Skip to content

Pull requests: NVIDIA/cutlass

New pull request New
179 Open 913 Closed
179 Open 913 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

Update MLA decode example
#3357 opened Jun 30, 2026 by Observer007 Contributor Loading…
Add: helper function for rollback and verification.
#3356 opened Jun 29, 2026 by Soojal1807 Loading…
[CuTeDSL] Drop Ampere/Ada from cvt_i8_bf16_intrinsic.supported_archs
#3355 opened Jun 29, 2026 by Hughshine Loading…
1
Fix c_pointer mem leak
#3352 opened Jun 26, 2026 by Difers Loading…
Warn on deprecated CooperativeGroup alignment argument
#3347 opened Jun 24, 2026 by xiufanl Contributor Loading…
[SM120] Emit grouped NVFP4 (ue4m3) block-scaled GEMM in cutlass_library
#3346 opened Jun 23, 2026 by waynehacking8 Loading…
1
Fix SM120 blockscaled FP32 epilogue store layout
#3345 opened Jun 23, 2026 by NVIDIA-JerryChen Loading…
fixes for #3340, stale comments on thread layout shape
#3341 opened Jun 22, 2026 by SpookOrSpooky Loading…
1
Return error for invalid runtime datatypes
#3339 opened Jun 21, 2026 by fallintoplace Loading…
Harden Python profiler command handling
#3338 opened Jun 21, 2026 by fallintoplace Loading…
Fix typo in 0t_mma_atom.md
#3335 opened Jun 19, 2026 by dukallis Loading…
[CuTeDSL] Fix make_ptr ZeroDivisionError on sub-byte dtypes
#3334 opened Jun 19, 2026 by waynehacking8 Loading…
[CuTeDSL] Fix cute.is_major crash on List[int] mode
#3333 opened Jun 19, 2026 by waynehacking8 Loading…
[CuTeDSL] Fix unsigned integer constant lowering above INT64_MAX (#3312)
#3332 opened Jun 19, 2026 by waynehacking8 Loading…
1
[CuTeDSL] fix mha support for thor
#3323 opened Jun 16, 2026 by HuPengsheet Loading…
insignificant typo
#3320 opened Jun 15, 2026 by igodlab Loading…
[CuteDSL] Fix source location info for cute.arch.elect_one
#3317 opened Jun 14, 2026 by pchen7e2 Loading…
[Tutorial] Fix race condition in 2SM TMEM alloc for Cute Blackwell Tutorial 04/05
#3316 opened Jun 14, 2026 by pchen7e2 Loading…
[docs/examples] Blackwell tutorial 04 (2SM MMA): fix stale B-operand print annotations (the 2x1SM atom splits B N/2-per-CTA) and document the expect-tx pitfalls that deadlock
#3314 opened Jun 11, 2026 by cfregly Loading…
[docs/examples] Blackwell cute tutorials: narrow TMEM_LOAD atoms (32dp32b1x) carry a large per-load lowering cost — prefer wider atoms (32dp32b32x) in t2r epilogues
#3313 opened Jun 11, 2026 by cfregly Loading…
[CuTeDSL] Fix _ScalarData internal methods triggering its own struct.scalar deprecation
#3311 opened Jun 10, 2026 by Johnsonms Contributor Loading…
quick fix for fmha_bwd regressions
#3309 opened Jun 9, 2026 by Jie-Fang Loading…
1
Fix CUTLASS Blackwell FMHA register spills on DRIVE Thor (sm_110a)
#3308 opened Jun 8, 2026 by pzhao-eng Loading…
Fixed integer overflow in make_cute_packed_stride batch stride computation
#3307 opened Jun 7, 2026 by a123pal Loading…
1
Fix QNX fill_workspace status
#3306 opened Jun 6, 2026 by fallintoplace Loading…
ProTip! Filter pull requests by the default branch with base:main.