commit | 767779f538ad64c3e5012c24e58e5e3e8e802b75 | [log] [tgz] |
---|---|---|
author | Rafal Slawik <rslawik@google.com> | Wed Apr 29 21:05:02 2020 +0100 |
committer | Rafal Slawik <rslawik@google.com> | Wed Apr 29 21:05:02 2020 +0100 |
tree | 57790f6d22fa993a2c92cd3ed85948a28d60740b | |
parent | 399c0c13aea991caf8d8bf6776327672506e7709 [diff] |
Optimise read time_in_state Replace std::map that was used as cache for time in state. Triming the map by selecting a random element in a loop was too slow. Instead keep a fixed size vector and evict cache entries on hash collision. Introduce additional check on ctime + stime from /proc/pid/stat before reading time_in_state for all its threads. If cpu time does not change, there is no need to check time_in_state. With this we can skip about 2/3 of processes. With these optimisations, typical read should go from ~300 ms to ~100 ms. Except the very first one where all threads are scanned. Bug: 155038236 Bug: 149734495 Test: perfetto_unittests Test: take a trace with metatrace enabled and compare timings Change-Id: Ia607d91702465992422929f7b43196b538fda4cd
Perfetto is an open-source project for performance instrumentation and tracing of Linux/Android/Chrome platforms and user-space apps.
See www.perfetto.dev for docs.
See /docs/contributing.md for instructions.
The source-of-truth repo is Android's Gerrit. The GitHub repo is a read-only mirror.
You can reach us on our Discord channel. If you prefer using IRC we have an experimental Discord <> IRC bridge synced with #perfetto-dev
on Freenode.