[EP] Add cuda kernel for moe_ep_post_reorder (#6837)

Co-authored-by: luoyuan.luo <luoyuan.luo@antgroup.com>
2025-06-05 15:33:47 +08:00
parent 0166403c20
commit 43baba649e
7 changed files with 377 additions and 4 deletions
--- a/sgl-kernel/python/sgl_kernel/init.py
+++ b/sgl-kernel/python/sgl_kernel/init.py
@@ -49,6 +49,7 @@ from sgl_kernel.gemm import (
 from sgl_kernel.grammar import apply_token_bitmask_inplace_cuda
 from sgl_kernel.moe import (
    cutlass_fp4_group_mm,
+    ep_moe_post_reorder,
    ep_moe_pre_reorder,
    fp8_blockwise_scaled_grouped_mm,
    moe_align_block_size,