Skip to content

Pull requests: triton-lang/triton

New pull request New
88 Open 4,916 Closed
88 Open 4,916 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

[BACKEND] Fix layout picked during TMA store pipelining
#6978 opened May 29, 2025 by ThomasRaoux Loading…
[AMD] Skip InThreadTransposeOp in scheduleGlobalLoadLocalStore
#6976 opened May 29, 2025 by zhanglx13 • Draft
Cast bf16 to fp32 for sum/reduce
#6974 opened May 29, 2025 by dshi7 Loading…
1
[Frontend] Add @tl.aggregate which autogenerates a Triton type
#6970 opened May 28, 2025 by Mogball Loading…
5
[AMD]Fix Matmul Tests for the Benchmark on AMD gfx950
#6965 opened May 28, 2025 by knwng Loading…
[PIPELINER] Reintroduce epilogue peeling
#6962 opened May 28, 2025 by pawelszczerbuk Loading…
Add a default-off option to promote bf16 to fp32 for accumulation in tl.sum
#6961 opened May 28, 2025 by dshi7 Loading…
6
Fix assertion in setOptimizedGatherLayout for 1D tensor
#6959 opened May 28, 2025 by chengjunlu Loading…
2
[AMD] Restrict fp8 dtype usage in dot products at the semantic level
#6955 opened May 28, 2025 by njriasan Loading…
3
Update FetchContent_Populate -> FetchContent_MakeAvailable in googletest.cmake
#6953 opened May 27, 2025 by anmyachev Loading…
[DOCS] Update Triton version in conf.py
#6952 opened May 27, 2025 by anmyachev Loading…
[BACKEND] Fix SoftwarePipeliner dumpIntermediateSteps for MLIR_ENABLE_DUMP=name
#6933 opened May 26, 2025 by plotfi Loading…
5 tasks done
6
Fix: PY_SSIZE_T_CLEAN macro must be defined for '#' formats
#6928 opened May 24, 2025 by AdityaMayukhSom Loading…
5 of 7 tasks
1
6
[WIP] Try a llvm build
#6923 opened May 23, 2025 by antiagainst • Draft
[IR] Remove boundary check and padding option from load and store
#6912 opened May 22, 2025 by peterbell10 Loading…
[BENCH] Use TMA for block scales even when performing W @ X
#6905 opened May 21, 2025 by csullivan Loading…
2
[PROTON-DEV] Add Sched Barrier Pass To Prevent Instruction Reordering Outside Proton Record Regions
#6904 opened May 21, 2025 by CRobeck Loading…
Fix RelWithDebInfo build
#6901 opened May 21, 2025 by ayshuha Loading…
5 of 7 tasks
4
[BENCH] add renormalize knob in routing, so that the kernel can be used for qwen and mixtral moe family
#6896 opened May 21, 2025 by zyongye Loading…
5 of 7 tasks
[DNR] Live buffer analysis
#6874 opened May 19, 2025 by Mogball • Draft
[BC-Breaking] Remove block_ptr in favor of tensor descriptor
#6847 opened May 16, 2025 by peterbell10 Loading…
1
[DRAFT][BENCH] Initialize simple multi-gpu moe baseline
#6842 opened May 16, 2025 by Jokeren Loading…
[RFC] Proof of concept of swizzling the scratch space in reduction.
#6826 opened May 15, 2025 by nhukc Loading…
5 of 7 tasks
3
Fix broken transpose match in OptimizeDotOperands
#6819 opened May 14, 2025 by saagarjha Loading…
5 of 7 tasks
1
9
[AMD] disable pattern rollback
#6793 opened May 12, 2025 by makslevental • Draft
2
ProTip! Adding no:label will show everything without a label.