Skip to content

[Bugfix] Fix illegal memory access in FP8 MoE kernel #6382

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
merged 1 commit into from
Jul 12, 2024
Merged

[Bugfix] Fix illegal memory access in FP8 MoE kernel #6382

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 @@ -492,12 +492,14 @@ def fused_experts(hidden_states: torch.Tensor,
if tokens_in_chunk == 0:
break

if tokens_in_chunk < CHUNK_SIZE:
# will only happen in the last chunk
if tokens_in_chunk < CHUNK_SIZE and chunk > 0:
# Adjust the intermediate cache size and config for the last
# chunk. Note that in most cases we only have one chunk
# so the cache size and config are already set correctly and
# do not need to be adjusted.
intermediate_cache1 = intermediate_cache1[:tokens_in_chunk]
intermediate_cache2 = intermediate_cache2[:tokens_in_chunk]
intermediate_cache3 = intermediate_cache3[:tokens_in_chunk]
# reload config to get better performance on the last chunk
config = get_config_func(tokens_in_chunk)

curr_topk_ids = topk_ids[begin_chunk_idx:end_chunk_idx]
Expand Down