We are running a survey to help us improve the experience for all of our members. If you see the survey appear, please take the time to tell us about your experience if you can.
Hi,
How to calculate MCPS using Streamline data?
Regards,
abhishek
Hi abhishek,
Streamline arrives at this figure by tapping into the Performance Monitoring Unit hardware counters of a processor (see here for a functional description of the PMU on a Cortex-A7 as an example): PMU functional description
In other words, it's sampled directly from the hardware.
Thanks,
Joe
Hi Joe,
Thanks for the response. How is self samples related to clock cycles?
No problem. Since Streamline is sample based (you decide how often you want Streamline to sample, up to 1 ms accuracy), it periodically checks which function a program is in and reports back. This allows you to see where in the code a system is spending the most time.
Streamline also periodically samples the PMU to collect a value for clock cycles, so what you see is a representative value.
If you're looking for a more in-depth explanation, I'm sure my colleagues wadecherry or paulbarrick from the Streamline team could help out.
Cheers,
To message another member on the community, you need to be following one another. They'll jump in on this thread though with an explanation potentially.