Skip to content

[Bugfix][Kernel] fix bias adding in triton kernel implemented fused moe #31676

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

Merged
mgoin merged 3 commits into from
Jan 7, 2026
Merged

[Bugfix][Kernel] fix bias adding in triton kernel implemented fused moe #31676

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
73234d9
fix bias adding for triton implemented fused_moe_kernel
xuebwang-amd Jan 4, 2026
08bb240
a tiny code lint issue
xuebwang-amd Jan 4, 2026
a80b38a
Merge remote-tracking branch 'origin/main' into xuebin_fix_triton_fus…
xuebwang-amd Jan 7, 2026
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
13 changes: 8 additions & 5 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 @@ -518,11 +518,7 @@ def fused_moe_kernel(
# Advance the ptrs to the next K block.
a_ptrs += BLOCK_SIZE_K * stride_ak
b_ptrs += BLOCK_SIZE_K * stride_bk
if HAS_BIAS:
accumulator = accumulator + bias[None, :]
if MUL_ROUTED_WEIGHT:
moe_weight = tl.load(topk_weights_ptr + offs_token, mask=token_mask, other=0)
accumulator = accumulator * moe_weight[:, None]

if use_int8_w8a16:
accumulator = (accumulator * b_scale).to(compute_type)
elif use_fp8_w8a8 or use_int8_w8a8:
Expand All @@ -533,6 +529,13 @@ def fused_moe_kernel(
else:
accumulator = accumulator.to(compute_type)

# Since bias is typically not quantized, it's added after dequantization.
if HAS_BIAS:
accumulator = accumulator + bias[None, :]
if MUL_ROUTED_WEIGHT:
moe_weight = tl.load(topk_weights_ptr + offs_token, mask=token_mask, other=0)
accumulator = accumulator * moe_weight[:, None]

# -----------------------------------------------------------
# Write back the block of the output
offs_cn = pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)
Expand Down