[1/2] Speed up prefill mla attention (#10156)

2025-09-09 00:00:33 +08:00
parent 2c2b19b18b
commit 0096798ed6
6 changed files with 130 additions and 0 deletions
--- a/sgl-kernel/python/sgl_kernel/init.py
+++ b/sgl-kernel/python/sgl_kernel/init.py
@@ -23,6 +23,7 @@ from sgl_kernel.cutlass_moe import cutlass_w4a8_moe_mm, get_cutlass_w4a8_moe_mm_
 from sgl_kernel.elementwise import (
    FusedSetKVBufferArg,
    apply_rope_with_cos_sin_cache_inplace,
+    concat_mla_k,
    copy_to_gpu_no_ce,
    downcast_fp8,
    fused_add_rmsnorm,