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

103 Open 823 Closed
103 Open 823 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

[QST]how to use one threadblock process one matrix multiplication? ? - Needs Triage question Question
#1617 opened Jul 5, 2024 by alephchang
[FEA] BFloat16x2 Atomics ? - Needs Triage feature request New feature or request
#1616 opened Jul 4, 2024 by HanGuo97
[QST] about swizzle_layout ? - Needs Triage question Question
#1615 opened Jul 3, 2024 by bluebabyxp
[FEA] Hopper group-gemm of mixed type ? - Needs Triage feature request New feature or request
#1614 opened Jul 3, 2024 by jcao-ai
What is the layout of the Batched Stride GEMM when A C is row major and B is col major？ ? - Needs Triage question Question
#1613 opened Jul 2, 2024 by Tangesion
[BUG] coalesce with Tensor dispatch to coalesce with Shape ? - Needs Triage bug Something isn't working
#1612 opened Jul 2, 2024 by cloudhan
4
[QST] LDSM Copy for int8 ? - Needs Triage question Question
#1611 opened Jul 2, 2024 by carlguo866
11
[QST] Benchmark #15 error. "void cutlass::arch::ldsm not implemented" ? - Needs Triage question Question
#1610 opened Jul 1, 2024 by dhjoo98
1
[QST] Calling sgemm kernel in cpp utilizing installed cutlass library ? - Needs Triage question Question
#1609 opened Jul 1, 2024 by dscotthunter
1
cudaMemcpy result = misaligned address ? - Needs Triage bug Something isn't working
#1608 opened Jun 30, 2024 by Tangesion
[QST] Last 10 Slides of Cutlass talk ? - Needs Triage question Question
#1607 opened Jun 30, 2024 by ericauld
[QST] ? - Needs Triage question Question
#1605 opened Jun 28, 2024 by ganeshcolfax
3
cutlass fails during tensorflow assembly ? - Needs Triage question Question
#1603 opened Jun 27, 2024 by zkbitcoin
5
[QST] How to make Main loop fusion for floats ? - Needs Triage question Question
#1602 opened Jun 27, 2024 by satyabhagavan
4
[QST] can group conv support simt version ? ? - Needs Triage question Question
#1599 opened Jun 26, 2024 by tengdecheng
[QST] does cutlass depthwise conv support 1x1 filter ? - Needs Triage question Question
#1598 opened Jun 25, 2024 by tengdecheng
[QST]41_fused_multi_head_attention on sm89 ? - Needs Triage question Question
#1597 opened Jun 21, 2024 by Sunny-bot1
2
[QST] Available Fusion Options in EVT ? - Needs Triage question Question
#1595 opened Jun 20, 2024 by satyabhagavan
2
[BUG] Convolution examples fail to compile ? - Needs Triage bug Something isn't working
#1591 opened Jun 14, 2024 by ogoidmatos
[QST]What is the difference between TensorOp and WmmaTensorOp ? - Needs Triage question Question
#1590 opened Jun 13, 2024 by sleepwalker2017
2
[QST]error: too few arguments for class template "cutlass::epilogue::collective::DefaultEpilogue" ? - Needs Triage question Question
#1589 opened Jun 13, 2024 by sleepwalker2017
[QST]Question about the cutlass 3.0 API ? - Needs Triage question Question
#1588 opened Jun 13, 2024 by sleepwalker2017
[QST]Why does profiler run so many kernels? ? - Needs Triage question Question
#1587 opened Jun 13, 2024 by sleepwalker2017
2
[QST] CUTLASS support for sparse matrix multiplication for X*W=Y with GPU sparse tensor core ? - Needs Triage question Question
#1585 opened Jun 12, 2024 by YukeWang96
[QST] How to improve skinny matrix perf over Ampere like 3090?
#1582 opened Jun 11, 2024 by leiwen83
ProTip! Find all open issues with in progress development work with linked:pr.