0800)]
OpenBLAS 0.2.8 rc1.

11 years agoMerge branch 'hotfix-v0.2.8' into develop
Zhang Xianyi [Wed, 31 Jul 2013 06:46:56 +0000 (14:46 +0800)]
Merge branch 'hotfix-v0.2.8' into develop

11 years agoRefs #266. Fixed the compiling bug with Open64 5.0.
Zhang Xianyi [Wed, 31 Jul 2013 06:41:39 +0000 (14:41 +0800)]
Refs #266. Fixed the compiling bug with Open64 5.0.

11 years agoadded generic trmm kernels and modified Makefile.L3
wernsaar [Tue, 30 Jul 2013 18:18:57 +0000 (20:18 +0200)]
added generic trmm kernels and modified Makefile.L3

11 years agoFixed #264 the memory leak bug in dtrtri_U.
Zhang Xianyi [Mon, 29 Jul 2013 15:21:10 +0000 (23:21 +0800)]
Fixed #264 the memory leak bug in dtrtri_U.

11 years agoFixed the FMA3 detection bug.
Zhang Xianyi [Sat, 27 Jul 2013 14:37:57 +0000 (22:37 +0800)]
Fixed the FMA3 detection bug.

11 years agoFixed #261. Use strncmp instead of a comparing trick.
Zhang Xianyi [Fri, 26 Jul 2013 15:43:54 +0000 (23:43 +0800)]
Fixed #261. Use strncmp instead of a comparing trick.

11 years agoFixed typo in getarch_2nd.c.
Zhang Xianyi [Mon, 29 Jul 2013 07:42:00 +0000 (15:42 +0800)]
Fixed typo in getarch_2nd.c.

11 years agoadded dtrsm_kernel_LT_8x2_bulldozer.S
wernsaar [Sun, 28 Jul 2013 14:47:58 +0000 (16:47 +0200)]
added dtrsm_kernel_LT_8x2_bulldozer.S

11 years agoRefs #263. Rollback bulldozer and piledriver kernels to barcelona kernels.
Zhang Xianyi [Sun, 28 Jul 2013 09:39:24 +0000 (17:39 +0800)]
Refs #263. Rollback bulldozer and piledriver kernels to barcelona kernels.

11 years agoMerge branch 'develop' into bulldozer
Zhang Xianyi [Sun, 28 Jul 2013 04:38:25 +0000 (06:38 +0200)]
Merge branch 'develop' into bulldozer

Conflicts:
kernel/x86_64/KERNEL.BULLDOZER

11 years agoRefs #262. Added executable stack markings.
Zhang Xianyi [Sat, 27 Jul 2013 16:09:40 +0000 (00:09 +0800)]
Refs #262. Added executable stack markings.

11 years agoMerge branch 'sfabbro-ldflags' into develop
Zhang Xianyi [Sat, 27 Jul 2013 15:03:07 +0000 (23:03 +0800)]
Merge branch 'sfabbro-ldflags' into develop

11 years agoFixed #260. Fixed generating 32-bit shared library on previous commit.
Zhang Xianyi [Sat, 27 Jul 2013 15:01:36 +0000 (23:01 +0800)]
Fixed #260. Fixed generating 32-bit shared library on previous commit.

11 years agoFixed the FMA3 detection bug.
Zhang Xianyi [Sat, 27 Jul 2013 14:37:57 +0000 (22:37 +0800)]
Fixed the FMA3 detection bug.

11 years agoMerge branch 'ldflags' of https://github.com/sfabbro/OpenBLAS into sfabbro-ldflags
Zhang Xianyi [Sat, 27 Jul 2013 14:19:54 +0000 (22:19 +0800)]
Merge branch 'ldflags' of https://github.com/sfabbro/OpenBLAS into sfabbro-ldflags

11 years agoFixed #261. Use strncmp instead of a comparing trick.
Zhang Xianyi [Fri, 26 Jul 2013 15:43:54 +0000 (23:43 +0800)]
Fixed #261. Use strncmp instead of a comparing trick.

11 years agoRespect user's LDFLAGS
Sebastien Fabbro [Wed, 24 Jul 2013 16:37:16 +0000 (09:37 -0700)]
Respect user's LDFLAGS

11 years agoMerge branch 'develop' v0.2.7
Zhang Xianyi [Thu, 25 Jul 2013 17:34:45 +0000 (01:34 +0800)]
Merge branch 'develop'

11 years agoRefs #259. Fixed missing LAPACK functions in shared library.
Zhang Xianyi [Thu, 25 Jul 2013 17:32:32 +0000 (01:32 +0800)]
Refs #259. Fixed missing LAPACK functions in shared library.

11 years agoMerge branch 'develop'
Zhang Xianyi [Tue, 23 Jul 2013 05:40:08 +0000 (13:40 +0800)]
Merge branch 'develop'

11 years agoMerge pull request #257 from staticfloat/develop
Zhang Xianyi [Tue, 23 Jul 2013 05:35:29 +0000 (22:35 -0700)]
Merge pull request #257 from staticfloat/develop

Add in return value for `interface/trtri.c`

11 years agoFix xianyi/OpenBLAS#256
Elliot Saba [Tue, 23 Jul 2013 00:02:06 +0000 (17:02 -0700)]
Fix xianyi/OpenBLAS#256

11 years agoRefs #255. Didn't use f77 compiler.
Zhang Xianyi [Mon, 22 Jul 2013 03:34:43 +0000 (11:34 +0800)]
Refs #255. Didn't use f77 compiler.

11 years agoUpdate CONTRIBUTORS.md
Zhang Xianyi [Sat, 20 Jul 2013 15:32:23 +0000 (23:32 +0800)]
Update CONTRIBUTORS.md

11 years agoMerge branch 'develop'
Zhang Xianyi [Sat, 20 Jul 2013 15:05:36 +0000 (23:05 +0800)]
Merge branch 'develop'

11 years agoFixed #253. Update doc for v0.2.7 version.
Zhang Xianyi [Sat, 20 Jul 2013 15:05:12 +0000 (23:05 +0800)]
Fixed #253. Update doc for v0.2.7 version.

11 years agoMerge branch 'loongson3b' into develop
Zhang Xianyi [Sat, 20 Jul 2013 14:33:35 +0000 (22:33 +0800)]
Merge branch 'loongson3b' into develop

11 years agoMerge branch 'loongson3a' into develop
Zhang Xianyi [Sat, 20 Jul 2013 14:32:38 +0000 (22:32 +0800)]
Merge branch 'loongson3a' into develop

Conflicts:
Makefile.system

11 years agoFixed #254. Added the date of changes in contributors file.
Zhang Xianyi [Sat, 20 Jul 2013 03:35:27 +0000 (11:35 +0800)]
Fixed #254. Added the date of changes in contributors file.

11 years agocreate contributor file.
Zhang Xianyi [Fri, 19 Jul 2013 00:38:03 +0000 (08:38 +0800)]
create contributor file.

11 years agoFixed a computational error in zgemm_kernel_4x4_sandy.S file.
wangqian [Thu, 18 Jul 2013 12:23:21 +0000 (20:23 +0800)]
Fixed a computational error in zgemm_kernel_4x4_sandy.S file.

11 years agoEnsure the correct stack alignment on Win32.
Zhang Xianyi [Wed, 17 Jul 2013 07:19:07 +0000 (15:19 +0800)]
Ensure the correct stack alignment on Win32.

11 years agoFixed typo in generating shared library on x86_64.
Zhang Xianyi [Tue, 16 Jul 2013 15:18:18 +0000 (23:18 +0800)]
Fixed typo in generating shared library on x86_64.

11 years agoModified Makefile to avoid redundant echo.
Zhang Xianyi [Tue, 16 Jul 2013 14:44:27 +0000 (22:44 +0800)]
Modified Makefile to avoid redundant echo.

11 years agoModified Makefile.install
Zhang Xianyi [Tue, 16 Jul 2013 09:45:00 +0000 (17:45 +0800)]
Modified Makefile.install

11 years agoRefs #225. Fixed a bug in GEMM OpenMP threading.
Zhang Xianyi [Mon, 15 Jul 2013 01:56:19 +0000 (09:56 +0800)]
Refs #225. Fixed a bug in GEMM OpenMP threading.

11 years agoRefs #191. A walk around for dtrtri_U single thread bug.
Zhang Xianyi [Sun, 14 Jul 2013 14:16:30 +0000 (22:16 +0800)]
Refs #191. A walk around for dtrtri_U single thread bug.

This function caused the failure of ERKALE serial test.
I replaced it with LAPACK source code.

11 years agoChanged makefile for lapack.
Zhang Xianyi [Sun, 14 Jul 2013 02:41:54 +0000 (10:41 +0800)]
Changed makefile for lapack.

11 years agoUpdated travis.
Zhang Xianyi [Fri, 12 Jul 2013 13:41:12 +0000 (21:41 +0800)]
Updated travis.

11 years agoUpdate build matrix for Travis CI.
Zhang Xianyi [Thu, 11 Jul 2013 15:49:29 +0000 (23:49 +0800)]
Update build matrix for Travis CI.

11 years agoFixed the typo.
Zhang Xianyi [Thu, 11 Jul 2013 15:47:07 +0000 (23:47 +0800)]
Fixed the typo.

11 years agoFixed generating dll bug in last commit.
Zhang Xianyi [Thu, 11 Jul 2013 14:24:50 +0000 (22:24 +0800)]
Fixed generating dll bug in last commit.

11 years agoFixed #251. Merge branch 'grisuthedragon-develop' into develop
Zhang Xianyi [Thu, 11 Jul 2013 13:41:44 +0000 (21:41 +0800)]
Fixed #251. Merge branch 'grisuthedragon-develop' into develop

11 years agocreate openblas_get_parallel to retrieve information which
grisuthedragon [Thu, 11 Jul 2013 11:39:27 +0000 (13:39 +0200)]
create openblas_get_parallel to retrieve information which
parallelization model is used by OpenBLAS.

11 years agoRefs #214, #221, #246. Fixed the getrf overflow bug on Windows.
Zhang Xianyi [Wed, 10 Jul 2013 19:20:02 +0000 (03:20 +0800)]
Refs #214, #221, #246. Fixed the getrf overflow bug on Windows.

I used a smaller threshold since the stack size is 1MB on windows.