Yongye ZhuandClaude Opus 5.5 7566d83bd3 [Attention][MiniMax-M3] NVFP4 KV cache on the MSA sparse attention path (#59300) ...

Signed-off-by: Yongye Zhu <zyy1102000@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

2026-10-01 14:39:55 -07:00

..

attention

…

core

…

cutlass_extensions

…

gdn

…

kimi_k3

…

mamba

…

moe

[Refactor] Remove dead kernel code (#57621)

2026-09-20 10:39:23 -07:00

quantization

[Perf][Kernel] Vectorized flat abs-max for dynamic per-tensor FP8 quantization (#58194)

2026-09-25 09:19:48 -06:00

activation_kernels.cu

…

async_util.cuh

…

cache_kernels_fused.cu

…

cache_kernels.cu

[Bugfix][SM120][MLA] Support NoPE sparse MLA (GLM-5.3-Flash) on the FlashInfer SM120 backend (#55277)

2026-09-23 02:03:20 -07:00

concat_mla_q.cuh

…

cooperative_topk.cu

…

cooperative_topk.cuh

…

cub_helpers.h

…

cuda_utils_kernels.cu

…

cuda_vec_utils.cuh

[Refactor] Remove dead kernel code (#57621)

2026-09-20 10:39:23 -07:00

cuda_view.cu

…

custom_all_gather_reduce_scatter_ops.cpp

[Perf][Distributed] Add low-SM multimem reduce-scatter for SM100/SM103 (#55072)

2026-09-25 03:27:21 +00:00

custom_all_gather_reduce_scatter.cu

[Perf][Distributed] Add low-SM multimem reduce-scatter for SM100/SM103 (#55072)

2026-09-25 03:27:21 +00:00

custom_all_reduce.cu

…

dispatch_utils.h

…

dsv3_fused_a_gemm.cu

…

fp32_router_gemm_entry.cu

…

fp32_router_gemm.cu

…

fused_deepseek_v4_qnorm_rope_kv_insert_kernel.cu

…

fused_kimi_k3_mla_key_concat_kv_cache_kernel.cu

[Bugfix] Fix SM100 fp8_ds_mla cache scales (#49435)

2026-09-19 17:15:07 -07:00

fused_minimax_m3_qknorm_rope_kv_insert_kernel.cu

[Attention][MiniMax-M3] NVFP4 KV cache on the MSA sparse attention path (#59300)

2026-10-01 14:39:55 -07:00

fused_qknorm_rope_kernel.cu

…

hisparse_kernels.cu

[Bugfix][HiSparse] Resolve MTP verification rows with a union residency kernel (#59235)

2026-09-30 03:01:45 +00:00

launch_bounds_utils.h

…

layernorm_kernels.cu

…

layernorm_quant_kernels.cu

…

minimax_reduce_rms_kernel.cu

…

minimax_reduce_rms_kernel.h

…

ngram_embedding_kernels.cu

…

nvfp4_ds_mla_cache_kernels.cu

…

nvfp4_ds_mla_cache.h

…

nvfp4_kv_cache_kernels.cu

…

ops.h

[Attention][MiniMax-M3] NVFP4 KV cache on the MSA sparse attention path (#59300)

2026-10-01 14:39:55 -07:00

persistent_topk.cuh

…

pos_encoding_kernels.cu

…

sampled_topk.cuh

…

sampler.cu

[Refactor] Remove dead kernel code (#57621)

2026-09-20 10:39:23 -07:00

topk_histogram_4096.cuh

…

topk.cu

…

torch_bindings.cpp

[Attention][MiniMax-M3] NVFP4 KV cache on the MSA sparse attention path (#59300)

2026-10-01 14:39:55 -07:00

torch_utils.h

…

type_convert.cuh

…

vocab_parallel_embedding_kernels.cu

…