Skip to content

Pull requests: NVIDIA/cutlass

New pull request New
90 Open 371 Closed
90 Open 371 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Add high-performance FP16 GEMM kernel with TMA for Blackwell(SM120a)
#3187 opened Apr 26, 2026 by FY-26 Loading…
1
doc: fix stride typo in layout algebra composition example
#3186 opened Apr 26, 2026 by leonardHONG Loading…
1
CuTe DSL: add SM120 NVFP4 warp MMA unpack support
#3185 opened Apr 24, 2026 by alecco Loading…
Add Snake activation functor for EVT
#3184 opened Apr 24, 2026 by emre570 Loading…
1
1
Fix CuTe composition stride-divisibility check (#3177)
#3181 opened Apr 21, 2026 by jduprat Loading…
1
[CuTeDSL] Fix random_ and normal_ ops to support torch.compile fullgraph
#3175 opened Apr 19, 2026 by Flink-ddd Loading…
1
2
Fix Thor MLA decode arch dispatch
#3173 opened Apr 17, 2026 by iloveai8086 Loading…
WIP: OSS CI Testing for v4.5
#3171 opened Apr 16, 2026 by zekunf-nv Collaborator Loading…
3
[CMake] Fix SM70/72 support range upper bound for CUDA 13.0
#3166 opened Apr 13, 2026 by Flink-ddd Loading…
1
4
[CuTeDSL] Update atomic_max_float32 to atomic_fmax in blockscaled GEMM amax example
#3155 opened Apr 8, 2026 by questa-wang Contributor Loading…
1
Fix incorrect example paths in CuTeDSL docstrings
#3151 opened Apr 6, 2026 by Weili-0234 Loading…
[CuTeDSL] Fix dynamic shape args not passed to JIT kernel
#3148 opened Apr 5, 2026 by Flink-ddd Loading…
1
3
clamp max_workers to at least 1 for single-cpu systems
#3146 opened Apr 5, 2026 by knQzx Loading…
1
[Fix] distributed gemm all reduce and reduce scatter examples
#3143 opened Apr 1, 2026 by Nicyzk Loading…
Fix Hopper FMHA performance regression on CUDA < 13.1
#3137 opened Mar 31, 2026 by arvin-chou Loading…
5 of 6 tasks
feat(CuTeDSL): print benchmark time from Blackwell dense_gemm CLI
#3136 opened Mar 30, 2026 by aidando73 Contributor Loading…
1
[CuTeDSL] Add a render function hook to allow render layout natively
#3135 opened Mar 30, 2026 by kainzhong Loading…
Fix incorrect warp arrangement in PitchLinearWarpStripedThreadMap inactive-30d
#3131 opened Mar 25, 2026 by HaoYuan-Gao Loading…
1
Fix elementwise_apply.py inactive-30d
#3129 opened Mar 25, 2026 by HydraQYH Contributor Loading…
1
[CuTeDSL] Add SM103 grouped block-scaled GEMM kernel and tests inactive-30d
#3124 opened Mar 23, 2026 by Johnsonms Contributor Loading…
1
Enable strict C++ compiler warnings with -Werror inactive-30d
#3123 opened Mar 22, 2026 by maxwbuckley Loading…
3 of 4 tasks
1
[FMHA] Add SM110 support for Blackwell FMHA example (77_blackwell_fmha) inactive-30d
#3112 opened Mar 18, 2026 by LiangSu8899 Loading…
1
fix(CuTeDSL): correct FP4 tensor K dimension in grouped blockscaled GEMM inactive-30d
#3102 opened Mar 13, 2026 by Hale423 Loading…
1
1
[CuTe] adding DSL INT8 MMA support on SM80+, with CuTe C++ example
#3097 opened Mar 9, 2026 by ayghri Loading…
2
Add dlopen-based dynamic kernel loading for profiler inactive-30d
#3088 opened Mar 3, 2026 by Wazrrr Loading…
1
ProTip! Type g p on any issue or pull request to go back to the pull request listing page.