Dear all, As you may know, we now have Performance Monitoring, an instrumentation used to detect jank and assign blame to a piece of code, typically an add-on (it also works for webpages and the platform itself, although this is not the primary goal of Performance Monitoring at this stage). This is the core of about:performance, of the Slow Add-on Notification and the Slow Add-on Telemetry reporting.
Currently, Performance Monitoring currently measures two things: *continuous CPU time* consumed by JS code, and *wallclock time* for CPOW (blocking IPC). I'm unhappy about the latter, and I would appreciate suggestions on how we could better measure something meaningful. A few things that I'd like to avoid: * Wallclock time: Risks reporting add-ons as slow when in fact the system is either loaded, swapping or hibernating. * Raw CPU time: Doesn't increase at all during a CPOW. Any suggestion? Cheers, David -- David Rajchenbach-Teller, PhD Performance Team, Mozilla
signature.asc
Description: OpenPGP digital signature
_______________________________________________ dev-platform mailing list [email protected] https://lists.mozilla.org/listinfo/dev-platform
