floraachy/OpenBLAS
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
1,286 Commits 47 Branches 64 Tags
41aad0407f285d0542cebdf1fef601973bab5d1f
Commit Graph

1286 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Zhang Xianyi
41aad0407f Merge pull request #482 from jeromerobert/develop 
Allow to do gemv and ger buffer allocation on the stack
 2015-01-02 02:26:17 +08:00
Zhang Xianyi
f8f2e84659 Merge pull request #486 from wernsaar/develop 
Optimizations for steamroller
 2014-12-31 02:36:23 +08:00
Werner Saar
34633fef01 Merge branch 'develop' of github.com:wernsaar/OpenBLAS into develop 2014-12-30 20:16:53 +08:00
Werner Saar
ddf983d643 added optimizations for steamroller 2014-12-30 20:14:45 +08:00
Zhang Xianyi
17b9db20f1 Merge pull request #483 from wernsaar/develop 
added Steamroller as a  cpu target
 2014-12-29 12:00:16 +08:00
Werner Saar
0dc559ed30 bugfix in dynamic.c 2014-12-28 17:15:42 +01:00
Werner Saar
9566f5fdb0 added Steamroller as a target processor 2014-12-28 13:45:19 +01:00
Werner Saar
4319769b79 added target processor STEAMROLLER 2014-12-28 20:16:46 +08:00
Jerome Robert
e9d9a8eae3 Allow to do gemv and ger buffer allocation on the stack 
ger and gemv call blas_memory_alloc/free which in their turn
call blas_lock. blas_lock create thread contention when matrices
are small and the number of thread is high enough. We avoid
call blas_memory_alloc by replacing it with stack allocation.
This can be enabled with:
make -DMAX_STACK_ALLOC=2048
The given size (in byte) must be high enough to avoid thread contention
and small enough to avoid stack overflow.

Fix #478
 2014-12-27 14:33:12 +01:00
Zhang Xianyi
cbb3ab80e7 Merge pull request #481 from eschnett/develop 
Correct ilaver C declaration
 2014-12-26 10:09:19 +08:00
Erik Schnetter
cd9868b1b4 Correct ilaver C declaration 2014-12-25 17:41:17 -05:00
Zhang Xianyi
eb738148fe Merge pull request #479 from wernsaar/develop 
workaround for sandybridge zgemm kernel
 2014-12-23 00:59:41 +08:00
Werner Saar
587e16fba3 Ref #458: Backport, sandybrigde uses nehalem zgemm kernel 2014-12-22 17:01:18 +01:00
Werner Saar
4de7b9ae47 increased NMAX to 128 2014-12-22 14:04:27 +01:00
Werner Saar
887aed634d modified sources for OS Darwin 2014-12-19 12:40:46 +01:00
Werner Saar
6261342de3 small optimization on dgemm_kernel for N=1 2014-12-18 20:35:51 +01:00
Werner Saar
1e566223ed added code for the size of n 2014-12-17 15:02:11 +01:00
Werner Saar
113b48ca22 modified makefile for acml6.1 2014-12-17 14:12:21 +01:00
Zhang Xianyi
3e81c99b6b Fixed installation bug on Mac OSX. 2014-12-13 13:05:06 +08:00
Werner Saar
ec85c4a51d Increased the Threshold value in sep.in 2014-12-11 14:57:41 +01:00
Werner Saar
97de657d38 added tests to sep.as as workaround for gfortran-4.8.x 2014-12-11 13:53:59 +01:00
Zhang Xianyi
71966eba6c Merge pull request #475 from xantares/patch-2 
add OpenBLAS_VERSION to cmake config file
 2014-12-09 17:57:43 +08:00
Zhang Xianyi
a359979e17 Merge pull request #474 from xantares/patch-1 
set OPENBLAS_CMAKE_DIR to <prefix>/lib/cmake/<package_name>
 2014-12-09 17:57:16 +08:00
xantares
7a6a141bc4 add OpenBLAS_VERSION to cmake config file 2014-12-09 10:34:41 +01:00
xantares
b8ff6892f6 set OPENBLAS_CMAKE_DIR to <prefix>/lib/cmake/<package_name> 
usually these files are more often located in this subdir
 2014-12-09 10:18:18 +01:00
Zhang Xianyi
8fe7a9ce6f Merge pull request #473 from wernsaar/develop 
changed inline assembler labels to short form
 2014-12-08 13:22:18 +08:00
Werner Saar
bc5fff7085 changed inline assembler labels to short form 2014-12-07 12:38:54 +01:00
Zhang Xianyi
1943ea91a8 Merge branch 'develop' of github.com:xianyi/OpenBLAS into develop 2014-12-03 23:03:48 +08:00
Zhang Xianyi
f5424fc9de Update the doc for 0.2.13 version. 2014-12-03 23:00:29 +08:00
Zhang Xianyi
0cf29ba6d2 Fixed a bug of sgemm sandy bridge kernel. 
Reported by Julia project. JuliaLang/julia#9084
 2014-12-03 17:38:41 +08:00
Zhang Xianyi
50e18033e6 Merge pull request #471 from nolta/patch-4 
c_check: set $hostarch to x86_64 instead of amd64
 2014-12-03 12:53:20 +08:00
Zhang Xianyi
551b55d1c7 Merge pull request #470 from nolta/patch-3 
fix fortran compiler detection on FreeBSD
 2014-12-03 12:50:46 +08:00
Mike Nolta
271ceb8bae c_check: set $hostarch to x86_64 instead of amd64 
`uname -m` returns "amd64" on some systems.
 2014-12-02 21:23:23 -05:00
Mike Nolta
5f846be2e4 fix fortran compiler detection on FreeBSD 
On FreeBSD, passing extra options to `which` causes it to report a non-zero status:

```
$ which gfortran48 -m64
/usr/local/bin/gfortran48
$ echo $?
1
```

```
$ which gfortran48
/usr/local/bin/gfortran48
$ echo $?
0
```
 2014-12-02 20:47:40 -05:00
Zhang Xianyi
fe7dcf98f3 Refs #461. Provide OpenBLASConfig.cmake to support CMake. 
If you "make PREFIX=/path/to/OpenBLAS install" ,
The config file will be located in /path/to/OpenBLAS/cmake

Then, you can use "find_package(OpenBLAS)" at CMake.
cmake -DOpenBLAS_DIR=/path/to/OpenBLAS/cmake ..
 2014-11-29 02:16:40 +08:00
Zhang Xianyi
2fb02626da Update organization info. 2014-11-25 15:28:58 +08:00
Zhang Xianyi
a85c2785ae Refs #467. Added generic kernel file for x86_64. 2014-11-24 15:34:48 +08:00
Zhang Xianyi
4806715c97 Fixed #456. Merged the optimizations for APM's 
xgene-1 (aarch64).
Merge branch 'benedikt-huber-dave-patch' into develop
 2014-11-11 22:21:04 +08:00
Benedikt Huber
58c90d5937 # The first commit's message is: 
Optimizations for APM's xgene-1 (aarch64).

1) general system updates to support armv8 better.  Make all did not work, one needed to supply TARGET=ARMV8.
2) sgem 4x4 kernel in assembler using SIMD, and configuration changes to use it.
3) strmm 4x4 kernel in C.  Since the sgem kernel does 4x4, the trmm kernel must also do 4xN.

Added Dave Nuechterlein to the contributors list.
 2014-11-11 22:19:23 +08:00
Zhang Xianyi
2987bc7b40 refs #464. Fixed the bug of detecting L2 associative on x86. 2014-11-10 17:15:34 +08:00
Zhang Xianyi
695e0fa649 #463 fixed a compiling bug on AIX. 2014-11-10 14:39:56 +08:00
Zhang Xianyi
cbb23c46c2 Merge pull request #459 from tkelman/symbol-rename 
add SYMBOLPREFIX and SYMBOLSUFFIX makefile options
 2014-10-25 19:49:03 +08:00
Tony Kelman
0b4602b753 add SYMBOLPREFIX and SYMBOLSUFFIX makefile options 
for adding a prefix or suffix to all exported symbol names in the shared library
Useful to avoid conflicts with other BLAS libraries, especially when using
64 bit integer interfaces in OpenBLAS

Note that since OSX does not have the objcopy utility, setting these options
to non-empty values on Mac requires the objconv tool, available (GPL license)
from http://www.agner.org/optimize/#objconv
 2014-10-24 22:27:09 -07:00
Zhang Xianyi
ac5a7e1c1b Update dot to 0.2.12 version. 2014-10-13 17:10:12 +08:00
wernsaar
f1b9a4a1ca Ref #454: fixed bug in common_param.h 2014-09-23 11:34:29 +02:00
Zhang Xianyi
ae6b7caf32 Merge pull request #453 from wernsaar/develop 
Enabled GEMM3M functions
 2014-09-22 16:47:54 +08:00
wernsaar
f446d2368a updated cblas.h and cblas_noconst.h 2014-09-21 13:39:15 +02:00
wernsaar
dab4edd069 added benchmark for gemm3m functions 2014-09-21 12:00:41 +02:00
wernsaar
9d7057366d bugfix for GEMM3M functions 2014-09-21 11:41:43 +02:00
wernsaar
7f234f8ed1 added GEMM3M tests 2014-09-21 10:55:08 +02:00