floraachy/OpenBLAS
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
200 Commits 47 Branches 64 Tags
3bbe3ddb31934fa4e63e8ee883ef6e7cf4649c61
Commit Graph

200 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Xianyi Zhang
3bbe3ddb31 Merge branch 'develop' of github.com:xianyi/OpenBLAS into loongson3b 2011-11-09 19:08:29 +00:00
traz
a32e56500a Fix the compute error of gemv when incx and incy are negative numbers. 2011-11-04 19:32:21 +00:00
traz
c1e618ea2d Add complete gemv function on Loongson3a platform. 2011-11-03 13:53:48 +00:00
traits
19f5b5c132 Fixed #66 the bug in zgemv kernel with transpose matrix on 64-bit MingW (Windows). 2011-10-18 18:44:23 +08:00
traits
c852ce3981 Ref #65. Fixed 64-bit Windows calling convention bug in cdot and zdot. 
According to 64-bit Windows calling convention, the return value is in %rax instead of %xmm0 in cdot kernel.
In zdot, the caller allocates a memory space for return value and sets this memory address to the first hidden parameter. Thus, the callee (zdot) should assign the result to this memory space and return the memory address in %rax.
 2011-10-18 10:23:17 +08:00
Xianyi Zhang
ba31b19c00 Ref #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.
 2011-10-16 23:00:43 +08:00
traits
66a3c6df4e Ref #63. Fixed generating DLL bug on ming-w64. 2011-10-09 17:25:44 +08:00
Xianyi Zhang
57658a8c14 ref #62. Added the user friendly message with USE_OPENMP=1. The users should use OMP_NUM_THREADS. 
When OpenBLAS is compiled with USE_OPENMP=1, it ignores OPENBLAS_NUM_THREADS and GOTO_NUM_THREADS flags.Therefore, you should use OMP_NUM_THREADS.

Without setting OMP_NUM_THREADS, a process will use maximal number of threads on a computing node. Thus, if there are 2 processes on the computing node, the thread will contend against other threads on CPU cores. As a result, the application will hang.
 2011-10-09 15:14:48 +08:00
traz
9fe3049de6 Adding conditional compilation(#if defined(LOONGSON3A)) to avoid affecting the performance of other platforms. 2011-09-26 15:21:45 +00:00
traz
831858b883 Modify aligned address of sa and sb to improve the performance of multi-threads. 2011-09-23 20:59:48 +00:00
Xianyi
8de2ba67dd Merge branch 'hotfix-0.1alpha2.4' into develop 2011-09-18 17:00:29 +08:00
Xianyi
1d31c79dc9 Prepared the document for 0.1 alpha 2.4 version. 2011-09-18 16:51:09 +08:00
Xianyi
d40e5621e9 Change the installation folder into /include and /lib. 2011-09-18 05:07:00 +08:00
Xianyi
bcc7956216 Refs #57. Continue to fix absolute path issue about shared library on Mac OSX. 
Used $(CURDIR) instead of pwd in generating shared library on Mac OSX.
Add more tips about this issue. Thank Dr Kane O'Donnell.
 2011-09-18 01:35:12 +08:00
Xianyi
821cbb2995 Updated the document for 0.1 alpha 2.4. 2011-09-17 07:55:59 +08:00
Xianyi
74fa790354 Merge branch 'develop' into hotfix-0.1alpha2.4 2011-09-17 07:32:10 +08:00
Xianyi
756477bfe3 Output the installation tip after building complete. 2011-09-17 07:21:11 +08:00
Xianyi
864c68ffc5 Bump the version number. 2011-09-17 03:05:26 +08:00
Xianyi
68cae521df Refs #57. The bug about absolute path of shared library on Mac OSX. 
OSX cann't use relative path in shared library. Thank Mr.Kane for this patch. The detail is in this link (https://github.com/xianyi/OpenBLAS/issues/57).
 2011-09-17 02:58:01 +08:00
Xianyi
d0152ec8ca Fixed #61 a building bug about setting TARGET and DYNAMIC_ARCH at the same time. 2011-09-17 02:27:56 +08:00
traz
e08cfaf9ca Complete all the complex single-precision functions of level3, but the performance needs further improve. 2011-09-16 17:50:40 +00:00
traz
ee4bb8bd25 Add ctrmm part in cgemm_kernel_loongson3a_4x2_ps.S. 2011-09-16 16:08:39 +00:00
traz
7fa3d23dd9 Complete cgemm function, but no optimization. 2011-09-15 16:08:23 +00:00
traz
9679dd077e Fix some compute error. 2011-09-14 20:00:35 +00:00
traz
048742f38f Merge branch 'loongson3a' of github.com:xianyi/OpenBLAS into loongson3a 2011-09-14 16:32:36 +00:00
Zhang Xiianyi
7b410b7f0e Fixed #58 zdot SEGFAULT bug with GCC-4.6. Thank Mr. John for this patch. 
In i386 calling convention, the caller put the address of return value of zdot into the first hidden parameter.
Thus, the callee should delete this address before return.
Actually, I have fixed the same bug on x86/zdot_sse2.S (issue #32). However, that is not a good implementation which uses 3 instructions. Mr. John told me used "ret $0x4" to skip the first hidden address (4 bytes).
 2011-09-14 23:52:51 +08:00
traz
d238a768ab Use ps instructions in cgemm. 2011-09-14 15:32:25 +00:00
traits
260db9fb9e Merge branch 'hotfix-0.1alpha2.3' into develop 2011-09-09 00:57:47 +08:00
traits
e27b761d7c Merge branch 'hotfix-0.1alpha2.3' v0.1alpha2.3 2011-09-09 00:55:04 +08:00
Xianyi Zhang
16fc083322 Refs #47. Fixed the seting parameter bug on Loongson 3A single thread version. 2011-09-08 16:39:34 +00:00
Xianyi Zhang
3c856c0c1a Check the return value of pthread_create. Update the docs with known issue on Loongson 3A. 2011-09-06 18:27:33 +00:00
Xianyi Zhang
dc9c69db93 Merge branch 'develop' into loongson3a 2011-09-06 18:19:50 +00:00
traits
b1fe26c45a refs #55. Changed DTB_ENTRIES to DTB_DEFAULT_ENTRIES in x86 gemv_n kernel codes. 2011-09-06 14:14:07 +08:00
traz
0389b631fa Merge branch 'loongson3a' of github.com:xianyi/OpenBLAS into loongson3a 2011-09-05 16:31:40 +00:00
traz
64fa709d1f Fixed #46. Initialize variables in cblat3.f and zblat3.f. 2011-09-05 16:30:55 +00:00
Xianyi Zhang
4727fe8abf Refs #47. On Loongson 3A, set DGEMM_R parameter depending on different number of threads. It would improve double precision BLAS3 on multi-threads. 2011-09-05 15:13:52 +00:00
traits
90481ce742 Updated the doc about 0.1alpha2.3. 2011-09-05 17:40:55 +08:00
traits
9fc6764fa7 refs #55. Added DTB_ENTRIES into dynamic arch setting parameters. Now, it can read DTB_ENTRIES on runtime. 2011-09-05 17:37:07 +08:00
traz
74d4cdb81a Fix an illegal instruction for strmm_RTLU. 2011-09-02 19:41:06 +00:00
traz
7906146836 Fix an error for strmm_LLTN. 2011-09-02 16:57:33 +00:00
traz
3274ff47b8 Fix an error for strmm_LLTN. 2011-09-02 16:50:50 +00:00
traz
a059c553a1 Fix a compute error for strmm. 2011-09-02 16:00:04 +00:00
traz
23e182ca7c Fix stack-pointer bug for strmm. 2011-09-02 15:28:01 +00:00
traz
a15bc95824 Add strmm part. 2011-09-02 09:15:09 +00:00
traz
74a3f63489 Tuning mb, kb, nb size to get the best performance. 2011-09-01 17:15:28 +00:00
traz
09f49fa891 Using PS instructions to improve the performance of sgemm and it is 4.2Gflops now. 2011-08-31 21:24:03 +00:00
Xianyi Zhang
b9d89f8aaa Fixed the bug about installation. f77blas.h works OK now. 2011-08-31 18:21:37 +08:00
traz
cb0214787b Modify compile options. 2011-08-30 20:57:00 +00:00
traz
2e8cdd1542 Using ps instruction. 2011-08-30 20:54:19 +00:00
traz
b29d327d14 Merge branch 'loongson3a' of github.com:xianyi/OpenBLAS into loongson3a 2011-07-18 17:06:53 +00:00