Skip to content

Pull requests: flashinfer-ai/flashinfer

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

[Draft][SM120] Add W4A8 CuTeDSL MegaMoE backend
#4632 opened Aug 20, 2026 by Euynaheh Loading…
Update Docker CI tags to 20260819-f092274 automated docker
#4629 opened Aug 20, 2026 by flashinfer-bot Collaborator Loading…
Low latency glm5 op: moe
#4627 opened Aug 19, 2026 by yijingl-nvidia Loading…
5 tasks done
fix(cudnn): honor q.dtype and add return_lse in cudnn decode op: attention
#4625 opened Aug 19, 2026 by vedaanta Contributor Loading…
5 tasks done
Enable top_k_varlen on Rubin (SM107)
#4621 opened Aug 19, 2026 by dhiraj113 Collaborator Loading…
5 tasks
feat(moe): add typed unified activation parity op: moe
#4613 opened Aug 19, 2026 by feih-nv Collaborator Draft
1 of 5 tasks
feat(moe): add remaining CUTLASS unified MoE runners op: moe run-ci
#4610 opened Aug 19, 2026 by feih-nv Collaborator Loading…
5 tasks done
test: skip SM120 FMHA v2 tests on SM121 op: attention
#4606 opened Aug 19, 2026 by nvamyt Contributor Loading…
5 tasks
MXFP8 x BF16 integration
#4604 opened Aug 19, 2026 by djns99 Contributor Draft
5 tasks
moe_ep: SM107 (Rubin) block-scaled mega-MoE backends
#4601 opened Aug 19, 2026 by mhoqueanik Collaborator Loading…
Adding var-window FMHA context support for PrimsTS. op: attention
#4599 opened Aug 19, 2026 by mingxu1067 Loading…
5 tasks done
ProTip! Find all pull requests that aren't related to any open issues with -linked:issue.