Skip to content

[bugfix] [test] update WCPB prelude frontier in split-prefill graph test #154

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
bjf-frz merged 1 commit into from
Oct 27, 2026
Merged

[bugfix] [test] update WCPB prelude frontier in split-prefill graph test #154

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
6 changes: 5 additions & 1 deletion tests/test_split_prefill.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1389,7 +1389,11 @@ def run(cuda_graphs):
pytest.fail("decode request did not reach its second recurrent step")
assert split_request.phase is RequestPhase.PREFILLING_RECURRENT
assert split_request.num_prefilled_tokens == 0
assert split_request.num_preluded_tokens == 2
# WCPB stages prompt chunks in a prelude wavefront: with
# total_ut_steps=4, prefill_chunk_size=2 and a 4-token budget, both
# of this prompt's chunks are staged together. num_prefilled_tokens
# stays 0 because the final recurrent depth has not completed.
assert split_request.num_preluded_tokens == 4

graph_counts = None
if e.model_runner.graphs is not None:
Expand Down
Loading