Commit Graph
-
0a696bd4cImproved the makefile for Intel compiler. by 2012-02-20 23:36:58 +0800 -
fda39c6cb(tag: v0.1alpha2.5) Updated the Changelog. by 2012-02-20 09:06:43 +0800 -
875da22a4Merge pull request #77 from nolta/master by 2012-02-19 16:44:35 -0800 -
765387af1(refs/pull/77/merge) Merge363a563ec2into0caa5616f2by 2012-02-19 11:08:43 -0800 -
363a563ec(refs/pull/77/head) fix #49 by 2012-02-19 14:07:34 -0500 -
8da6fdc2cMerge branch 'hotfix-0.1alpha2.5' into develop by 2012-02-19 23:11:06 +0800 -
0caa5616fMerge branch 'hotfix-0.1alpha2.5' by 2012-02-19 22:56:06 +0800 -
727e6d83cReleased 0.1 alpha 2.5. Updated the documents. by 2012-02-19 22:55:31 +0800 -
da3f101a7Merge branch 'develop' into hotfix-0.1alpha2.5 by 2012-02-19 22:31:09 +0800 -
fe613de8erefs #69. Auto-detect Intel Core i6/i7 (Sandy Bridge) CPU with Nehalem assembly kernels. by 2012-02-13 19:20:35 +0800 -
142e99d4eMerge branch 'master' into develop by 2012-01-20 21:32:13 +0800 -
7af0139a0Modify P Q R size of Loongson3b. by 2012-01-11 16:05:39 +0000 -
8e53b57bbAppending gemmkernel and trmmkernel C code in kernel/generic, this code can be used to execute on a new platform which dose not have optimized assemble kernel. by 2012-01-10 17:16:13 +0000 -
0d3647c39Merge pull request #76 from StefanKarpinski/patch-1 by 2012-01-01 05:57:25 -0800 -
233c7e4be(refs/pull/76/merge) Merge0d76196a09intofe7a932ab8by 2011-12-28 20:54:45 -0800 -
0d76196a0(refs/pull/76/head) Fix #68: don't require SystemStubs on OS X. by 2011-12-28 23:53:20 -0500 -
b281f3deeMerge remote branch 'origin/loongson3a' into loongson3b by 2011-12-06 13:49:39 +0000 -
a4292976eAdding detection of complex situations in symm.c, otherwise the buffer address of sb will overlap the end of sa. by 2011-12-05 14:54:25 +0000 -
c2dad58adAdding n32 multiple threads condition. by 2011-12-01 16:33:11 +0000 -
d5a6d789eFixed a typo in Makefile. by 2011-11-28 15:31:46 +0800 -
875dde437Merge branch 'lapack_3.4.0' into develop by 2011-11-28 15:28:54 +0800 -
5be22ca80Refs #72. Upgraded LAPACK to 3.4.0 version. by 2011-11-28 15:28:22 +0800 -
66904fc4eBLAS3 used standard MIPS instructions without extensions on Loongson 3B. by 2011-11-25 11:20:25 +0000 -
8163ab7e5Change the block size on Loongson 3B. by 2011-11-23 18:40:35 +0000 -
ef6f7f32aFixed mbind bug on Loongson 3B. Check the return value of my_mbind function. by 2011-11-23 17:17:41 +0000 -
285e69e2dDisable using simple thread level3 to fix a bug on Loongson 3B. by 2011-11-17 16:46:26 +0000 -
d1baf14a6Enable thread affinity on Loongson 3B. Fixed the bug of reading cycle counter. by 2011-11-11 17:49:41 +0000 -
0884f6b78Merge branch 'loongson3a' of github.com:xianyi/OpenBLAS into loongson3b by 2011-11-11 14:26:49 +0000 -
2d78fb05cAdd conjugate condition to gemv. by 2011-11-10 15:38:48 +0000 -
b95ad4cfaSupport detecting ICT Loongson-3B CPU. by 2011-11-09 19:28:22 +0000 -
3bbe3ddb3Merge branch 'develop' of github.com:xianyi/OpenBLAS into loongson3b by 2011-11-09 19:08:29 +0000 -
a32e56500Fix the compute error of gemv when incx and incy are negative numbers. by 2011-11-04 19:32:21 +0000 -
c1e618ea2Add complete gemv function on Loongson3a platform. by 2011-11-03 13:53:48 +0000 -
19f5b5c13Fixed #66 the bug in zgemv kernel with transpose matrix on 64-bit MingW (Windows). by 2011-10-18 18:44:23 +0800 -
c852ce398Ref #65. Fixed 64-bit Windows calling convention bug in cdot and zdot. by 2011-10-18 10:23:17 +0800 -
ba31b19c0Ref #62. In OpenMP implementation, check the return value of omp_get_max_threads(). It makes sure the return value as same as blas_cpu_numbers which is an internal global variable to store the number of threads in OpenBLAS. by 2011-10-16 22:56:19 +0800 -
66a3c6df4Ref #63. Fixed generating DLL bug on ming-w64. by 2011-10-09 17:25:44 +0800 -
57658a8c1ref #62. Added the user friendly message with USE_OPENMP=1. The users should use OMP_NUM_THREADS. by 2011-10-09 15:14:48 +0800 -
9fe3049deAdding conditional compilation(#if defined(LOONGSON3A)) to avoid affecting the performance of other platforms. by 2011-09-26 15:21:45 +0000 -
831858b88Modify aligned address of sa and sb to improve the performance of multi-threads. by 2011-09-23 20:59:48 +0000 -
8de2ba67dMerge branch 'hotfix-0.1alpha2.4' into develop by 2011-09-18 17:00:29 +0800 -
fe7a932ab(tag: v0.1alpha2.4) Merge branch 'hotfix-0.1alpha2.4' by 2011-09-18 16:57:28 +0800 -
1d31c79dcPrepared the document for 0.1 alpha 2.4 version. by 2011-09-18 05:46:08 +0800 -
d40e5621eChange the installation folder into /include and /lib. by 2011-09-18 05:07:00 +0800 -
bcc795621Refs #57. Continue to fix absolute path issue about shared library on Mac OSX. by 2011-09-18 01:35:12 +0800 -
821cbb299Updated the document for 0.1 alpha 2.4. by 2011-09-17 07:55:59 +0800 -
74fa79035Merge branch 'develop' into hotfix-0.1alpha2.4 by 2011-09-17 07:32:10 +0800 -
756477bfeOutput the installation tip after building complete. by 2011-09-17 07:21:11 +0800 -
864c68ffcBump the version number. by 2011-09-17 03:05:26 +0800 -
68cae521dRefs #57. The bug about absolute path of shared library on Mac OSX. by 2011-09-17 02:58:01 +0800 -
d0152ec8cFixed #61 a building bug about setting TARGET and DYNAMIC_ARCH at the same time. by 2011-09-17 02:27:56 +0800 -
e08cfaf9cComplete all the complex single-precision functions of level3, but the performance needs further improve. by 2011-09-16 17:50:40 +0000 -
ee4bb8bd2Add ctrmm part in cgemm_kernel_loongson3a_4x2_ps.S. by 2011-09-16 16:08:39 +0000 -
7fa3d23ddComplete cgemm function, but no optimization. by 2011-09-15 16:08:23 +0000 -
9679dd077Fix some compute error. by 2011-09-14 20:00:35 +0000 -
048742f38Merge branch 'loongson3a' of github.com:xianyi/OpenBLAS into loongson3a by 2011-09-14 16:32:36 +0000 -
7b410b7f0Fixed #58 zdot SEGFAULT bug with GCC-4.6. Thank Mr. John for this patch. by 2011-09-14 23:52:51 +0800 -
d238a768aUse ps instructions in cgemm. by 2011-09-14 15:32:25 +0000 -
260db9fb9Merge branch 'hotfix-0.1alpha2.3' into develop by 2011-09-09 00:57:47 +0800 -
e27b761d7(tag: v0.1alpha2.3) Merge branch 'hotfix-0.1alpha2.3' by 2011-09-09 00:55:04 +0800 -
16fc08332Refs #47. Fixed the seting parameter bug on Loongson 3A single thread version. by 2011-09-08 16:39:34 +0000 -
3c856c0c1Check the return value of pthread_create. Update the docs with known issue on Loongson 3A. by 2011-09-06 18:27:33 +0000 -
dc9c69db9Merge branch 'develop' into loongson3a by 2011-09-06 18:19:50 +0000 -
b1fe26c45refs #55. Changed DTB_ENTRIES to DTB_DEFAULT_ENTRIES in x86 gemv_n kernel codes. by 2011-09-06 14:14:07 +0800 -
0389b631fMerge branch 'loongson3a' of github.com:xianyi/OpenBLAS into loongson3a by 2011-09-05 16:31:40 +0000 -
64fa709d1Fixed #46. Initialize variables in cblat3.f and zblat3.f. by 2011-09-05 16:30:55 +0000 -
4727fe8abRefs #47. On Loongson 3A, set DGEMM_R parameter depending on different number of threads. It would improve double precision BLAS3 on multi-threads. by 2011-09-05 15:13:05 +0000 -
90481ce74Updated the doc about 0.1alpha2.3. by 2011-09-05 17:40:55 +0800 -
9fc6764farefs #55. Added DTB_ENTRIES into dynamic arch setting parameters. Now, it can read DTB_ENTRIES on runtime. by 2011-09-05 17:37:07 +0800 -
74d4cdb81Fix an illegal instruction for strmm_RTLU. by 2011-09-02 19:41:06 +0000 -
790614683Fix an error for strmm_LLTN. by 2011-09-02 16:57:33 +0000 -
3274ff47bFix an error for strmm_LLTN. by 2011-09-02 16:50:50 +0000 -
a059c553aFix a compute error for strmm. by 2011-09-02 16:00:04 +0000 -
23e182ca7Fix stack-pointer bug for strmm. by 2011-09-02 15:28:01 +0000 -
a15bc9582Add strmm part. by 2011-09-02 09:15:09 +0000 -
74a3f6348Tuning mb, kb, nb size to get the best performance. by 2011-09-01 17:15:28 +0000 -
09f49fa89Using PS instructions to improve the performance of sgemm and it is 4.2Gflops now. by 2011-08-31 21:24:03 +0000 -
b9d89f8aaFixed the bug about installation. f77blas.h works OK now. by 2011-08-31 18:21:37 +0800 -
cb0214787Modify compile options. by 2011-08-30 20:57:00 +0000 -
2e8cdd154Using ps instruction. by 2011-08-30 20:54:19 +0000 -
b29d327d1Merge branch 'loongson3a' of github.com:xianyi/OpenBLAS into loongson3a by 2011-07-18 17:06:53 +0000 -
c8360e3aeComplete all the plura single precision functions of level3 on Loongson3a, the performance is 2.3GFlops. by 2011-07-18 17:03:38 +0000 -
19d2ab485Merge branch 'hotfix-0.1alpha2.2' into develop by 2011-07-14 01:09:21 +0800 -
12d77deee(tag: v0.1alpha2.2) Merge branch 'hotfix-0.1alpha2.2' by 2011-07-14 01:03:09 +0800 -
043927c7dUpdate the documents for 0.1alpha2.2 version. by 2011-07-14 01:02:19 +0800 -
30947ea2dFixed #44 a makefile bug when DYNAMIC_ARCH=1 and INTERFACE64=1. by 2011-07-14 00:54:23 +0800 -
33313b022Merge branch 'develop' into loongson3a by 2011-07-07 14:25:51 +0800 -
a5300420eMerge branch 'hotfix-0.1alpha2.1' into develop by 2011-06-28 15:46:55 +0800 -
9b46bf1eb(tag: v0.1alpha2.1) Merge branch 'hotfix-0.1alpha2.1' by 2011-06-28 15:43:08 +0800 -
c06b7be32Refs #42. Output the error message when detecting fortran compiler failed. by 2011-06-28 15:42:09 +0800 -
68532fa9eMerge branch 'loongson3a' of github.com:xianyi/OpenBLAS into loongson3a by 2011-06-24 09:28:12 +0000 -
708d2b625Fix compute error in ztrmm. by 2011-06-24 09:27:41 +0000 -
e72113f06Add ztrmm and ztrsm part on loongson3a. The average performance is 2.2G. by 2011-06-23 21:11:00 +0000 -
fc21f7ad2Merge branch 'release-v0.1alpha2' into loongson3a by 2011-06-23 16:08:23 +0800 -
14f81da37Change prefetch length of A and B, the performance is 2.1G now. by 2011-06-23 10:46:58 +0000 -
ca8bf5abbMerge branch 'release-v0.1alpha2' into develop by 2011-06-23 16:07:34 +0800 -
4a73f5c5e(tag: v0.1alpha2) Merge branch 'release-v0.1alpha2' by 2011-06-23 15:18:40 +0800 -
6a0762949Fixed #38. Released v0.1 alpha2. by 2011-06-23 15:16:24 +0800 -
859b71645Refs #37. Updated REAME about the compatible issue with EKOPath compiler. by 2011-06-23 15:09:34 +0800 -
078bfd0b4Refs #39. Moved the shared lib (dll) to top directory in MingW64 compiler environment. by 2011-06-22 13:19:39 +0800