Skip to content

Pull requests: flashinfer-ai/flashinfer

New pull request New
191 Open 1,952 Closed
191 Open 1,952 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

Update Docker CI tags to 20260401-2c675fb automated docker
#2936 opened Apr 1, 2026 by flashinfer-bot Loading…
1
fix: vectorize get_shuffle_matrix_a_row_indices to eliminate CPU contention run-ci
#2935 opened Apr 1, 2026 by youkaichao Loading…
1
5
[Do not merge] Debugging nvshmem-cu12 3.6.5 fix op: comm
#2932 opened Apr 1, 2026 by bkryu • Draft
5 tasks
7
feat: add get flashinfer-trace interface .fi_trace op: attention op: gemm
#2931 opened Mar 31, 2026 by yyihuang • Draft
5 tasks
5
feat: SM121 (GB10) tile filtering and autotuner robustness run-ci
#2927 opened Mar 31, 2026 by askliar Loading…
2 tasks
15
doc: add CI triggering guide to CONTRIBUTING.md
#2924 opened Mar 31, 2026 by yongwww Loading…
5 tasks
13
ci: narrow skip scope in test_fmha_v2_prefill.py op: attention run-ci
#2922 opened Mar 31, 2026 by bobboli Loading…
10
feat: add delimiterless multi-item scoring (MaskMode::kMultiItemScoringV2)
#2921 opened Mar 31, 2026 by chanh Loading…
3 tasks
14
docs: document replay command in CLI reference
#2919 opened Mar 31, 2026 by ooooo-create Loading…
3 of 5 tasks
5
trtllmgen moe oss kernels run-ci
#2917 opened Mar 30, 2026 by aleozlx Loading…
5 tasks done
10
fix: Fix autotuner crash on meta-device tensor in trtllm_fp4_block_scale_routed_moe op: moe run-ci
#2916 opened Mar 30, 2026 by bkryu Loading…
5 tasks
1
@bkryu
9
perf: optimize top-k kernel for fp32
#2915 opened Mar 30, 2026 by Archie-wang Loading…
5
feat: Add cuBLASLt backend for mm_bf16 and enable multi-tactic autotuning for FP8/MXFP8 runners op: gemm
#2914 opened Mar 30, 2026 by vadiklyutiy Loading…
28
feat(moe): Add MxInt4 x FP8 grouped GEMM support for trtllm-gen fused MoE
#2912 opened Mar 29, 2026 by StudyingShao Loading…
5 tasks done
6
fix: segfault using packed topk_id/weight as trtllm_bf16_routed_moe input in DeepSeek routing
#2911 opened Mar 29, 2026 by rosenrodt Loading…
2 of 5 tasks
3
Yanqinz/dynamic shape unified api op: gemm run-ci
#2910 opened Mar 29, 2026 by yanqinz2 Loading…
3 of 5 tasks
30
feat(gdn): state checkpointing in chunk_gated_delta_rule
#2908 opened Mar 28, 2026 by feldsherov Loading…
4 of 5 tasks
5
fix(moe): make hidden_states_scale optional in trtllm_fp4_block_scale_moe
#2906 opened Mar 28, 2026 by kuttivicky Loading…
1 of 5 tasks
1
7
feat(gdn): separate input and output pool indices
#2905 opened Mar 28, 2026 by feldsherov Loading…
4 of 5 tasks
7
perf: Optimize CuTe-DSL fp4 and fp8 quantization kernels
#2904 opened Mar 27, 2026 by bkryu Loading…
3 of 5 tasks
@bkryu
51
fix: avoid re-downloading BMM export headers when flashinfer-cubin is installed op: moe run-ci
#2903 opened Mar 27, 2026 by yzh119 Loading…
1 of 2 tasks
@aleozlx
20
feat: add MXFP8 GEMM support for SM120 op: gemm run-ci
#2902 opened Mar 27, 2026 by samuellees Loading…
5 tasks done
1
27
bench: fix GPU power throttling in benchmark utilities
#2899 opened Mar 26, 2026 by Edenzzzz Loading…
4 of 5 tasks
6
fix: snap weight_scale_vec_size to handle block_scale_interleave padding for SM120 op: moe run-ci
#2898 opened Mar 26, 2026 by samuellees Loading…
3 tasks done
1
26
Add Relu2 activation support in CUTLASS MoE backend and fix autotuner async CUDA error handling op: gemm op: moe run-ci
#2897 opened Mar 26, 2026 by askliar Loading…
16
ProTip! Mix and match filters to narrow down what you’re looking for.