JieXin Liang 18efb5e8e0 [perf][sgl-kernel] extend cutlass_mla_decode to support num_head < 128 (#6929)

2025-06-08 19:37:34 -07:00

..

cutlass_sm100_mla

[perf][sgl-kernel] extend cutlass_mla_decode to support num_head < 128 (#6929)

2025-06-08 19:37:34 -07:00

cascade.cu

feat: adapt merge_state (#5337)

2025-04-12 21:14:04 -07:00

cutlass_mla_kernel.cu

[perf][sgl-kernel] extend cutlass_mla_decode to support num_head < 128 (#6929)

2025-06-08 19:37:34 -07:00

lightning_attention_decode_kernel.cu

…

merge_attn_states.cu

bugfix: fix merge_state_v2 cuda graph (#5419)

2025-04-15 10:18:47 -07:00

vertical_slash_index.cu

[Feat] QWen-1M context support[1/2]: Update block sparse attention backend utils kernel (#5847)

2025-04-28 11:03:17 -07:00