Skip to content

Issues: triton-lang/triton

Beta
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

746 Open 998 Closed
746 Open 998 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

Bad performance in InfoNCE loss between triton vs pytorch implementation
#6605 opened Apr 25, 2025 by rangehow
How to customize the CUDA kernel name with num_warps/num_stages
#6588 opened Apr 24, 2025 by houqi
2
tl.dot on transposed matrix tries to rearrange matrix in shared memory performance
#6569 opened Apr 23, 2025 by saagarjha
2
Triton does not detect libcuda.so inside Conda environment bug
#6558 opened Apr 22, 2025 by BrandonStudio
Release notes 🙏
#6557 opened Apr 22, 2025 by jeffra
2
Crash when MLIR_DUMP_PATH is set to a directory instead of a file bug
#6548 opened Apr 21, 2025 by peymanbr
1
Block pointer not compatible with atomic_add
#6539 opened Apr 19, 2025 by Edenzzzz
Internal error (Illegal instruction) bug
#6533 opened Apr 18, 2025 by Toheeb27
Triton matmul_ogs Kernel missing multiplying expert weight bug
#6527 opened Apr 17, 2025 by zyongye
Autotune bug: autotune result not equal to non-autotuned result bug
#6524 opened Apr 17, 2025 by huangrt01
Matmul performance on A100 performance
#6523 opened Apr 17, 2025 by QuqqU
The position of if-else significantly affect performance, which is unexpected. performance
#6491 opened Apr 15, 2025 by wenhaoli-xmu
Pip install failure bug
#6483 opened Apr 14, 2025 by HariprasadManimozhi
7
Unnecessary shared memory layout change
#6446 opened Apr 10, 2025 by ekuznetsov139
Triton BF16 mode is slower compared to FP32 mode. performance
#6427 opened Apr 8, 2025 by wenhaoli-xmu
2
Request: document changelog for 3.0,3.1,3.2 releases..
#6423 opened Apr 8, 2025 by oscarbg
Request: Make/Tag 3.0,3.1,3.2 releases..
#6422 opened Apr 8, 2025 by oscarbg
(tensor, tensor) raises TypeError("'constexpr' object is not subscriptable") when indexed at an inner function
#6414 opened Apr 7, 2025 by TongWang-AI4S
3
[AMD] BufferOps require additional annotation for grid dependent information performance
#6398 opened Apr 4, 2025 by njriasan
1
[FeatureRequest] Support passing triton.jit functions to triton.jit functions as arguments
#6367 opened Apr 2, 2025 by alexdremov
1
How to save the results of autotune？ question
#6355 opened Apr 1, 2025 by helloworldstone
4
Loading from TMA descriptor hangs bug question
#6354 opened Mar 31, 2025 by saagarjha
4
Hard coded .to(tl.floatXX) in 06-fused-attention example bug
#6322 opened Mar 27, 2025 by githubsgi
[AMD] Triton Kernel generates suboptimal s_waitcnt vmcnt when mixing buffer_load and global_load performance
#6310 opened Mar 26, 2025 by njriasan
@zhanglx13
2
Flash Attention 2 is slow when head dim is 128 performance
#6305 opened Mar 26, 2025 by SonicZun
3
ProTip! Type g i on any issue or pull request to go back to the issue listing page.