Skip to content

Issues: NVIDIA/cutlass

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

179 Open 982 Closed
179 Open 982 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
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

[BUG] seq_v unusued, but set causing complaints in upstream PyTorch compilation ? - Needs Triage bug Something isn't working
#1971 opened Nov 29, 2024 by Skylion007
[QST]Why can't n=24 or n=40 run the tma_warpSpecialized (including: pingpong/cooperative/native warpspecialized) scheduler? ? - Needs Triage question Question
#1970 opened Nov 29, 2024 by chenzhanyiczy2
[QST] understanding GroupedProblemVisitor.next_tile( ) in Grouped Gemm or Batched Gemm ? - Needs Triage question Question
#1969 opened Nov 29, 2024 by danielhua23
[BUG] ? - Needs Triage bug Something isn't working
#1967 opened Nov 27, 2024 by rasai
1
[QST] CUDA free failed when executing example 59_ampere_gather_scatter_conv ? - Needs Triage question Question
#1965 opened Nov 27, 2024 by IzanCatalan
2
[QST] int8 Conv2D for volta V100 ? - Needs Triage question Question
#1964 opened Nov 26, 2024 by sycz00
[QST] Row Contiguous C Matrix from TiledMMA? ? - Needs Triage question Question
#1963 opened Nov 26, 2024 by osayamenja
how does the threadblock_tile_offset read the global memory in gemm_splitk_parallel.h ? ? - Needs Triage question Question
#1957 opened Nov 21, 2024 by pily1
2
[QST] make_tiled_copy_B generates incompatible layouts ? - Needs Triage question Question
#1953 opened Nov 20, 2024 by phantaurus
9
[BUG] Example 09_turing_tensorop_conv2dfprop does not work ? - Needs Triage bug Something isn't working
#1952 opened Nov 19, 2024 by IzanCatalan
3
[QST] Modify how to load Activations and Filters ? - Needs Triage question Question
#1950 opened Nov 18, 2024 by IzanCatalan
[BUG] Wrong assertion in integer_subbyte.h ? - Needs Triage bug Something isn't working
#1949 opened Nov 18, 2024 by Algy
2
[QST]Question about vectorized memory accesses. ? - Needs Triage question Question
#1946 opened Nov 15, 2024 by leizhao1234
1
[QST] Can hopper_int4_fp8_gemm support Scale with zero-point mode? ? - Needs Triage question Question
#1944 opened Nov 15, 2024 by ZZBoom
[QST] Does TMA overlap memory copy from/to global memory address from another GPU return by cudaIpcGetMemHandle? ? - Needs Triage question Question
#1943 opened Nov 15, 2024 by umiswing
1
[QST] What does "l" in "mnkl" mean in cutlass? ? - Needs Triage question Question
#1939 opened Nov 13, 2024 by umiswing
3
[QST] FP8 with row-wise scaling on Ada-Lovelace ? - Needs Triage question Question
#1937 opened Nov 11, 2024 by vgoklani
6
[QST] bfloat16 x int8 GEMM ? - Needs Triage question Question
#1936 opened Nov 11, 2024 by sycz00
[QST] How to define a new custom kernel ? - Needs Triage question Question
#1930 opened Nov 8, 2024 by IzanCatalan
2
[QST] Why tma_load.get_slice(0) here always need 0? ? - Needs Triage question Question
#1929 opened Nov 8, 2024 by ziyuhuang123
[QST] Does CUTLASS 3.5.1 support int4 x float16 GEMMs natively? ? - Needs Triage question Question
#1928 opened Nov 7, 2024 by SimpleTheoryOfTypes
3
[QST] Question Regarding To The Use Of Swizzle ? - Needs Triage question Question
#1927 opened Nov 7, 2024 by Yanksi
1
[QST] Why did I get a wrong result from GemmGrouped? ? - Needs Triage question Question
#1924 opened Nov 7, 2024 by WangNorthSea
3
[QST] Is there a Cutlass GEMM example to read inputs with custom padding? ? - Needs Triage question Question
#1922 opened Nov 6, 2024 by ghostplant
5
[FEA] Better grid size for H100 GPU with SXM5 ? - Needs Triage feature request New feature or request
#1921 opened Nov 6, 2024 by zhipeng93
7
ProTip! no:milestone will show everything without a milestone.