Skip to content

[FR_attribution] Fix issues with stale entries #210

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

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
Open

[FR_attribution] Fix issues with stale entries #210

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
186865c
Fix issues with stale entries
sbak5 Oct 27, 2025
d341c41
Merge branch 'main' into sbak/fr_fix
sbak5 Nov 5, 2025
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
12 changes: 10 additions & 2 deletions src/nvidia_resiliency_ext/attribution/trace_analyzer/fr_attribution.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -446,8 +446,9 @@ def get_correct_seq_id(collective):
if c.state != 'scheduled':
continue
rank_counts['appeared'].append(c.file_id)
if get_correct_seq_id(c) <= max_completed_collective_seq_id:
rank_counts['mismatched'].append(c.file_id)
if self.args.use_stale_entries_as_failures:
if get_correct_seq_id(c) <= max_completed_collective_seq_id:
rank_counts['mismatched'].append(c.file_id)
appeared_rank_counts = Counter(rank_counts['appeared'])
# Ranks with less number of enqueued collectives than max_enqueued_collective_seq_id -> host not making expected progress
for rank_id in self.pg_configs[process_group]['ranks']:
Expand Down Expand Up @@ -975,6 +976,13 @@ def main():
help='Convert the trace file to json file, if the trace is binary, for debugging',
)

parser.add_argument(
'--use-stale-entries-as-failures',
action='store_true',
help='Use stale entries as failures in the analysis,'
'which are older than the last completed collective for the process group',
)

args = parser.parse_args()

analyzer = CollectiveAnalyzer(args)
Expand Down