fix: fix illegal cuda memory access at fused_moe_kernel (#4727)

Co-authored-by: yuethe <yuethe@tencent.com>
2025-04-03 15:07:32 +08:00
parent cccfc10e9c
commit e41549c3d6
1 changed files with 1 additions and 0 deletions
--- a/python/sglang/srt/layers/moe/fused_moe_triton/fused_moe.py
+++ b/python/sglang/srt/layers/moe/fused_moe_triton/fused_moe.py
@@ -152,6 +152,7 @@ def fused_moe_kernel(
        return
    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M)
    offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
+    offs_token = offs_token.to(tl.int64)
    token_mask = offs_token < num_valid_tokens

    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)) % N