Skip to content

Pull requests: triton-lang/triton

New pull request New
133 Open 6,127 Closed
133 Open 6,127 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

[KERNELS] Associate the correct file and line numbers with specialized kernels
#8387 opened Oct 7, 2025 by Jokeren Loading…
1
[mxfp] support EXPT_IS_INNER in non-persistent matmul
#8385 opened Oct 7, 2025 by jongsoo-openai • Draft
7 tasks
Support L2 cache clearing in do_bench_cudagraph
#8384 opened Oct 6, 2025 by yf225 Loading…
3 tasks done
2
fix: missing dependency TritonIR -> TritonGPUIR's tbl-gen headers
#8383 opened Oct 6, 2025 by SanaaHamel Loading…
5 of 7 tasks
3
[TRITON_KERNELS] added BitmatrixMetadta and RaggedTensorMetadata; deprecated triton_kernels.routing
#8375 opened Oct 5, 2025 by ptillet Loading…
3
[Backend] Optimize membar insertion on hopper
#8374 opened Oct 5, 2025 by lijinpei Loading…
3 tasks
24
[AMD][DRAFT] revamp range analysis
#8372 opened Oct 5, 2025 by yangshuxin • Draft
7 tasks
[AMD][BACKEND] Use PaddedLayout with AsyncCopy on gfx950 when pipelining
#8365 opened Oct 3, 2025 by AlexAUT Loading…
8
Bump actions/setup-python from 5 to 6
#8361 opened Oct 3, 2025 by anmyachev Loading…
[Backend] Bump to llvm/llvm-project@b92ff6b209
#8360 opened Oct 3, 2025 by yiqian1 • Draft
[AMD] Option to use custom scheduler in LLVM backend
#8326 opened Sep 30, 2025 by alefimov-amd • Draft
fix: missing dep NVHopperTransforms -> TritonGPUTransforms
#8318 opened Sep 29, 2025 by SanaaHamel Loading…
5 of 7 tasks
1
[KERNELS] Add NCHW BatchNorm forward (fp32 accum)
#8305 opened Sep 27, 2025 by Aminsed Loading…
1
[AMD] Support float8_e5m2 in tutorials/03-matrix-multiplication.py for gfx950
#8304 opened Sep 26, 2025 by matthiasdiener Loading…
5 of 7 tasks
7
[AMD] Add an option to restrict BufferOps to only use range analysis
#8303 opened Sep 26, 2025 by njriasan • Draft
1
[AMD] Add error message for invalid MFMA tile k-dimension size
#8301 opened Sep 26, 2025 by christopherpriebe Loading…
1 task done
8
[MULTI-GPU] Optimize reduce_scatter (except all-to-all) using custom triton kernels
#8300 opened Sep 26, 2025 by Jokeren Loading…
7
Persist valid rematerialisations from slice creation
#8292 opened Sep 25, 2025 by neildhar Loading…
5 of 7 tasks
7
Skip values with existing conversions in getConvertBackwardSlice
#8291 opened Sep 25, 2025 by neildhar Loading…
5 of 7 tasks
5
[WIP! Do not merge!] Remove optimizeLDSUsage pass and usage of padded layout in general co…
#8282 opened Sep 25, 2025 by plognjen Loading…
mxfp4 hbm gb200 layout
#8274 opened Sep 24, 2025 by aeng-openai • Draft
[BACKEND] Update LLVM version to https://github.com/llvm/llvm-project/commit/b8649098a7fcf598406d8d8b7d68891d1444e9c8
#8270 opened Sep 24, 2025 by olegshyshkov Loading…
[mxfp] swap grid shapes of quant/dequant kernels
#8250 opened Sep 21, 2025 by jongsoo-openai Loading…
7 tasks
1
[BACKEND] Improved swizzling when there is not enough vectorisation
#8240 opened Sep 19, 2025 by lezcano Loading…
6
Add support for fp8e4m3fnuz dtype in Triton
#8231 opened Sep 19, 2025 by arnav-garg1 Loading…
5 of 7 tasks
1
5
ProTip! What’s not been updated in a month: updated:<2025-09-07.