Skip to content

Pull requests: meta-pytorch/tritonbench

New pull request New
26 Open 693 Closed
26 Open 693 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

Guard against attribute error cla signed fb-exported meta-exported
#1042 opened Apr 27, 2026 by xuzhao9 Contributor Loading…
1
[tritonbench][AutoWS] Enable AutoWS matmul kernels on Hopper cla signed fb-exported meta-exported
#1025 opened Apr 21, 2026 by njriasan Contributor Loading…
1
[tk] add tk gb300 kernels cla signed
#1019 opened Apr 15, 2026 by xuzhao9 Contributor Loading…
Add mxfp8_blackwell_attentions benchmark to TritonBench cla signed fb-exported meta-exported
#1012 opened Apr 13, 2026 by njriasan Contributor Loading…
1
Security: Unsafe eval() when parsing input loader data (FP8 GEMM)
#1008 opened Apr 11, 2026 by tomaioo Loading…
10
Add attn_vis metrics and widgets (#999) cla signed fb-exported meta-exported
#999 opened Apr 8, 2026 by momochen Contributor Loading…
1
adding support for every-n test sampling cla signed fb-exported meta-exported
#992 opened Apr 2, 2026 by jhou-jpg Contributor Loading…
1
[BE] Remove hstu submodule and install as packages cla signed
#989 opened Apr 2, 2026 by xuzhao9 Contributor Loading…
[Flex Attention] Add a Triton version so we can test autoWS cla signed
#983 opened Mar 30, 2026 by manman-ren Contributor • Draft
Add inductor_flex_attention_bwd operator cla signed fb-exported meta-exported
#940 opened Mar 10, 2026 by OmarPavel Contributor Loading…
1
Add inductor_flex_attention_fwd operator cla signed fb-exported meta-exported
#939 opened Mar 10, 2026 by OmarPavel Contributor Loading…
2
[wip] support multi-mode benchmarking cla signed
#909 opened Mar 3, 2026 by xuzhao9 Contributor Loading…
[TLX][FA] update tritonbench version to support rescale opt cla signed
#884 opened Feb 21, 2026 by manman-ren Contributor Loading…
1
Add Diode Inductor max-autotune benchmarks to gemm, addmm, bmm cla signed fb-exported meta-exported
#758 opened Dec 23, 2025 by jananisriram Contributor Loading…
3
[DO NOT LAND] Test run for MLP bias accuracy issue cla signed
#609 opened Oct 31, 2025 by xuzhao9 Contributor Loading…
3
fix partitionK cla signed
#570 opened Oct 17, 2025 by tianrengao Contributor • Draft
Update gluon_attention_persistent_forward cla signed
#518 opened Oct 6, 2025 by neildhar Contributor • Draft
Add a Blackwell-specific scaled persistent + TMA template for GEMMs cla signed fb-exported meta-exported
#432 opened Sep 17, 2025 by jananisriram Contributor Loading…
2
Add trtlllm to triton bench cla signed fb-exported meta-exported
#379 opened Aug 29, 2025 by Aya-ZIbra Contributor Loading…
5
Add cutlass decode kernel to TritonBench cla signed fb-exported meta-exported
#376 opened Aug 28, 2025 by Aya-ZIbra Contributor Loading…
4
[gdpa] Turn on on-host TMA for TLX fwd cla signed
#365 opened Aug 28, 2025 by htyu Contributor Loading…
2
Validate exhaustive autotuning for FP8 Inductor templates cla signed fb-exported
#355 opened Aug 25, 2025 by jananisriram Contributor Loading…
5
[DO NOT LAND] Try always enabling cuda graph cla signed
#348 opened Aug 21, 2025 by xuzhao9 Contributor Loading…
[RoPE] update input settings cla signed
#345 opened Aug 20, 2025 by BoyuanFeng Contributor • Draft
1
Add amax as default per-row scaling factor for fp8_gemm benchmark cla signed fb-exported
#341 opened Aug 20, 2025 by jananisriram Contributor Loading…
15
ProTip! Adding no:label will show everything without a label.