Closed jasonneverstop closed 1 year ago
Thanks for bringing this up. Some hardware events dont translate perfectly to the pid mode of perf. I will look into whether this is the case or not for cpu utilization and get back to you / update pid mode
@jasonneverstop Thank you for reporting this issue. The metric CPU utilization uses reference cycles to identify the utilization % for a given pid. Since reference cycles does not account for dynamic changes to the clock frequency, the value can sometimes exceed 100% with the current implementation. We will add a fix for this issue in the next release.
@jasonneverstop We now released V1.2.3 with a the fix for this issue: https://github.com/intel/PerfSpect/releases/tag/1.2.3. Please try it out and let us know your feedback.
@jasonneverstop We now released V1.2.3 with a the fix for this issue: https://github.com/intel/PerfSpect/releases/tag/1.2.3. Please try it out and let us know your feedback. Thank you for helping to solve this problem. I will verify this issue in future work. Thank you for your contribution to this project, which has been very helpful to us.
When I use 'perfspect -p' to target only one process,the metric_CPU utilization is greater than 100.I want to know why and how to solve it.