Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[BugFix] [Kernel] Fix GPU SEGV occuring in fused_moe kernel #10385

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
+5 −3
Open

[BugFix] [Kernel] Fix GPU SEGV occuring in fused_moe kernel #10385

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 5 additions & 3 deletions vllm/model_executor/layers/fused_moe/fused_moe.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -105,16 +105,18 @@ def fused_moe_kernel(
num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
return
offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M)
offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(
tl.int64)
offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
token_mask = offs_token < num_valid_tokens

offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)) % N
offs_bn = (pid_n * BLOCK_SIZE_N +
tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
offs_k = tl.arange(0, BLOCK_SIZE_K)
a_ptrs = a_ptr + (offs_token[:, None] // top_k * stride_am +
offs_k[None, :] * stride_ak)

off_experts = tl.load(expert_ids_ptr + pid_m)
off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
b_ptrs = b_ptr + off_experts * stride_be + (offs_k[:, None] * stride_bk +
offs_bn[None, :] * stride_bn)
if use_int8_w8a16:
Expand Down