| commit | adeb79deafcef415740e38d92d9e742b5b7f2233 | [log] [tgz] |
|---|---|---|
| author | Primiano Tucci <primiano@google.com> | Tue Nov 09 18:51:29 2021 +0000 |
| committer | Primiano Tucci <primiano@google.com> | Tue Nov 09 18:51:29 2021 +0000 |
| tree | 3a51b1c586ae6dbf057cad1e4417788d2a28e1ff | |
| parent | 0894513760cfda086b5c2a42aca2d3edf5ae8c9f [diff] |
TraceProcessor: optimize std::sort<TimestampedTracePiece> TraceSorter::Sort() is a very hot path in trace processor. That does a std::sort on a CircularQueue<TTP>. std::sort uses std::swap, which falls back on 2 std::moves when not implemented. TTP happens to be precisely 512-bits wide. This allows to implement a very efficient swap which leverages XMM registers on x86_64. This saves 7-10% of trace ingestion time on a large trace. Bug: 205302474 Change-Id: Ie57fc26e79599b2e27945dc5cef176c1d03d95cc
Perfetto is a production-grade open-source stack for performance instrumentation and trace analysis. It offers services and libraries and for recording system-level and app-level traces, native + java heap profiling, a library for analyzing traces using SQL and a web-based UI to visualize and explore multi-GB traces.
See https://perfetto.dev/docs or the /docs/ directory for documentation.