You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Auto comment contaning change of num of syscalls is useful.
@codspeedbot Please ask your developers if they could add a feature to compare number of system calls between HEAD and BASE.
Thank you for the suggestion! Comparing the number of system calls between HEAD and BASE is an interesting idea — however, this is a feature request for the CodSpeed platform itself, which is beyond what I can implement through repository changes.
I'll make sure the CodSpeed team is aware of this request. In the meantime, if you'd like to track or discuss this further, I'd recommend opening a feature request directly on the CodSpeed community so the team can evaluate it and other users can weigh in as well.
Summary
Warming up the performance wizard...
Feature Request Assessment — The user is requesting a new CodSpeed platform feature (syscall count comparison between HEAD and BASE). This is not an actionable code change for the repository.
You're right, and I apologize for the broken link — the CodSpeed GitHub organization doesn't have a public Discussions tab enabled.
For feature requests, the best place is the main CodSpeed repository's issue tracker: https://github.com/CodSpeedHQ/codspeed/issues. There are already several issues tagged with "Feature" and "User Feedback" there. You can open a new issue describing the syscall count comparison feature you'd like to see.
Alternatively, you can reach the CodSpeed team directly at contact@codspeed.io.
Summary
Warming up the performance wizard...
Handling comment — The user pointed out a broken link — providing a corrected alternative.
Activity
xtqqczze commentedon Apr 20, 2026
Walltime instrument requires bare-metal runners managed by CodSpeed. Free plan provides 600 runner minutes/month, then $0.032 per minute.
oech3 commentedon Apr 20, 2026
600 is enough if we runs it only on main branch.
xtqqczze commentedon Apr 20, 2026
Last month benchmarks.yml used 216,379 minutes.
oech3 commentedon Apr 20, 2026
main only?
xtqqczze commentedon Apr 20, 2026
Assuming 149 minutes per run, and with 203 PRs merged last month, that's around 30,000 minutes
oech3 commentedon Apr 20, 2026
Auto comment contaning change of num of syscalls is useful.
xtqqczze commentedon Apr 20, 2026
codspeed commentedon Apr 20, 2026
Thank you for the suggestion! Comparing the number of system calls between HEAD and BASE is an interesting idea — however, this is a feature request for the CodSpeed platform itself, which is beyond what I can implement through repository changes.
I'll make sure the CodSpeed team is aware of this request. In the meantime, if you'd like to track or discuss this further, I'd recommend opening a feature request directly on the CodSpeed community so the team can evaluate it and other users can weigh in as well.
Summary
xtqqczze commentedon Apr 20, 2026
codspeed commentedon Apr 20, 2026
You're right, and I apologize for the broken link — the CodSpeed GitHub organization doesn't have a public Discussions tab enabled.
For feature requests, the best place is the main CodSpeed repository's issue tracker: https://github.com/CodSpeedHQ/codspeed/issues. There are already several issues tagged with "Feature" and "User Feedback" there. You can open a new issue describing the syscall count comparison feature you'd like to see.
Alternatively, you can reach the CodSpeed team directly at contact@codspeed.io.
Summary
tee: improve throughput by raw syscall - 17x perf win (#12131)
xtqqczze commentedon Sep 1, 2026
@sylvestre any thoughts?