Skip to content

Conversation

@Zqy11
Copy link

@Zqy11 Zqy11 commented Nov 6, 2025

The DeepEP implementation for SBO (DeepEP #390) will be merged into the antgroup-opt branch.

for (int token_idx = offset + sub_warp_id; token_idx < offset + num_tokens_to_send; token_idx += num_warps_per_group) {
if (overlap or (not is_rank_masked<true>(mask_buffer_ptr, dst_rank))) {
auto token_start_idx = overlap ? local_expert_signal_idx * block_m : offset;
auto token_end_idx = overlap ? min((local_expert_signal_idx + 1) * block_m, num_tokens_per_expert) : (offset + num_tokens_to_send);
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hi~ I have tried this great feature, and foud that combine_send was slower than the non-overlap situation, maybe block_m is too big for one SM in each itertions? I

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You may consider increasing num_sms from the default value of 3 to 4-6. When block_m is set to 64, given that num_warps is 32: if num_token <= 32, one SM sends a single round; if 32 < num_token <= 64, one SM needs to send two rounds. Two rounds may take slightly longer than the origin combined send, but you can increase parallelism by raising num_sms.

Copy link
Collaborator

@wangfakang wangfakang left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM. Thanks.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants