Skip to content
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.

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

eliminate wall-clock time in favor of counters in core-sharded analyzers using the scheduler #6971

Open
derekbruening opened this issue Sep 9, 2024 · 0 comments
Assignees

Comments

@derekbruening
Copy link
Contributor

Today the memtrace scheduler uses wall-clock time when there is no simulator tracking time, which happens with analyzers. This leads to extra non-determinism and complex cases like #6966 and makes analyzers look a little bit different from simulators. I originally tried a counter but had trouble due to the idle cases incrementing it much more quickly; the plan is to revisit with a weighted counter.

@derekbruening derekbruening self-assigned this Sep 9, 2024
derekbruening added a commit that referenced this issue Sep 9, 2024
Only call gettimeofday once per next_record for instruction quanta,
rather than every get_output_time() call.

Future work will try to revisit using a counter to eliminate the
non-determinism of wall-clock time and make analyzers behavior more
like simulators, easing testing.

Issue: #6971
derekbruening added a commit that referenced this issue Sep 10, 2024
Only call gettimeofday once per next_record for instruction quanta,
rather than every get_output_time() call.

Future work will try to revisit using a counter to eliminate the
non-determinism of wall-clock time and make analyzers behavior more like
simulators, easing testing.

Issue: #6971
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

1 participant