Skip to content

Actions: flashinfer-ai/flashinfer

Actions

Claude Code Review

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
claude-code-review.yml
2,500+ workflow runs
2,500+ workflow runs

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
improve gdn mtp bf16 state perf for BS<=8 with LDG.128 Claude Code Review #2783: Pull request #3143 opened by ameynaik-hub
11s ameynaik-hub:ameyn/wide_vec_tile_v
ameynaik-hub:ameyn/wide_vec_tile_v
11s
feat: add get flashinfer-trace interface .fi_trace Claude Code Review #2782: Pull request #2931 synchronize by yyihuang
8s yyihuang:fi_trace
yyihuang:fi_trace
8s
[feat] Trtllm-gen Per-token Nvfp4 MoE Claude Code Review #2781: Pull request #3027 synchronize by IwakuraRein
10s IwakuraRein:per-token-fp4
IwakuraRein:per-token-fp4
10s
[codex] add CLI helper for flashinfer-jit-cache wheel installs Claude Code Review #2780: Pull request #3142 opened by dierksen
9s dierksen:codex/install-jit-cache-wheel-cli
dierksen:codex/install-jit-cache-wheel-cli
9s
Add TGV NVFP4 GEMM tactic to mm_fp4 cute-dsl backend (SM100/SM103) Claude Code Review #2779: Pull request #3141 synchronize by Sinestro38
11s Sinestro38:tgv-nvfp4-mm-fp4-backend
Sinestro38:tgv-nvfp4-mm-fp4-backend
11s
Add TGV NVFP4 GEMM tactic to mm_fp4 cute-dsl backend (SM100/SM103) Claude Code Review #2778: Pull request #3141 synchronize by Sinestro38
Action required Sinestro38:tgv-nvfp4-mm-fp4-backend
Sinestro38:tgv-nvfp4-mm-fp4-backend
Action required
Add TGV NVFP4 GEMM tactic to mm_fp4 cute-dsl backend (SM100/SM103) Claude Code Review #2777: Pull request #3141 opened by Sinestro38
Action required Sinestro38:tgv-nvfp4-mm-fp4-backend
Sinestro38:tgv-nvfp4-mm-fp4-backend
Action required
fix(gemm): skip FP4 cuDNN override-shape path on SM120/SM121 (NaN regression from #2910) Claude Code Review #2776: Pull request #3140 opened by Kh4L
Action required Kh4L:fix/sm120-121-fp4-cudnn-override-shape-guard
Kh4L:fix/sm120-121-fp4-cudnn-override-shape-guard
Action required
unifying all reduce memory allocation for single-node and multi-node nvlink Claude Code Review #2775: Pull request #2955 synchronize by Amir-19
8s Amir-19:nvl_mem_uni
Amir-19:nvl_mem_uni
8s
Fix 10 bugs in BF16 XQA MLA kernel for SM120/SM121 Claude Code Review #2774: Pull request #2689 synchronize by blake-snc
8s blake-snc:feat-sm120-bf16-mla
blake-snc:feat-sm120-bf16-mla
8s
[CuTe DSL] Fix FP8 MLA persistent perf regression and ProxyKind cu13 wheel breakage Claude Code Review #2773: Pull request #3132 synchronize by saltyminty
9s pgera:pgera/fp8-mma-unroll-fix
pgera:pgera/fp8-mma-unroll-fix
9s
feat: add get flashinfer-trace interface .fi_trace Claude Code Review #2772: Pull request #2931 synchronize by yyihuang
9s yyihuang:fi_trace
yyihuang:fi_trace
9s
feat: add get flashinfer-trace interface .fi_trace Claude Code Review #2771: Pull request #2931 synchronize by yyihuang
8s yyihuang:fi_trace
yyihuang:fi_trace
8s
[Perf] Add FMHAv2 to flashinfer_benchmark.py and eliminate unnecessary H2D Claude Code Review #2770: Pull request #2841 synchronize by jimmyzho
7s jimmyzho:fmhav2-bench
jimmyzho:fmhav2-bench
7s
feat: add get flashinfer-trace interface .fi_trace Claude Code Review #2769: Pull request #2931 synchronize by yyihuang
10s yyihuang:fi_trace
yyihuang:fi_trace
10s
feat(MoE): FP8 MoE per-channel quant support (note: need cubin!) Claude Code Review #2768: Pull request #2809 synchronize by raayandhar
11s raayandhar:users/rdhar/moe-per-channel-quant
raayandhar:users/rdhar/moe-per-channel-quant
11s
perf(gdn): fix bf16_state T=1 per-call overhead and add pool+padding … Claude Code Review #2767: Pull request #3118 synchronize by ameynaik-hub
11s ameynaik-hub:ameyn/bf16_baseline_fix
ameynaik-hub:ameyn/bf16_baseline_fix
11s
feat: add get flashinfer-trace interface .fi_trace Claude Code Review #2766: Pull request #2931 synchronize by yyihuang
9s yyihuang:fi_trace
yyihuang:fi_trace
9s
feat(moe-a2a): Update nvlink onesided all-to-all Claude Code Review #2765: Pull request #3139 synchronize by trevor-m
8s trevor-m:feat/b802a3818-float32-combine-accumulators
trevor-m:feat/b802a3818-float32-combine-accumulators
8s
feat: add get flashinfer-trace interface .fi_trace Claude Code Review #2764: Pull request #2931 synchronize by yyihuang
8s yyihuang:fi_trace
yyihuang:fi_trace
8s
feat: add get flashinfer-trace interface .fi_trace Claude Code Review #2763: Pull request #2931 synchronize by yyihuang
10s yyihuang:fi_trace
yyihuang:fi_trace
10s
feat(moe-a2a): Update nvlink onesided all-to-all Claude Code Review #2762: Pull request #3139 opened by trevor-m
8s trevor-m:feat/b802a3818-float32-combine-accumulators
trevor-m:feat/b802a3818-float32-combine-accumulators
8s
feat: implement configurable tie_break for filtered topk Claude Code Review #2761: Pull request #3095 synchronize by ziang-and
10s zianglih:topk-tie
zianglih:topk-tie
10s
feat: add delimiterless multi-item scoring (MaskMode::kMultiItemScoringV2) Claude Code Review #2760: Pull request #2921 synchronize by chanh
Action required chanh:cnguyen/delimiterless-mis-v2
chanh:cnguyen/delimiterless-mis-v2
Action required
feat: Enable FP8 (E4M3/E5M2) in concat_mla_k for optimize long-context prefill performance and refactor type dispatch for BF16/FP16 Claude Code Review #2759: Pull request #3129 synchronize by qiching
7s qiching:fix/concat-mla-k-fp8-support
qiching:fix/concat-mla-k-fp8-support
7s