gaspersic/triton
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
278 Commits 14 Branches 9 Tags
29bdf7f546453b1ce8b48e3ac7197c433a45ce23
Commit Graph

50 Commits

Author SHA1 Message Date
Philippe Tillet
29bdf7f546 Code quality: made the backend static 2015-07-30 21:15:37 -07:00
Philippe Tillet
89ee015f7f General: Bugfixes here and there 2015-07-27 11:37:19 -07:00
Philippe Tillet
a2b533b9a8 Driver: made cl and cu attributes private in Handle<> 2015-07-23 09:40:18 -07:00
Philippe Tillet
cbe930398e Code quality: Cleaned directory folder, variable names and mingw compliance 2015-07-21 13:29:23 -04:00
Philippe Tillet
f4615446c5 GEMM: More optimizations 2015-07-18 17:23:53 -04:00
Philippe Tillet
6ccf32904a GEMM: Still optimizing 2015-07-18 16:06:17 -04:00
Philippe Tillet
753a9b1f3e Benchmarks: now benchmaring all AlexNet sizes 2015-07-14 13:33:23 -04:00
Philippe Tillet
281fa9c7a6 Benchmarks: Now testing AlexNet's size 2015-07-10 16:05:28 -04:00
Philippe Tillet
e7cabf65ac Tuning: Merged tune branch. 
- Much cleaner and more concise source
- Better exceptions handling
- Checks local minima to see if retuning is needed.

Resolved conflicts:
	bench/blas.cpp
	include/isaac/backend/templates/mproduct.h
	include/isaac/driver/buffer.h
	lib/array.cpp
	lib/backend/templates/mproduct.cpp
	lib/driver/buffer.cpp
	python/setup.py
	tune/pysrc/autotune.py
	tune/pysrc/dataset.py
	tune/pysrc/misc_tools.py
 2015-06-28 17:53:16 -07:00
Philippe Tillet
b32de3ac76 C++: More clBLAS routines 2015-06-25 08:12:16 -07:00
Philippe Tillet
9f7e34ba5d C++: Added clBLAS sGEMM ABI (still buggy) 2015-06-24 07:51:27 -07:00
Philippe Tillet
a67476671d Bench: Removed warnings in bench-blas when no external blas is defined 2015-04-29 16:11:32 -04:00
Philippe Tillet
cf5028d55b Squashed feature branch: 
* Added CUDA support
 * Performance improvements
 * API improvements
 * Added "depth" parameter to GEMM
 * Android cross-compilation
 2015-04-29 15:52:21 -04:00
Philippe Tillet
5ff16bfcb6 Added cublas sgemm 2015-02-13 04:31:42 -05:00
Philippe Tillet
e453031094 More efficient access pattern in the GEMV kernel 2015-02-11 02:06:16 -05:00
Philippe Tillet
37fc98c532 Fixed bug on marix-vector products with vectorization 2015-02-10 03:09:41 -05:00
Philippe Tillet
a89f6d88be Fix bug in operation-specific tuning 2015-02-09 01:58:32 -05:00
Philippe Tillet
a6d7671831 removing C++11 interface 2015-02-08 23:19:38 -05:00
Philippe Tillet
85fb438806 More convenient use of specific runtime tuning 2015-02-08 14:23:38 -05:00
Philippe Tillet
9c68704f09 Now using a list of event instead of a single one 2015-02-08 00:56:24 -05:00
Philippe Tillet
b768e913c9 Now using events to time autotuning 2015-02-06 22:11:03 -05:00
Philippe
7fc2348924 Fixed CUDA benchmark 2015-02-05 23:42:31 -05:00
Philippe Tillet
8f8b01938b Rearranged benchmarking script 2015-02-05 23:11:16 -05:00
Philippe Tillet
e214927b16 Better control flow through options 2015-02-05 04:43:50 -05:00
Philippe Tillet
bbf2f0188e Ported to C++11 2015-02-05 04:43:40 -05:00
Philippe Tillet
3a296ae3b7 Added a control flow API 2015-02-03 15:25:01 -05:00
Philippe Tillet
939ce15b45 Cleaner benchmarking code 2015-02-02 00:03:48 -05:00
Philippe Tillet
2afc574724 Implemented simple operation cache 2015-02-01 23:56:05 -05:00
Philippe Tillet
535706f35a Some renaming; lower overhead in benchmark 2015-02-01 22:28:49 -05:00
Philippe Tillet
3b61842528 Lower overhead in the benchmarking source code 2015-02-01 18:59:27 -05:00
Philippe Tillet
3f1fa822f8 save 2015-02-01 15:58:05 -05:00
Philippe Tillet
d29f1252ad Clearer array_expression with hopefully lower overhead. 
Also removed pyc's
 2015-01-31 22:01:48 -05:00
Philippe Tillet
f488274269 Fixed relational operators tests 2015-01-29 16:01:46 -05:00
Philippe Tillet
c7665021d1 reducing overhead; reverted custom CL/ header because CL/cl.hpp was buggy 2015-01-28 23:04:19 -05:00
Philippe Tillet
1246fbe9a8 More portable synchronization in blas-bench 2015-01-28 20:06:41 -05:00
Philippe Tillet
a6c513014f Silly bugfix in cublas saxpy 2015-01-28 19:54:36 -05:00
Philippe Tillet
0dcf4d3617 Better exception handling, lowered CMake requirement ; blas-bench now benchmarks square matrices 2015-01-28 17:08:39 -05:00
Philippe Tillet
c37d8a2a81 Now using system CL include 2015-01-27 16:19:50 -05:00
Philippe Tillet
c12ec4cebd tentative cuda benchmark integration 
x
 2015-01-27 15:32:59 -05:00
Philippe Tillet
a96c897cb3 Various fixes 2015-01-27 02:41:27 -05:00
Philippe Tillet
6d7f103ab1 Added GEMM benchmark 2015-01-25 18:19:19 -05:00
Philippe Tillet
da9defac04 No more random order in default initialization 
Added linking to MKL/CBLAS
 2015-01-24 14:51:48 -05:00
Philippe Tillet
69311b7982 Now ATIDLAS is standalone. Everything dynamic.... 2015-01-12 13:24:06 -05:00
Philippe Tillet
8dd492de23 Now using ViennaCL for reference instead of "naive" 2014-11-06 16:15:22 -05:00
Philippe Tillet
146f54b32c Improved benchmark suite 2014-11-06 07:07:27 -05:00
Philippe Tillet
e576d12d66 Added ViennaCL time 2014-10-30 14:20:14 -04:00
Philippe Tillet
f0a4255daa Now actually benchmarking axpy 2014-10-30 14:11:50 -04:00
Philippe Tillet
71b4bde3ec Better benchmarking 2014-10-30 13:04:33 -04:00
Philippe Tillet
83bce648a0 Now using the median for benchmarking 2014-10-29 17:03:24 +01:00
Philippe Tillet
740ba08c5c Added basic BLAS1 benchmark 2014-10-27 05:35:04 -04:00