Commit Graph

  • 1e4703cbab Support parallel split K mode for porfiling (#277) Fujun Han 2022-01-27 23:37:37 +08:00
  • c3353add63 Merge pull request #388 from depaulmillz/fix/headersonly Dustyn Blasig 2022-01-26 14:22:51 -06:00
  • ac8825b941 Minor fix to change from LIBRARY_INIT to LIBRARY dePaul Miller 2022-01-26 15:17:46 -05:00
  • 8fd94806e5 Update PUBLICATIONS.md Haicheng Wu 2022-01-17 00:08:18 -05:00
  • d7c9cbf0b9 Fix typo in scripts/library.py (wrong data size for u8) (#393) Masahiro Masuda 2022-01-08 03:29:56 +09:00
  • c2ee13a0fe Add epilogue functor for residual block fusion (#391) masahi 2021-12-30 12:53:40 +09:00
  • f78994bb40 add the missing pieces (#392) Haicheng Wu 2021-12-25 07:29:54 -05:00
  • dceabd4c5a Support half precision sigmoid activation (#378) masahi 2021-12-23 04:45:06 +09:00
  • 86fa1dc30b Fix utils include not being installed in header only dePaul Miller 2021-12-21 12:10:26 -05:00
  • 288af365db Added missing synchronization to avoid WAR hazards between tiles. (#386) Andrew Kerr 2021-12-20 11:34:08 -05:00
  • 0dc3ba60b3 Refactor GELU and Sigmoid epilogue to use a common template (and add SiLu, Hardswish epilogue) (#379) masahi 2021-12-19 04:58:15 +09:00
  • ec4f7e5194 Updates to fused epilogue (#383) Andrew Kerr 2021-12-17 16:04:43 -05:00
  • 4e666e1dfd Updated README and added issue templates. (#382) Andrew Kerr 2021-12-17 09:26:20 -05:00
  • 3799e12f25 Merge pull request #381 from Peter9606/update-makefile-version Haicheng Wu 2021-12-16 21:54:57 -05:00
  • fc3bc85db8 Update project version to 2.8.0 in CMakeLists.txt Peter Han 2021-12-17 02:23:31 +00:00
  • 49c0a58d50 Set theme jekyll-theme-minimal Matthew Nicely 2021-12-15 14:51:24 -05:00
  • 5fe09c2d67 Updated GEMM performance plot with CUTLASS 2.8 compiled with CUDA 11.5 Toolkit (#375) Andrew Kerr 2021-12-06 14:21:33 -05:00
  • 6b69c79ac3 Fixed contributor formatting. (#365) Andrew Kerr 2021-11-22 14:30:53 -05:00
  • 62e438f450 Listed Matthew Nicely as the CUTLASS product manager.. (#364) Andrew Kerr 2021-11-19 20:51:21 -05:00
  • 808c25337a CUTLASS 2.8 (#363) Manish Gupta 2021-11-19 13:26:35 -08:00
  • 6fc5008803 Update quickstart.md Haicheng Wu 2021-11-11 09:53:46 -05:00
  • a3bcc6981d Merge pull request #331 from reed-lau/feature/fix-wmma-shape-typo Haicheng Wu 2021-09-28 10:20:29 -04:00
  • 3b28642801 fix wmma shape typo reed-lau 2021-09-28 19:04:09 +08:00
  • 538592dea4 example 23 gemm operand reduction fusion (#325) Manish Gupta 2021-09-20 13:34:47 -07:00
  • 2e07c4cc2f CUTLASS 2.7 (#318) Manish Gupta 2021-09-20 11:02:22 -07:00
  • 9ac255863f Merge pull request #246 from mengchihe/master Haicheng Wu 2021-09-08 11:40:53 -04:00
  • 59e2aa505a refine the implementation Haicheng Wu 2021-09-08 13:14:08 +00:00
  • 4e8af93da1 Merge remote-tracking branch 'origin/master' into small_alignment Haicheng Wu 2021-09-07 20:39:38 +00:00
  • 6c2f8f2fb8 CUTLASS 2.6.1 - functional and performance enhancements to strided DGRAD, fixes, and tuning Manish Gupta 2021-09-03 10:26:15 -07:00
  • 598e35401c Merge remote-tracking branch 'origin/master' into small_alignment Haicheng Wu 2021-08-16 07:49:08 -07:00
  • a01feb93d9 Merge pull request #308 from dongxiao92/patch-1 Manish Gupta 2021-08-08 11:54:42 -07:00
  • d36f331b44 fix typo in doc dongxiao 2021-08-08 16:44:22 +08:00
  • 69abafb85a Merge pull request #306 from NVIDIA/fix-profiler-cmd-doc Haicheng Wu 2021-07-30 14:36:54 -04:00
  • 68a078fbbf cleanup Haicheng Wu 2021-07-30 11:27:21 -07:00
  • 10709dbb64 clean profiler cmd and doc Haicheng Wu 2021-07-30 11:02:17 -07:00
  • 1227351079 Merge pull request #305 from NVIDIA/fix_epilogue_spill Manish Gupta 2021-07-29 14:30:11 -07:00
  • a77c658439 fix epilogue register spill Haicheng Wu 2021-07-29 14:25:48 -07:00
  • 4516b833ce Merge pull request #303 from Peter9606/doc_typo Haicheng Wu 2021-07-28 20:49:06 -04:00
  • 64dd1e1915 Doc typo Peter Han 2021-07-29 08:43:27 +08:00
  • 1ac4559d12 Cutlass 2.6 Update 1 (#301) Manish Gupta 2021-07-27 17:58:30 -07:00
  • e5d51840e8 CUTLASS 2.6 (#298) Manish Gupta 2021-07-22 21:40:53 -07:00
  • 6c29fe20ba Merge pull request #285 from tjingrant/patch-1 Haicheng Wu 2021-07-05 22:51:19 -04:00
  • e3c56b0d6b Update predicated_tile_iterator.h Tian Jin 2021-07-05 12:11:53 -04:00
  • 4647c57243 Update predicated_tile_iterator.h Tian Jin 2021-07-05 12:06:41 -04:00
  • 856d4db3fb Update basic_gemm.cu Haicheng Wu 2021-06-15 09:08:36 -04:00
  • 6a1064093f Merge pull request #274 from mani-ananth/master Haicheng Wu 2021-06-02 13:17:39 -04:00
  • c5f1ef4dff update contributors Manikandan Ananth 2021-06-02 10:11:42 -07:00
  • 47ebfccbec bug fixes Manikandan Ananth 2021-06-02 10:08:25 -07:00
  • ad9486684f Merge pull request #272 from BernardoCovas/master Haicheng Wu 2021-05-28 17:18:27 -04:00
  • 1d8372a8e2 fix typo in reference conv3d Bernardo Covas 2021-05-28 21:06:59 +01:00
  • 9cb7d63424 Merge pull request #266 from mani-ananth/master Haicheng Wu 2021-05-19 15:15:22 -04:00
  • da2f110906 Fixes for public issue #265 Manikandan Ananth 2021-05-19 10:16:52 -07:00
  • b68113f5be Merge pull request #264 from zheng95z/patch-3 Haicheng Wu 2021-05-17 10:03:30 -04:00
  • a68d7cd6f1 Adds NoBetaScaling for LinearCombination Zheng Zeng 2021-05-12 22:23:55 +08:00
  • 38e8b29f56 Merge pull request #259 from hzfan/ignore_pr Haicheng Wu 2021-05-10 20:07:53 -04:00
  • ee7349c94f fix Haozheng Fan 2021-05-10 16:39:04 +08:00
  • 8cdd4293d4 add gitignore Haozheng Fan 2021-05-10 16:37:59 +08:00
  • f58b843951 Merge pull request #239 from KeDengMS/kedeng/gelu Haicheng Wu 2021-05-08 12:51:42 -04:00
  • 5fc142296f Merge pull request #237 from Peter9606/issue_236_typo Haicheng Wu 2021-05-08 07:51:19 -04:00
  • 233d69aa6d Merge pull request #235 from Peter9606/issue_233_tranpose_update Haicheng Wu 2021-05-07 07:14:30 -04:00
  • 9840d25269 Merge pull request #256 from zheng95z/patch-2 Haicheng Wu 2021-05-06 11:02:49 -04:00
  • b878c96421 Fixes some typos in utilities.md Zheng Zeng 2021-05-06 22:37:37 +08:00
  • 8f8a80cad5 Merge pull request #251 from zheng95z/patch-1 Haicheng Wu 2021-04-25 22:09:44 -04:00
  • a8f6f8eb07 add a missing 'device_memory::' before a function Zheng Zeng 2021-04-25 20:05:39 +08:00
  • f4b0a33633 add unit test for non int4 load mengchi.hmc 2021-04-23 14:33:46 +08:00
  • 7c783adf53 Merge pull request #247 from xue-fc/patch-1 Haicheng Wu 2021-04-22 09:27:40 -04:00
  • 4000df9567 fix a wrong description xue-fc 2021-04-22 20:28:28 +08:00
  • bb35a3ba6f support setting load granularity for conv2d fprop mengchi.hmc 2021-04-22 15:20:57 +08:00
  • 7ec3a87f22 support unalignment input for conv2d fprop stage=2 Fix for issue #242 mengchi.hmc 2021-04-21 14:28:58 +08:00
  • 0b74c8f473 Address CR KeDengMS 2021-04-19 23:36:06 +00:00
  • 83036ed646 More clean up KeDengMS 2021-04-18 04:29:20 +00:00
  • b7e43f5eb9 Clean up KeDengMS 2021-04-18 04:24:25 +00:00
  • 5c62d892fa Add test KeDengMS 2021-04-18 04:09:34 +00:00
  • 41a31b404b Fixes to Gelu for half and fusion KeDengMS 2021-04-17 22:10:19 +00:00
  • 7320aee17d Typo fix issue#236 Peter Han 2021-04-15 15:08:35 +08:00
  • 2142a05d9d tranpose.h update based on issue#233 Peter Han 2021-04-14 19:58:00 +08:00
  • c77a524459 Merge pull request #230 from mani-ananth/master Haicheng Wu 2021-04-09 14:45:55 -04:00
  • fac6680f31 Merge branch 'master' of github.com:NVIDIA/cutlass Manikandan Ananth 2021-04-09 11:36:31 -07:00
  • 08993707da fixing functional bug in fused epilogue Manikandan Ananth 2021-04-09 11:36:03 -07:00
  • c805593ebe Merge pull request #228 from mani-ananth/master Haicheng Wu 2021-04-08 10:08:13 -04:00
  • 26556d7206 fix a broken sparse gemm example. found by the community. Manikandan Ananth 2021-04-07 13:32:55 -07:00
  • 4839b6cb61 add 2stage fprop 3d into default file Manikandan Ananth 2021-04-07 13:29:32 -07:00
  • d97214987a Merge pull request #220 from Peter9606/wrong-stride-array-definition Haicheng Wu 2021-04-02 08:43:52 -04:00
  • b0bbc6d548 Merge pull request #219 from mani-ananth/master Haicheng Wu 2021-04-02 08:42:09 -04:00
  • 7074047a54 Bugfix: typo, make reduction device cases passed Peter Han 2021-04-02 09:35:23 +08:00
  • 75a4737cfe Fix for public issue #211 - Add a slice-K tile size to the profiler - fix num warps calculations in implicit gemm header Manikandan Ananth 2021-04-01 14:42:00 -07:00
  • 8a3e4b8d02 Merge pull request #214 from Peter9606/separate-stream-error Haicheng Wu 2021-03-24 12:09:01 -04:00
  • 6a6b4028bd Revert wrong fix of params.update in GemmUniversalBase Peter Han 2021-03-23 23:20:40 +08:00
  • 92393b2676 Bugfix: memsetAsync uses wrong default stream Peter Han 2021-03-23 21:11:42 +08:00
  • 50bf00e5f2 Merge pull request #193 from Peter9606/public_shape_type_from_Mma_HFMA2 Haicheng Wu 2021-03-05 21:38:59 -05:00
  • 4cd004ead1 fix test name to optimized and instance large tile sizes to speed unit tests Manish Gupta 2021-03-05 13:32:36 -08:00
  • 6c4539e372 Make arch tag of test cases more precisely to SM60 Peter Han 2021-03-05 10:53:26 +08:00
  • a3639ab1a0 Append fp16 test case to verify Mma_HFMA2 Peter Han 2021-03-04 18:17:57 +08:00
  • 169181f30f Make Shape public from Mma_HFMA2. Peter Han 2021-03-04 11:05:16 +08:00
  • 0f1056390d Create PUBLICATIONS.md (#189) Haicheng Wu 2021-03-03 14:17:40 -05:00
  • 34a42e5620 Update generator.py (#192) Haicheng Wu 2021-03-02 15:21:48 -05:00
  • 8f09b82b12 Merge pull request #187 from NVIDIA/cutlass_2.5 Dustyn Blasig 2021-02-26 23:56:04 -06:00
  • 200a5a5146 Enabled reduction unit tests. Andrew Kerr 2021-02-26 15:46:57 -05:00
  • 746b7b3247 Enabled tensor reduction kernels. Andrew Kerr 2021-02-26 15:32:19 -05:00
  • abdf16a4d9 Updated release notes. Andrew Kerr 2021-02-26 13:55:04 -05:00