Skip to content

fix(flux): start each MLPerf run with an empty result file #1071

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

Closed
gphuang wants to merge 2 commits into from
Closed

fix(flux): start each MLPerf run with an empty result file #1071

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
78f283e
fix(flux): start each MLPerf run with an empty result file
Sep 1, 2026
f0f240f
fix(flux): create the result file's directory before truncating it
Sep 1, 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
6 changes: 6 additions & 0 deletions examples/mlperf/flux1/megatron/run_and_time.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -51,6 +51,12 @@ export RESULTS_DIR RUN_INDEX PRIMUS_SEED MLLOG_OUTPUT_FILE

mkdir -p "${RESULTS_DIR}"

# The MLLOG logger opens this append-only, so a relaunch under the same
# RUN_INDEX concatenates two runs into one file and the compliance checker
# reports every once-per-run key twice. One file is one run.
mkdir -p "$(dirname "${MLLOG_OUTPUT_FILE}")"
: > "${MLLOG_OUTPUT_FILE}"

# --- Cold start -------------------------------------------------------------
# cache_clear is a claim about the machine, so drop the caches here and report
# what actually happened rather than what was requested. Dropping caches needs
Expand Down