wernsaar [Sun, 18 May 2014 17:36:32 +0000 (19:36 +0200)]
enabled and tested optimized zgetrf lapack function
wernsaar [Sun, 18 May 2014 17:07:51 +0000 (19:07 +0200)]
enabled and tested optimized dgetrf function
wernsaar [Sun, 18 May 2014 12:09:22 +0000 (14:09 +0200)]
added optimized lapack files from OpenBLAS
wernsaar [Sat, 17 May 2014 11:00:36 +0000 (13:00 +0200)]
bugfix for ARMV6
wernsaar [Sat, 17 May 2014 09:18:26 +0000 (11:18 +0200)]
enable debug for lapack testing
wernsaar [Fri, 16 May 2014 18:37:41 +0000 (20:37 +0200)]
some modifications regarding lapack test
wernsaar [Fri, 16 May 2014 18:34:48 +0000 (20:34 +0200)]
changed threshold to 50.0
wernsaar [Fri, 16 May 2014 12:36:24 +0000 (14:36 +0200)]
changed default optimization flag from O3 to O2 for ARM
wernsaar [Fri, 16 May 2014 12:32:10 +0000 (14:32 +0200)]
changed threshold for 50.0 to 54.0 in svd.in
wernsaar [Thu, 15 May 2014 09:37:38 +0000 (11:37 +0200)]
changed YIELDING for BULLDOZER
wernsaar [Wed, 14 May 2014 13:16:21 +0000 (15:16 +0200)]
Modified lapack-test, using lapack_testing.py to run tests
wernsaar [Wed, 14 May 2014 13:01:03 +0000 (15:01 +0200)]
added FCOMMON_OPT for lapack
wernsaar [Wed, 14 May 2014 11:08:05 +0000 (13:08 +0200)]
changed label lapack-test
wernsaar [Wed, 14 May 2014 10:59:20 +0000 (12:59 +0200)]
bugfixs for ARM regarding lapack tests
wernsaar [Tue, 13 May 2014 15:25:19 +0000 (17:25 +0200)]
added ARMV5 as reference platform
wernsaar [Tue, 13 May 2014 14:25:21 +0000 (16:25 +0200)]
added ZERO check to zscal.c because bug in lapack-test
wernsaar [Tue, 13 May 2014 13:36:03 +0000 (15:36 +0200)]
added ZERO check because bug in lapack-test
wernsaar [Sun, 11 May 2014 10:09:34 +0000 (12:09 +0200)]
removed lapack objects from interface/Makefile
wernsaar [Sun, 11 May 2014 09:21:23 +0000 (11:21 +0200)]
checked, that zhpr is OK
wernsaar [Sun, 11 May 2014 09:11:05 +0000 (11:11 +0200)]
checked, that zhpr2 is OK
wernsaar [Sun, 11 May 2014 08:46:48 +0000 (10:46 +0200)]
checked, that zhpmv is OK
wernsaar [Sun, 11 May 2014 08:36:34 +0000 (10:36 +0200)]
checked, that zher is OK
wernsaar [Sun, 11 May 2014 08:26:05 +0000 (10:26 +0200)]
checked, that zher2 is OK
wernsaar [Sun, 11 May 2014 08:15:06 +0000 (10:15 +0200)]
checked, that zhemv is OK
wernsaar [Sun, 11 May 2014 07:58:16 +0000 (09:58 +0200)]
marked zhbmv as smp bug
wernsaar [Sun, 11 May 2014 07:30:18 +0000 (09:30 +0200)]
checked, that zscal and zswap are OK
wernsaar [Sun, 11 May 2014 07:11:20 +0000 (09:11 +0200)]
checked, that trtri is OK
wernsaar [Sun, 11 May 2014 06:59:33 +0000 (08:59 +0200)]
checked, that trsm is OK
wernsaar [Sun, 11 May 2014 06:47:44 +0000 (08:47 +0200)]
checked, that trmv is OK
wernsaar [Sun, 11 May 2014 06:35:34 +0000 (08:35 +0200)]
checked, that tpmv is OK
wernsaar [Sun, 11 May 2014 06:22:00 +0000 (08:22 +0200)]
checked, that tbmv is OK
wernsaar [Sun, 11 May 2014 06:10:25 +0000 (08:10 +0200)]
checked, that syrk is OK
wernsaar [Sun, 11 May 2014 05:46:22 +0000 (07:46 +0200)]
checked, that syr is OK
wernsaar [Sat, 10 May 2014 23:04:46 +0000 (01:04 +0200)]
check, that syr2k is OK
wernsaar [Sat, 10 May 2014 22:48:49 +0000 (00:48 +0200)]
checked, that syr2 is OK
wernsaar [Sat, 10 May 2014 22:36:56 +0000 (00:36 +0200)]
checked, that symv is OK
wernsaar [Sat, 10 May 2014 22:22:40 +0000 (00:22 +0200)]
checked, that symm is OK
wernsaar [Sat, 10 May 2014 22:07:07 +0000 (00:07 +0200)]
checked, that spr is OK
wernsaar [Sat, 10 May 2014 21:55:43 +0000 (23:55 +0200)]
checked, that spr2 is OK
wernsaar [Sat, 10 May 2014 21:18:35 +0000 (23:18 +0200)]
marked as smp bug
wernsaar [Sat, 10 May 2014 20:55:44 +0000 (22:55 +0200)]
removed gesv.c from interface
wernsaar [Sat, 10 May 2014 20:49:53 +0000 (22:49 +0200)]
checked, that ger is OK
wernsaar [Sat, 10 May 2014 20:38:09 +0000 (22:38 +0200)]
cheched, that gbmv is OK
wernsaar [Sat, 10 May 2014 20:24:05 +0000 (22:24 +0200)]
checked, that gemv is OK
wernsaar [Sat, 10 May 2014 20:09:49 +0000 (22:09 +0200)]
checked that axpy is OK
wernsaar [Sat, 10 May 2014 19:51:44 +0000 (21:51 +0200)]
checked that gemm.c is OK
wernsaar [Sat, 10 May 2014 19:27:47 +0000 (21:27 +0200)]
added C files in interface
wernsaar [Sat, 10 May 2014 19:25:17 +0000 (21:25 +0200)]
removed all C files in interface
wernsaar [Sat, 10 May 2014 09:24:07 +0000 (11:24 +0200)]
replaced sgemm and cgemm kernels because lapack bugs
wernsaar [Sat, 10 May 2014 08:52:37 +0000 (10:52 +0200)]
replaced dgemm-kernel because bug in lapack
wernsaar [Fri, 9 May 2014 11:17:47 +0000 (13:17 +0200)]
changed test ratio from 30.0 to 40.0
wernsaar [Thu, 8 May 2014 09:06:32 +0000 (11:06 +0200)]
fixed incorrect parameter 2 errors
wernsaar [Wed, 7 May 2014 12:36:49 +0000 (14:36 +0200)]
added log file for lapack development
wernsaar [Wed, 7 May 2014 09:33:02 +0000 (11:33 +0200)]
changed makefiles for lapack development
wernsaar [Tue, 6 May 2014 13:29:39 +0000 (15:29 +0200)]
replaced files for sdot, sgemv_n and sgemv_t for bug #348
Zhang Xianyi [Fri, 2 May 2014 09:55:08 +0000 (17:55 +0800)]
Merge pull request #367 from xantares/patch-2
Makefile typo
Zhang Xianyi [Fri, 2 May 2014 09:54:22 +0000 (17:54 +0800)]
Merge pull request #366 from xantares/patch-1
Install dll to prefix/bin instead of prefix/lib
wernsaar [Fri, 2 May 2014 09:34:26 +0000 (11:34 +0200)]
fixed 2 bugs as reported by Brendan Tracey
xantares [Fri, 2 May 2014 06:54:22 +0000 (08:54 +0200)]
Update Makefile
xantares [Thu, 1 May 2014 19:48:26 +0000 (21:48 +0200)]
Install dll to prefix/bin instead of prefix/bin
Zhang Xianyi [Sun, 27 Apr 2014 05:05:26 +0000 (13:05 +0800)]
Refs #63. delete prefix for mingw64 toolchain.
wernsaar [Thu, 24 Apr 2014 12:09:26 +0000 (14:09 +0200)]
reduced stack usage on windows to 16K
Zhang Xianyi [Sat, 22 Mar 2014 07:08:18 +0000 (15:08 +0800)]
Refs #355. Fixed ARM detection bug.
Zhang Xianyi [Thu, 6 Mar 2014 09:44:03 +0000 (17:44 +0800)]
#351. Release 0.2.9 rc2.
Zhang Xianyi [Thu, 6 Mar 2014 04:53:10 +0000 (12:53 +0800)]
Merge branch 'develop' of github.com:xianyi/OpenBLAS into develop
wernsaar [Fri, 28 Feb 2014 13:59:36 +0000 (14:59 +0100)]
bugfix for sdsdot
wernsaar [Fri, 28 Feb 2014 13:43:28 +0000 (14:43 +0100)]
rewrote rotmg.c instead of modifying very old code
Zhang Xianyi [Tue, 25 Feb 2014 16:54:01 +0000 (00:54 +0800)]
Merge pull request #345 from ogrisel/fix-non-smp-server-pthread_atfork-reference
Do not reference pthread_atfork in non-SMP_SERVER mode
Zhang Xianyi [Tue, 25 Feb 2014 16:38:29 +0000 (00:38 +0800)]
Merge branch 'develop' of github.com:xianyi/OpenBLAS into develop
Olivier Grisel [Tue, 25 Feb 2014 09:42:39 +0000 (10:42 +0100)]
Add cast to function pointer to remove warning
Olivier Grisel [Thu, 20 Feb 2014 08:59:18 +0000 (09:59 +0100)]
Do not reference pthread_atfork in non-SMP_SERVER mode
Zhang Xianyi [Mon, 24 Feb 2014 22:43:30 +0000 (06:43 +0800)]
Merge pull request #346 from ogrisel/fix-openblas_config.h
More robust OPENBLAS_ prefixing of macros in openblas_config.h
Olivier Grisel [Mon, 24 Feb 2014 12:21:06 +0000 (13:21 +0100)]
More robust OPENBLAS_ prefixing of macros in openblas_config.h
Zhang Xianyi [Thu, 20 Feb 2014 23:15:35 +0000 (07:15 +0800)]
Merge branch 'develop' of github.com:xianyi/OpenBLAS into develop
Zhang Xianyi [Wed, 19 Feb 2014 22:58:27 +0000 (06:58 +0800)]
Merge pull request #343 from ogrisel/fix-294-fork-safe-pthread
FIX #294: fork-safe pthread mode
Olivier Grisel [Wed, 19 Feb 2014 18:25:48 +0000 (19:25 +0100)]
Do not reference pthread_atfork under windows
Olivier Grisel [Wed, 19 Feb 2014 18:14:13 +0000 (19:14 +0100)]
Make sure that fork_test.c is not built under windows
Olivier Grisel [Tue, 18 Feb 2014 09:57:01 +0000 (10:57 +0100)]
FIX #294: make OpenBLAS thread-pool resilient to fork via pthread_atfork
Olivier Grisel [Wed, 19 Feb 2014 17:32:54 +0000 (18:32 +0100)]
Revert "Refs #294. Used pthread_atfork to avoid hang after a Unix fork."
This reverts commit
3617c22a5679c52b1e760d739d76a7bde9e84d35.
Zhang Xianyi [Tue, 18 Feb 2014 07:53:57 +0000 (15:53 +0800)]
Merge branch 'develop' of github.com:xianyi/OpenBLAS into develop
Zhang Xianyi [Tue, 18 Feb 2014 07:36:04 +0000 (15:36 +0800)]
Refs #294. Used pthread_atfork to avoid hang after a Unix fork.
The problem is the mutex we used in blas_server. Thus, we must clear
the mutex before the fork and re-init them at parent and child process.
If you used OpenMP, GOMP has the same problem by now. Please try other OpenMP
implemantation.
wernsaar [Sun, 16 Feb 2014 10:45:47 +0000 (11:45 +0100)]
checked in bugfixes for ARM
Zhang Xianyi [Sat, 15 Feb 2014 15:23:15 +0000 (23:23 +0800)]
Refs #340. Fixed SEGFAULT bug of dgemv_n on OSX.
wangqian [Tue, 4 Feb 2014 13:47:51 +0000 (21:47 +0800)]
Fixed computational error of dgemv_n.
Zhang Xianyi [Tue, 28 Jan 2014 08:40:19 +0000 (16:40 +0800)]
Used SwitchToThread for YIELDING on AMD piledriver with Windows.
Zhang Xianyi [Fri, 24 Jan 2014 07:39:46 +0000 (15:39 +0800)]
Refs #329 #287. Only disable -fopenmp for LAPACK Fortran codes on Windows.
Zhang Xianyi [Thu, 23 Jan 2014 18:05:59 +0000 (02:05 +0800)]
Refs #338. Added OPENBLAS_VERBOSE environment variable on runtime
By default, OpenBLAS doesn't output the warning message. You can set
OPENBLAS_VERBOSE (e.g. export OPENBLAS_VERBOSE=1) to enable the warning
message on runtime.
Zhang Xianyi [Mon, 20 Jan 2014 17:27:51 +0000 (01:27 +0800)]
Refs #333. Detect the wrong parameter for zherk/zher2k.
Zhang Xianyi [Sun, 19 Jan 2014 15:28:11 +0000 (23:28 +0800)]
Fixed #334 a makefile bug in lapacke.
Zhang Xianyi [Wed, 8 Jan 2014 03:16:21 +0000 (11:16 +0800)]
Refs #335. Added the fallback of L2 size detection for some virtual machines.
Zhang Xianyi [Sun, 5 Jan 2014 15:44:29 +0000 (23:44 +0800)]
Refs #332. Added addtional Intel Ivy Bridge and Haswell CPU-id.
Zhang Xianyi [Tue, 31 Dec 2013 07:53:13 +0000 (15:53 +0800)]
Refs #307. Delete debug printf.
Zhang Xianyi [Mon, 16 Dec 2013 12:31:17 +0000 (20:31 +0800)]
Refs #330. Fixed the compatible issue with clang on Mac OSX.
Zhang Xianyi [Fri, 13 Dec 2013 12:48:05 +0000 (20:48 +0800)]
Release 0.2.9 rc1 version.
Zhang Xianyi [Fri, 13 Dec 2013 08:17:41 +0000 (16:17 +0800)]
Refs #325 #324. Imported carkl's patch for building LAPACKE on Windows.
Zhang Xianyi [Thu, 12 Dec 2013 15:07:33 +0000 (23:07 +0800)]
Refs #310. Fixed Segfault bug on nehalem when Julia calling dgeqrt3 on OSX.
Please also check JuliaLang/julia#4099
Julia test script:
A=rand(256, 256)
qrfact(A)
I found this was a bug in kernel/x86_64/dgemm_ncopy_8.S.
However, I cannot use gdb with julia. Thus, this is a walkaround fix.
Zhang Xianyi [Wed, 11 Dec 2013 14:16:40 +0000 (06:16 -0800)]
Merge pull request #327 from lucasb-eyer/develop
Added support for NO_SHARED in make install.
lucasb-eyer [Tue, 10 Dec 2013 16:10:13 +0000 (17:10 +0100)]
Added support for NO_SHARED in make install.
Zhang Xianyi [Mon, 9 Dec 2013 15:05:02 +0000 (07:05 -0800)]
Merge pull request #326 from cicku/develop
Update Makefile.install
Christopher Meng [Mon, 9 Dec 2013 11:56:37 +0000 (19:56 +0800)]
Add DESTDIR support for easier building on RPM based distros.
Use install command instead of cp to install files with permissions control.