wernsaar [Sun, 1 Dec 2013 12:52:48 +0000 (13:52 +0100)]
changed level3.c
wernsaar [Sun, 1 Dec 2013 12:16:41 +0000 (13:16 +0100)]
Merge remote branch 'origin/develop' into armv7
wernsaar [Sun, 1 Dec 2013 09:12:07 +0000 (10:12 +0100)]
Merge branch 'develop' into armv7
wernsaar [Fri, 29 Nov 2013 16:06:33 +0000 (17:06 +0100)]
added complex gemv kernels for ARMV6 and ARMV7
wernsaar [Thu, 28 Nov 2013 16:40:21 +0000 (17:40 +0100)]
optimized gemv_n_vfp.S
wernsaar [Thu, 28 Nov 2013 13:35:07 +0000 (14:35 +0100)]
added optimized ctrmm kernel for ARMV6
wernsaar [Thu, 28 Nov 2013 12:41:06 +0000 (13:41 +0100)]
added optimized ztrmm kernel for ARMV6
wernsaar [Thu, 28 Nov 2013 11:45:38 +0000 (12:45 +0100)]
optimized strmm kernel for ARMV6
wernsaar [Thu, 28 Nov 2013 11:32:12 +0000 (12:32 +0100)]
optimized dtrmm kernel for ARMV7
wernsaar [Thu, 28 Nov 2013 10:54:38 +0000 (11:54 +0100)]
add optimized cgemm kernel for ARMV6
wernsaar [Thu, 28 Nov 2013 09:04:43 +0000 (10:04 +0100)]
optimized zgemm kernel for ARMV6
wernsaar [Thu, 28 Nov 2013 07:33:44 +0000 (08:33 +0100)]
optimized zgemm_kernel_2x2_vfp.S
wernsaar [Thu, 28 Nov 2013 07:08:08 +0000 (08:08 +0100)]
optimized sgemm kernel for ARMV6
wernsaar [Wed, 27 Nov 2013 17:38:32 +0000 (18:38 +0100)]
optimized sgemm kernel for ARMV6
wernsaar [Wed, 27 Nov 2013 16:37:38 +0000 (17:37 +0100)]
optimized dgemm kernel for ARMV6
wernsaar [Wed, 27 Nov 2013 15:15:06 +0000 (16:15 +0100)]
added zgemm_kernel for ARMV6
wernsaar [Mon, 25 Nov 2013 18:31:27 +0000 (19:31 +0100)]
added optimized gemv_t kernel for ARMV6
wernsaar [Mon, 25 Nov 2013 11:25:58 +0000 (12:25 +0100)]
add optimized axpy kernel
wernsaar [Sun, 24 Nov 2013 15:19:01 +0000 (16:19 +0100)]
changed default optimization from -O2 to -O3 for ARMV8
wernsaar [Sun, 24 Nov 2013 14:47:00 +0000 (15:47 +0100)]
added experimental support for ARMV8
wernsaar [Sun, 24 Nov 2013 12:22:49 +0000 (13:22 +0100)]
changed default optimization flag for ARM from -O2 to -O3
wernsaar [Sun, 24 Nov 2013 11:03:28 +0000 (12:03 +0100)]
add single and double precision gemv_n kernel for ARMV6
wernsaar [Sun, 24 Nov 2013 07:46:47 +0000 (08:46 +0100)]
added ncopy kernels for ARMV6
wernsaar [Sat, 23 Nov 2013 17:09:41 +0000 (18:09 +0100)]
added optimized sgemm and strmm kernel for ARMV6
wernsaar [Sat, 23 Nov 2013 15:24:52 +0000 (16:24 +0100)]
added optimized dgemm and dtrmm kernel for ARMV6
wernsaar [Sat, 23 Nov 2013 13:35:19 +0000 (14:35 +0100)]
fixed bug in SAVE macros, that are not found by any test routine
wernsaar [Sat, 23 Nov 2013 11:35:31 +0000 (12:35 +0100)]
small optimizations for zgemv kernels
wernsaar [Fri, 22 Nov 2013 16:21:10 +0000 (17:21 +0100)]
added nrm2 kernel for ARMV6
wernsaar [Thu, 21 Nov 2013 23:20:25 +0000 (00:20 +0100)]
renamed some ncopy and tcopy files
wernsaar [Thu, 21 Nov 2013 22:42:54 +0000 (23:42 +0100)]
removed obsolete file sdot_vfpv3.S
wernsaar [Thu, 21 Nov 2013 22:34:51 +0000 (23:34 +0100)]
added sdot_vfp.S
wernsaar [Thu, 21 Nov 2013 22:19:32 +0000 (23:19 +0100)]
renamed swap_vfpv3.S to swap_vfp.S
wernsaar [Thu, 21 Nov 2013 22:07:51 +0000 (23:07 +0100)]
renamed some dot kernels
wernsaar [Thu, 21 Nov 2013 21:52:24 +0000 (22:52 +0100)]
delete rot_vfpv3.S
wernsaar [Thu, 21 Nov 2013 21:49:28 +0000 (22:49 +0100)]
renamed rot_vfpv3.S to rot_vfp.S
wernsaar [Thu, 21 Nov 2013 21:26:27 +0000 (22:26 +0100)]
renamed asum_vfpv3.S to asum_vfp.S
wernsaar [Thu, 21 Nov 2013 21:03:36 +0000 (22:03 +0100)]
renamed scal_vfpv3.S to scal_vfp.S
wernsaar [Thu, 21 Nov 2013 20:12:33 +0000 (21:12 +0100)]
renamed iamax assembler kernel
wernsaar [Thu, 21 Nov 2013 19:48:57 +0000 (20:48 +0100)]
renamed some BLAS kernels, which are compatible to ARMV6
wernsaar [Thu, 21 Nov 2013 19:18:51 +0000 (20:18 +0100)]
added cpu detection and target ARMV6, used in raspberry pi
wernsaar [Tue, 19 Nov 2013 14:07:20 +0000 (15:07 +0100)]
added gemv_n kernel for single and double precision
wernsaar [Tue, 19 Nov 2013 08:55:54 +0000 (09:55 +0100)]
added gemv_t kernel for single and double precision
wernsaar [Sat, 16 Nov 2013 15:17:17 +0000 (16:17 +0100)]
added nrm2 kernel for all precisions
wernsaar [Fri, 15 Nov 2013 13:08:57 +0000 (14:08 +0100)]
added rot kernel for all precisions
wernsaar [Fri, 15 Nov 2013 10:56:43 +0000 (11:56 +0100)]
added scal kernel for all precisions
wernsaar [Thu, 14 Nov 2013 18:06:19 +0000 (19:06 +0100)]
added swap-kernel for all precisions
wernsaar [Thu, 14 Nov 2013 12:52:47 +0000 (13:52 +0100)]
added max- und min-kernels for all precisions
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.
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.
wernsaar [Mon, 11 Nov 2013 14:47:56 +0000 (15:47 +0100)]
small optimizations on dot-kernels
wernsaar [Mon, 11 Nov 2013 13:20:59 +0000 (14:20 +0100)]
added asum_kernel for all precisions and complex
wernsaar [Fri, 8 Nov 2013 08:08:11 +0000 (09:08 +0100)]
added blas level1 dot kernels for complex and double complex
wernsaar [Thu, 7 Nov 2013 16:22:03 +0000 (17:22 +0100)]
added optimized blas level1 dot kernels for single and double precision
wernsaar [Thu, 7 Nov 2013 16:18:56 +0000 (17:18 +0100)]
added optimized blas level1 copy kernels
wernsaar [Thu, 7 Nov 2013 16:15:50 +0000 (17:15 +0100)]
added cgemm_tcopy_2_vfpv3.S and zgemm_tcopy_2_vfpv3.S
Zhang Xianyi [Thu, 7 Nov 2013 05:06:42 +0000 (13:06 +0800)]
Avoid argument list too long issue in make clean.
Zhang Xianyi [Thu, 7 Nov 2013 00:12:03 +0000 (08:12 +0800)]
Refs #314. Fixed clang compiling bug on OSX.
wernsaar [Wed, 6 Nov 2013 19:01:18 +0000 (20:01 +0100)]
added dgemm_tcopy_4_vfpv3.S and sgemm_tcopy_4_vfpv3.S
Zhang Xianyi [Wed, 6 Nov 2013 17:08:39 +0000 (01:08 +0800)]
Refs #309. Fixed trtri_U single thread computational bug.
wernsaar [Tue, 5 Nov 2013 19:21:35 +0000 (20:21 +0100)]
added cgemm_ncopy_2_vfpv3.S and made assembler labels unique
wernsaar [Tue, 5 Nov 2013 18:31:22 +0000 (19:31 +0100)]
added zgemm_ncopy_2_vfpv3.S and made assembler labels unique
wernsaar [Sun, 3 Nov 2013 10:54:39 +0000 (11:54 +0100)]
added missing file kernel/arm/Makefile
wernsaar [Sun, 3 Nov 2013 10:19:32 +0000 (11:19 +0100)]
added missing file arm/Makefile in lapack/laswp
wernsaar [Sun, 3 Nov 2013 10:04:16 +0000 (11:04 +0100)]
added missing file cblas_noconst.h to the armv7 branch
wernsaar [Sun, 3 Nov 2013 09:34:04 +0000 (10:34 +0100)]
redefined functions for TIMING and YIELDING for ARMV7 processor
wernsaar [Sat, 2 Nov 2013 12:12:21 +0000 (13:12 +0100)]
deleted obsolete dgemm_kernel and dtrmm_kernel
wernsaar [Sat, 2 Nov 2013 12:06:11 +0000 (13:06 +0100)]
small optimizations on sgemm_kernel for ARMV7
wernsaar [Sat, 2 Nov 2013 08:43:53 +0000 (09:43 +0100)]
minor optimizations on zgemm_kernel for ARMV7
Zhang Xianyi [Sat, 2 Nov 2013 07:59:00 +0000 (15:59 +0800)]
Fixed #315. Added OPENBLAS_ prefix to openblas_config.h.
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.
wernsaar [Fri, 1 Nov 2013 17:22:27 +0000 (18:22 +0100)]
added sgemm_ncopy routine and made some improvements on cgemm_kernel for ARMV7
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
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.
Zhang Xianyi [Tue, 22 Oct 2013 05:24:47 +0000 (13:24 +0800)]
Added gfortran dependency for LSB/lsbcc.
wernsaar [Wed, 16 Oct 2013 17:04:42 +0000 (19:04 +0200)]
moved compiler flags from Makefile.rule to Makefile.arm
wernsaar [Wed, 16 Oct 2013 16:04:34 +0000 (18:04 +0200)]
optimized param.h
wernsaar [Wed, 16 Oct 2013 16:00:41 +0000 (18:00 +0200)]
added kernels for cgemm, ctrmm, zgemm and ztrmm
wernsaar [Mon, 14 Oct 2013 06:22:27 +0000 (08:22 +0200)]
added sgemm- and strmm_kernel
wernsaar [Sat, 12 Oct 2013 14:48:29 +0000 (16:48 +0200)]
added dgemm_ncopy_4_vfpv3.S
wernsaar [Sat, 12 Oct 2013 07:42:18 +0000 (09:42 +0200)]
minor optimizations on dgemm_kernel
wangqian [Thu, 10 Oct 2013 19:20:20 +0000 (03:20 +0800)]
Fixed buffer overflow bug in kernel/x86_64/dgemv_t.S file.
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
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.
wernsaar [Sat, 5 Oct 2013 10:59:44 +0000 (12:59 +0200)]
Changed kernels for dgemm and dtrmm
wernsaar [Mon, 30 Sep 2013 16:03:56 +0000 (18:03 +0200)]
changed some values for arm
wernsaar [Mon, 30 Sep 2013 15:31:23 +0000 (17:31 +0200)]
updated dgemm_kernel_8x2_vfpv3.S
wernsaar [Sun, 29 Sep 2013 17:42:33 +0000 (19:42 +0200)]
add modified c_check perl program
wernsaar [Sun, 29 Sep 2013 16:55:21 +0000 (18:55 +0200)]
added Makefile.arm
wernsaar [Sun, 29 Sep 2013 15:46:23 +0000 (17:46 +0200)]
changed dgemm_kernel to use fused multiply add
wernsaar [Sat, 28 Sep 2013 17:13:47 +0000 (19:13 +0200)]
modified Makefile.L3 for ARM
wernsaar [Sat, 28 Sep 2013 17:10:32 +0000 (19:10 +0200)]
common files modified for ARM
wernsaar [Sat, 28 Sep 2013 17:02:25 +0000 (19:02 +0200)]
initial checkin of kernel/arm
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
Zhang Xianyi [Thu, 5 Sep 2013 07:39:45 +0000 (15:39 +0800)]
Added backers.
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.
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
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
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.
Lars Buitinck [Wed, 28 Aug 2013 14:52:24 +0000 (16:52 +0200)]
get rid of the generated cblas_noconst.h file
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!