Skip to content

Pull requests: triton-lang/triton

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
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort

Pull requests list

[CI] Bump actions/cache from v4 to v6
#11206 opened Aug 7, 2026 by antiagainst Member Loading…
Fix byte-level race tracking in GSan
#11205 opened Aug 7, 2026 by roman-openai Collaborator Draft
Remove redundant Triton Python guards
#11204 opened Aug 7, 2026 by roman-openai Collaborator Draft
[ConSan] Preserve cluster reader visibility
#11203 opened Aug 7, 2026 by jeffniu-openai Collaborator Loading…
[ConSan] Track multicast barrier recipients
#11202 opened Aug 7, 2026 by jeffniu-openai Collaborator Loading…
[Gluon] Include bases in TensorMemoryLinearLayout mangling
#11201 opened Aug 7, 2026 by 39ali Loading…
4 of 7 tasks
[ConSan] Distinguish readers from peer observers
#11200 opened Aug 6, 2026 by jeffniu-openai Collaborator Loading…
[GSan] Fix race in gdc_wait handling
#11199 opened Aug 6, 2026 by peterbell10 Contributor Loading…
[gluon] Allow offline FP4 descriptor construction
#11198 opened Aug 6, 2026 by aeng-openai Collaborator Loading…
Add JUnit output to the test runner
#11197 opened Aug 6, 2026 by CliveUnger Contributor Loading…
[AMD] Re-enable true16 feature for gfx1100
#11188 opened Aug 5, 2026 by saeid-rostami Contributor Draft
5 of 7 tasks
[AMD] Align TDM descriptor layouts with allocations
#11184 opened Aug 5, 2026 by agron911 Loading…
4 of 7 tasks
Support independent activation scale TMA modes
#11177 opened Aug 5, 2026 by lyu-oai Collaborator Loading…
[NVIDIA] Fix st.async handling in lowering and ConSan
#11175 opened Aug 4, 2026 by jeffniu-openai Collaborator Loading…
fix(autotuner): release CUDA memory retained by failed configs
#11174 opened Aug 4, 2026 by lequytra Loading…
5 of 7 tasks
[KERNELS] Fix Blackwell NVFP4 x dense matmul tile selection
#11171 opened Aug 4, 2026 by amyyang-openai Collaborator Loading…
[triton-ext] Coalesce .extend_with() into a single top-level function
#11167 opened Aug 4, 2026 by abrown Contributor Loading…
[BACKEND] Restrict broadcast-mul-reduce combine to f32
#11164 opened Aug 4, 2026 by Chennesxu Contributor Loading…
[DNR][BACKEND] Reduce layout conversions at memory boundaries
#11161 opened Aug 4, 2026 by jeffniu-openai Collaborator Draft
1 task
Skip LLVM jump threading for functions with convergent ops
#11156 opened Aug 3, 2026 by 39ali Loading…
5 of 7 tasks
[SM120] Support mixed FP8 and FP4 scaled dot
#11154 opened Aug 3, 2026 by sjoerdmeijer Loading…
5 of 7 tasks
Preserve AxisInfo hints on inlined jit call results
#11150 opened Aug 2, 2026 by 39ali Loading…
5 of 6 tasks
ProTip! Add no:assignee to see everything that’s not assigned.