-
Notifications
You must be signed in to change notification settings - Fork 2k
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
CPU flamegraphs using perf show wrong sample counts #326
Comments
IMHO we should bring back the old behavior and move the "weighting" behavior behind a command line option.
I understand that there's demand to have the samples weighted by the amount the cycle counter increased between the last sample and this sample. So my proposal would is:
What do you think? I'll be happy to create the PR. |
Issue
When using the instruction from https://www.brendangregg.com/FlameGraphs/cpuflamegraphs.html the number of samples according to perf and the number of samples according to the flame graph differ greatly.
Example numbers:
The flame graph reports billions of samples:
Notes
The example flamegraph on https://www.brendangregg.com/FlameGraphs/cpuflamegraphs.html does report correct numbers.
First Analysis
This change has been introduced with #165 in PR #250
From there on, samples are weighted by a factor but still reported as "samples", making the output of perf and flamegraph contradictory.
The text was updated successfully, but these errors were encountered: