Commit Graph
-
9dca578c7(refs/pull/2891/head) Cleanup by 2020-10-13 10:14:08 +0200 -
1e7eb7b7aFix typos in currently unused sections by 2020-10-13 09:17:15 +0200 -
84949754aFix bfloat16 conditional by 2020-10-13 09:11:36 +0200 -
2ae878560Add a POWER9 build with BFLOAT16 enabled by 2020-10-13 09:07:50 +0200 -
e05af6575Fix some overlooked "SHBLAS" entries by 2020-10-13 09:05:04 +0200 -
c1643006aMerge pull request #97 from xianyi/develop by 2020-10-13 09:01:49 +0200 -
8d2df7d06Revert special handling of Windows xNRM2 and enable C+intrinsics kernel for SSUM/DSUM by 2020-10-13 00:14:29 +0200 -
08929430cMerge pull request #2886 from martin-frbg/issue_2767 by 2020-10-13 00:04:35 +0200 -
0c84ffe05Merge pull request #2881 from mattip/fninit by 2020-10-12 23:50:41 +0200 -
36bd6ba6c(refs/pull/2887/head) Use the new universal intrinsics for s/dSUM across all platforms, and generic C c/zSUM on Windows by 2020-10-12 23:45:49 +0200 -
cb4274e3aMerge pull request #2888 from Qiyu8/usimd-sum by 2020-10-12 23:22:08 +0200 -
fac9afe64(refs/pull/2889/head) Reset the FPU stack on Windows to work around a bug in Windows10.19041 by 2020-10-12 19:04:01 +0200 -
403eb513a(refs/pull/2881/head) use emms instead, add WIN guards by 2020-10-12 18:15:01 +0300 -
cb839575e(refs/pull/2886/head) Convert the prototypes of the unimplemented BFLOAT16 functions to the new naming scheme by 2020-10-12 14:44:33 +0200 -
0ed1f0766(refs/pull/2888/head) Optimize the performance of sum by using universal intrinsics by 2020-10-12 19:48:53 +0800 -
600054b0aUse generic kernels for xSUM on Windows by 2020-10-12 08:24:51 +0200 -
bb74dd29dRestore -msse3 by 2020-10-12 00:42:05 +0200 -
629c497b6common_sh.h renamed to common_sb.h by 2020-10-12 00:27:11 +0200 -
2c552f107Change "HALF" and "sh" to "BFLOAT16" and "sb" by 2020-10-12 00:11:31 +0200 -
7ae9e8960Change "HALF" and "sh" to "BFLOAT16" and "sb" by 2020-10-12 00:08:29 +0200 -
e3a29f6b5Change "HALF" and "sh" to "BFLOAT16" and "sb" by 2020-10-12 00:07:37 +0200 -
006c7f667Change "HALF" and "sh" to "BFLOAT16" and "sb" by 2020-10-12 00:06:06 +0200 -
85154c2e1Change "HALF" and "sh" to "BFLOAT16" and "sb" by 2020-10-12 00:05:05 +0200 -
ae1ab5bfdChange "HALF" and "sh" to "BFLOAT16" and "sb" by 2020-10-12 00:03:21 +0200 -
052f31bc3Change "HALF" and "sh" to "BFLOAT16" and "sb" by 2020-10-12 00:02:16 +0200 -
3aecafad8Change "HALF" and "sh" to "BFLOAT16" and "sb" by 2020-10-12 00:00:55 +0200 -
756062afaRename "HALF" and "sh" to "BFLOAT16" and "sb" by 2020-10-11 23:56:17 +0200 -
2061f7fdfRename "HALF" and "sh" to "BFLOAT16" and "sb" by 2020-10-11 23:54:53 +0200 -
dc8a1afa6Rename "HALF" and "sh" to "BFLOAT16" and "sb" by 2020-10-11 23:53:50 +0200 -
32733ded0Rename "HALF" and "sh" to "BFLOAT16" and "sb" by 2020-10-11 23:52:45 +0200 -
3bc8e8c33Rename "HALF" and "sh" to "BFLOAT16"and "sb" by 2020-10-11 23:51:34 +0200 -
573508f0eRename common_sh.h to common_sb.h by 2020-10-11 23:50:54 +0200 -
ca31c3269Rename "HALF" and "sh" to "BFLOAT16" and "sb" by 2020-10-11 23:49:22 +0200 -
5800758b4Rename "HALF" and "sh" to "BFLOAT16" and "sb" by 2020-10-11 23:44:38 +0200 -
924fd806dRename "HALF" and "sh" to "BFLOAT16" and "sb" by 2020-10-11 23:43:36 +0200 -
4db09c6ceRename compare_sgemm_shgemm.c to compare_sgemm_sbgemm.c by 2020-10-11 23:42:45 +0200 -
fd9423604Rename "HALF" and "sh" to "BFLOAT16" and "sb" by 2020-10-11 23:42:07 +0200 -
68ce719faRename shdot_microk_cooperlake.c to sbdot_microk_cooperlake.c by 2020-10-11 23:41:13 +0200 -
d7dd9b396Rename shdot.c to sbdot.c by 2020-10-11 23:40:43 +0200 -
9ae80490erename "HALF" and "sh" to "BFLOAT16" and "sb" by 2020-10-11 23:39:42 +0200 -
d314d1f49Rename shgemm_kernel_power10.c to sbgemm_kernel_power10.c by 2020-10-11 23:37:38 +0200 -
f0883740eMerge pull request #96 from xianyi/develop by 2020-10-11 23:34:36 +0200 -
1c0b03efbMerge branch 'develop' into develop by 2020-10-11 23:34:14 +0200 -
c589c3e2aMerge pull request #2882 from martin-frbg/issue2709 by 2020-10-11 22:22:30 +0200 -
ec638a82bMerge pull request #2852 from martin-frbg/issue2588-cmake by 2020-10-11 22:21:33 +0200 -
caa0d757c(refs/pull/2852/head) repair TABs by 2020-10-11 18:29:34 +0200 -
6154f72d6Copy BUILD_ settings to the LAPACK make.inc by 2020-10-11 18:25:16 +0200 -
ae8b0d257Set BUILD_ options to 1 instead of just defining them by 2020-10-11 18:08:21 +0200 -
1da32cc1fAdd cblas_xerbla interface by 2020-10-11 17:45:41 +0200 -
8c5e08076If none of the BUILD_ options is set, enable them all by 2020-10-11 17:33:51 +0200 -
5f23bdf43remove debug output by 2020-10-11 17:23:08 +0200 -
b593e6b65Merge pull request #2885 from martin-frbg/ifexists by 2020-10-11 15:45:24 +0200 -
082c86a53Merge pull request #2884 from martin-frbg/sse_fixup by 2020-10-11 15:14:03 +0200 -
e396ec8b5Allow building support for only a subset of variable types by 2020-10-11 15:11:15 +0200 -
68e6823d3Adapt for supporting only a subset of variable types by 2020-10-11 15:01:32 +0200 -
887e00fd7Adapt for supporting only a subset of variable types by 2020-10-11 14:58:57 +0200 -
886a8e319Adapt for supporting only a subset of variable types by 2020-10-11 14:57:32 +0200 -
0f7d73ff6Allow supporting only a subset of variable types by 2020-10-11 14:53:26 +0200 -
6b6adf8a4Allow compiling only a subset of kernels for specific variable types by 2020-10-11 14:52:09 +0200 -
a6570108cAdd Makefile support for enabling only some variable types by 2020-10-11 14:49:58 +0200 -
ef552bc57Add Makefile support for enabling only some variable types by 2020-10-11 14:49:06 +0200 -
efe1ad470Add Makefile support for enabling only some variable types by 2020-10-11 14:48:23 +0200 -
b27ca78a2Adapt to having only a subset of variable types supported by 2020-10-11 14:46:24 +0200 -
93454022aAdapt to having only a subset of variable types supported by 2020-10-11 14:45:40 +0200 -
20cf1d773Adapt to having only a subset of variable types supported by 2020-10-11 14:44:56 +0200 -
5c657fffaAdapt to having only a subset of variable types supported by 2020-10-11 14:44:13 +0200 -
b26205805Adapt to having only a subset of variable types supported by 2020-10-11 14:43:13 +0200 -
bc319cee8Adapt to having only a subset of variable types supported by 2020-10-11 14:42:26 +0200 -
e5966f860Adapt to having only a subset of variable types supported by 2020-10-11 14:41:43 +0200 -
9df12eb08Adapt to having only a subset of variable types supported by 2020-10-11 14:40:51 +0200 -
cf53970bcAdapt to having only a subset of variable types supported by 2020-10-11 14:40:06 +0200 -
dcd51d5c7Adapt to having only a subset of variable types supported by 2020-10-11 14:39:19 +0200 -
b8f95354cAdapt to having only a subset of variable types supported by 2020-10-11 14:38:25 +0200 -
d33de97d6Adapt to having only a subset of variable types supported by 2020-10-11 14:36:45 +0200 -
6a83c591dAdapt for having only a subset of variable types by 2020-10-11 14:34:12 +0200 -
f6d2827d0Adapt ctests to having only a subset of types in the build by 2020-10-11 14:32:00 +0200 -
08f4749ebAdapt tests to having only a subset of types in the build by 2020-10-11 14:25:24 +0200 -
63d7dad04Adapt utests for builds supportin only some variable types by 2020-10-11 14:15:35 +0200 -
ac653c94fMerge branch 'develop' into issue2588-cmake by 2020-10-11 13:57:07 +0200 -
190b74dd2Add files via upload by 2020-10-11 13:26:05 +0200 -
9d43140d6(refs/pull/2885/head) Improve check for conflicting config_kernel.h by 2020-10-11 12:58:17 +0200 -
8ef600f1aMerge pull request #95 from xianyi/develop by 2020-10-11 12:53:18 +0200 -
88928650cMerge pull request #2883 from martin-frbg/issue2872 by 2020-10-11 10:30:33 +0200 -
7a5312848(refs/pull/2884/head) Add whitelist of DYNAMIC_ARCH kernels for which -msse3 needs to be enabled by 2020-10-11 01:06:46 +0200 -
0c773b820Do not rely on HAVE_SSE3 in DYNAMIC_ARCH builds by 2020-10-11 01:04:57 +0200 -
fbda20c85Merge pull request #94 from xianyi/develop by 2020-10-11 01:03:00 +0200 -
82a497ec5(refs/pull/2883/head) restore PRESCOTT default for DYNAMIC_LIST by 2020-10-11 00:43:09 +0200 -
de27e4f5fStop DYNAMIC_ARCH build if the toplevel source contains a stray config_kernel.h from a gmake build by 2020-10-11 00:40:22 +0200 -
e1b7123bbMerge pull request #2867 from Qiyu8/usimd-floatdot by 2020-10-10 12:10:25 +0200 -
f32d34a01(refs/pull/2867/head) add sse3 compiler flag by 2020-10-10 10:36:15 +0800 -
599777ecbMerge pull request #2879 from martin-frbg/issue2839 by 2020-10-06 23:26:52 +0200 -
781248609(refs/pull/2882/head) Use generic C for D/Z nrm2 kernels on Windows to work around fpu exception bug by 2020-10-06 21:33:16 +0200 -
a5b164946add fninit to reset fpu registers before assembler routines by 2020-10-05 22:13:25 +0300 -
a5feea661(refs/pull/2879/head) make BLAS3_MEM_ALLOC_THRESHOLD configurable on non-Windows by 2020-10-04 23:01:06 +0200 -
dc8e4e195Reduce the BLAS3 heap allocation threshold to 32 and mark it as configurable by 2020-10-04 22:59:24 +0200 -
cccd1438dMerge pull request #93 from xianyi/develop by 2020-10-04 22:57:11 +0200 -
f032d8966Merge pull request #2874 from Flamefire/memory_fixes by 2020-10-04 15:16:51 +0200 -
f6e4cf2f9Merge pull request #2876 from Flamefire/omp_fork_fix by 2020-10-03 22:52:17 +0200 -
9828343e1Merge pull request #2878 from brada4/asms by 2020-10-03 22:51:49 +0200 -
d2333e784(refs/pull/2878/head) aarch64 fix std=c18 compilation by 2020-10-03 18:00:34 +0300