NVIDIA/cutlass
1
0
Fork 0
mirror of https://github.com/NVIDIA/cutlass.git synced 2026-05-12 09:15:56 +00:00
Code Issues Packages Projects Releases Wiki Activity
776 Commits 21 Branches 52 Tags
ebf3165efba92012b4f0eff001bc22d9a4e2af39
Commit Graph

776 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Qi Yuhang
ebf3165efb [Bug Fix]Bypass launch grids for SM120 Kernel with SM90 Mainloop & SM100 TileScheduler (#2865) 
* Delete unused #ifdef/#endif. Bypass sm120 case.

* Add todo.

* Fix pingpong.

* Revert "Add todo."

This reverts commit 246cb42091.

* Refine name.

Refine name again.

* Apply suggestions from code review

Skip `is_last_tile` for all sm120 kernels.

Co-authored-by: Junkai-Wu <junkaiw@nvidia.com>

* Skip early stop for sm120 kernel.

* Fix typo.

---------

Co-authored-by: Junkai-Wu <junkaiw@nvidia.com>
 2025-12-18 08:51:38 +08:00
Haicheng Wu
d4e16f5d4e Bump version from 4.2.1 to 4.3.3 2025-12-11 23:58:38 -05:00
Junkai-Wu
d3a5492381 v4.3.3 update. (#2868) 2025-12-11 00:26:58 -05:00
Amin Sedaghat
49bd6bf1ba fix print_layout printf format in device code (#2688) 
* fix print_layout printf format in device code

* Replace %.*s format specifier with explicit loop
* Remove unused delim variable

The printf format %.*s with dynamic width does not work correctly
in CUDA device code, causing literal %.*s to appear in output.

Fixes #2496

* Update include/cute/util/print_tensor.hpp

Co-authored-by: Cris Cecka <ccecka@users.noreply.github.com>

* Update include/cute/util/print_tensor.hpp

Co-authored-by: Cris Cecka <ccecka@users.noreply.github.com>

---------

Co-authored-by: Cris Cecka <ccecka@users.noreply.github.com>
 2025-12-10 08:57:56 +08:00
Junkai-Wu
c6dfdf8375 Merge pull request #2719 from HydraQYH/dev_support_pdl_for_sm90_gemm_array_tma_ws 
Support PDL for SM90 Array TMA GEMM
 2025-12-09 18:32:15 +08:00
HydraQYH
95f8beb44c Revert "Remove unnecessary #ifdef #endif for general gemm." 
This reverts commit 17ffd56dfe.
 2025-12-09 11:52:23 +08:00
HydraQYH
17ffd56dfe Remove unnecessary #ifdef #endif for general gemm. 2025-12-06 10:20:35 +08:00
HydraQYH
ff7f2dcdfb Remove duplicated cutlass::arch::wait_on_dependent_grids(); 2025-12-06 10:20:35 +08:00
HydraQYH
929e1e0259 Remove unnecessary #ifdef / #endif for launch_dependent_grids. 2025-12-06 10:20:35 +08:00
HydraQYH
b6ad6db219 Delete unnecessary #ifdef / #endif. 2025-12-06 10:20:35 +08:00
HydraQYH
e1b2ec57e3 Hoist waits above the warp specialized region. 2025-12-06 10:20:35 +08:00
HydraQYH
1e5f95cbbe Support PDL in sm90_gemm_array_tma_warpspecialized_cooperative 2025-12-06 10:20:35 +08:00
HydraQYH
acf5990cc2 Refine position for wait_on_dependent_grids. 2025-12-06 10:20:35 +08:00
HydraQYH
91de7891a5 Support PDL in sm90_gemm_array_tma_warpspecialized_pingpong.hpp 2025-12-06 10:20:35 +08:00
Haicheng Wu
1e7d3e030b Update CHANGELOG for CuTe DSL enhancements 
Added new environment variable for CuTe DSL cache directory.
 2025-12-05 13:49:57 -05:00
Haicheng Wu
c4744f706e Bump version from 4.2.1 to 4.3.2 2025-12-05 13:45:16 -05:00
Linfeng Zheng
f6402fcd5e add pytest support for tutorial gemm (#2826) 
* add pytest support for tutorial gemm

* add license
 2025-12-05 08:45:01 -05:00
bangyu shen
7252a2d17e remove internal comment (#2841) 
Co-authored-by: bangyus <bangyus@nvidia.com>
 2025-12-04 10:36:21 -05:00
Junkai-Wu
bc680c7f67 v4.3.2 update. (#2839) 2025-12-04 10:14:32 -05:00
bangyu shen
52ae719eda [examples][CuTeDSL] init commit for distirbuted examples (#2806) 
* init commit for distirbuted examples

* better OOB protection

* and try import to nvshmem for better error message and a READMME.md to introduce nvshmem and multimem instructions

* add some lamport explanation

* enhance f8 output and warn that f8 output can have nan in it

* tell user why we need complicate data conversions in ref check part

* tell user we don't support nvshmem device function

---------

Co-authored-by: bangyus <bangyus@nvidia.com>
 2025-12-01 22:25:40 -05:00
Haicheng Wu
5e847d37c4 Bump version from 4.2.1 to 4.3.1 2025-12-01 22:13:19 -05:00
Haicheng Wu
f16068b4db Bump version from 4.2.0 to 4.3.1 2025-12-01 22:12:20 -05:00
Haicheng Wu
1acfe141af Bump version from 4.2.1 to 4.3.1 2025-12-01 22:11:13 -05:00
Haicheng Wu
f11375bf91 Bump CUTLASS patch version to 1 2025-12-01 22:08:52 -05:00
Shreya Gaur
af8d5dfa54 bug fix for example 92 (#2830) 
Co-authored-by: Shreya Gaur <shgaur@dc2-container-xterm-012.prd.it.nvidia.com>
Co-authored-by: Shreya Gaur <shgaur@2u2g-spr-0015.ipp4a1.colossus.nvidia.com>
 2025-12-01 22:02:59 -05:00
drazi
ec8daf642d Merge pull request #2809 from whatdhack/patch-1 
Update notebook title from 'Tour to' to 'Tour of'
 2025-11-28 18:07:34 +08:00
drazi
5016493cc0 Merge pull request #2813 from fengxie/ftse/fix/example 
Refactor TVM FFI examples and update doc
 2025-11-28 09:07:15 +08:00
Fung Xie
8588d099e4 refactored doc 2025-11-27 17:04:20 -08:00
Fung Xie
8fc9bc5dda update doc 2025-11-27 17:03:51 -08:00
Fung Xie
f71892b824 update doc 2025-11-27 17:03:03 -08:00
Fung Xie
03aa211310 update doc 2025-11-27 17:02:59 -08:00
Fung Xie
286781a1fb add requirements.txt 2025-11-27 17:02:27 -08:00
Fung Xie
2664cac685 enhanced the example for tvm-ffi 2025-11-27 17:02:26 -08:00
Fung Xie
b9154d65b3 update examples for tvm-ffi 2025-11-27 17:02:26 -08:00
Fung Xie
afe2f71522 reorganize examples for tvm-ffi 2025-11-27 17:02:26 -08:00
Fung Xie
739fffce27 fix TVM FFI doc and update example 2025-11-27 17:02:26 -08:00
Junkai-Wu
1de3a576cc v4.3.1 update. (#2817) 2025-11-27 09:49:30 -05:00
Shreya Gaur
2052fd3885 Blockscaled Ragged Contiguous Grouped Gemm for MoEs (#2790) 
* Adding blockscaled ragged contiguous grouped gemm for MoEs

* cleaning up the example

* introduction to example improved

---------

Co-authored-by: Shreya Gaur <shgaur@dc2-container-xterm-012.prd.it.nvidia.com>
 2025-11-26 20:16:49 -05:00
whatdhack
4a55379686 Update notebook title from 'Tour to' to 'Tour of' 
Grammar check . LLM's can quickly clean up such issues.
 2025-11-24 20:11:14 -08:00
Haicheng Wu
e67e63c331 Bump version from 4.2.1 to 4.3.0 v4.3.0 2025-11-24 16:36:06 -05:00
Haicheng Wu
ddaf12c1b1 Bump version from 4.2.0 to 4.3.0 2025-11-24 16:35:27 -05:00
Haicheng Wu
7967ce5f83 Bump version to 4.3.0 2025-11-24 16:34:45 -05:00
Junkai-Wu
8cd5bef43a v4.3 tag release update. (#2789) 2025-11-20 20:49:44 -05:00
Linfeng Zheng
406e078b29 add a notebook for tour to sol gemm (#2780) 
* add tour to sol gemm notebook

* change some typos

* change some typos
 2025-11-20 09:41:01 -05:00
Zekun Fan
a2439551c7 Fixed editable install to depend on CuTeDSL/requirements.txt (#2768) 
To guarantee wheel version alignment of the source code.
 2025-11-14 15:31:49 -08:00
drazi
bd96096d58 Merge pull request #2758 from limin2021/delete-pdl-example 
[cute-dsl][fix]remove cute dsl pdl example.
 2025-11-10 22:56:57 +08:00
Mindy Li
06b6bd7d7b remove cute dsl pdl example. 2025-11-09 21:47:00 -08:00
Linfeng Zheng
2252254ce2 Add tutorial fp16_gemm_1 (#2750) 
* Add tutorial fp16_gemm_1

* refine

* refine

* refine

* revert changes in fp16_gemm_0.py
 2025-11-06 22:40:09 -05:00
Ali Hassani
d1ef0e87f2 DistGEMM bug fixes (#2713) 
* Blackwell DistGEMM bug fixes

1. If using preferred cluster, there needs to be a branch so that
   the universal GEMM wrapper finds the correct base params.
2. Workspace sizes can change depending on problem shape in Blackwell,
   and DistGEMM was previously using the per-device shape to evaluate
   workspace size instead of the per-gemm shape.
3. Flattened size used to initialize host tensors can overflow (in
   Hopper example as well)
4. Preferred and fallback cluster args need to be set explicitly,
   otherwise if someone modifies the example to use preferred cluster,
   it will just fail.

* Fix example runtimes

* Set default fallback cluster shapes to the static ones
 2025-11-06 13:31:24 -05:00
ANIKET SHIVAM
020c700e97 support for K=0 for sm100 GG (#2746) 2025-11-04 11:25:39 -05:00