Skip to content

Pull requests: NVIDIA/cutlass

New pull request New
98 Open 782 Closed
98 Open 782 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

[CuTeDSL] Distributed example, using TMALDG to access remote memory rank-by-rank, reducing in cta, broadcast result to all ranks by multimem TMASTG
#2970 opened Jan 21, 2026 by aragorn-guan Loading…
1
[NVVM API] Update cutedsl nvvm api change
#2969 opened Jan 21, 2026 by XiaoSong9905 Loading…
3
[Bug Fix]Set NumSplitsM to 1 when TileShapeM < 128 in sm90 fp8 blockwise scaling CollectiveMma
#2965 opened Jan 19, 2026 by HydraQYH Loading…
2
fix(examples): fix device compatibility check for Ada FP8 GEMM
#2954 opened Jan 13, 2026 by w1ndseeker Loading…
Fix out-of-bounds TMA access in wgmma_tma_sm90 tutorial
#2945 opened Jan 10, 2026 by Johnsonms Loading…
1
Update profiler.md with how to use generator.py
#2943 opened Jan 10, 2026 by aidando73 Loading…
cutlass profiler - align emitted SFA/SFB kernel naming with typical convention
#2942 opened Jan 10, 2026 by aidando73 Loading…
[docs] Add additional tip for generating less kernels in blockwise
#2940 opened Jan 9, 2026 by aidando73 Loading…
2
Fix Warp Memory Access Arrangement in Epilogue: Upper Bound memory access width by output tile width
#2938 opened Jan 8, 2026 by lukas-ruettgers Loading…
1
add missing Mma-specialization for m16n8k32.s32.s4.s4.s32
#2936 opened Jan 8, 2026 by dxqb • Draft
Refactor binary_op functions to remove unused result parameter
#2919 opened Jan 2, 2026 by pbelevich Loading…
2
feat(examples/test_run): use runtime sm arch
#2916 opened Dec 31, 2025 by tpoisonooo Loading…
4
[Doc]Fix Mode Name and Stride in 0t_mma_atom.md
#2910 opened Dec 27, 2025 by HydraQYH Loading…
3
docs: note when DSL dumps are populated
#2891 opened Dec 20, 2025 by ColinPeppler Loading…
4
Fix finding cuDNN
#2890 opened Dec 19, 2025 by TLescoatTFX Loading…
1
3
docs: Add FP16 GEMM documentation to sgemm_sm80.cu - Fixes #1686
#2870 opened Dec 10, 2025 by blueberrycongee Loading…
1
1
[WIP]Unit tests for Kernels that perform BF16 x BF16 = MXFP8 and MXFP8 x MXFP8 = BF16
#2857 opened Dec 8, 2025 by Shreya-gaur Loading…
Remove redundant "from" from comment inactive-30d
#2853 opened Dec 8, 2025 by crcrpar Loading…
1
add SM75_16x8x8_F16F16F16F16_TN inactive-30d
#2851 opened Dec 6, 2025 by jinzhen-lin Loading…
1
1
use cp.async.bulk for per-row data; quiets synccheck inactive-30d
#2850 opened Dec 5, 2025 by v0i0 Loading…
2
Remove deprecated newshape argument. inactive-30d
#2844 opened Dec 4, 2025 by Artem-B Loading…
1
Fix TMEM address read/write race in example 77 inactive-30d
#2835 opened Dec 3, 2025 by milesvant • Draft
4
[FIX] Update nvidia-cutlass-dsl requirements version from 4.3.0 to 4.3.1 inactive-30d
#2823 opened Nov 29, 2025 by jeromeku Loading…
1
[CuTeDSL] Feature/fp8e4m3 to fp16 conversion inactive-30d
#2822 opened Nov 28, 2025 by arseniivanov Loading…
2
Fix processing of relative imports in AST preprocessing
#2821 opened Nov 28, 2025 by danieldk Loading…
4
ProTip! Filter pull requests by the default branch with base:main.