chromium/src
0
Fork 0
Code Activity

Add examples to microbenchamrks documentation to provide more context

Browse Source 
Change-Id: Id51ed68e6d6c3cd4ffd60bc4257d25d9d82a4f8e
Reviewed-on: https://chromium-review.googlesource.com/c/1450035
Commit-Queue: Majid Valipour <majidvp@chromium.org>
Reviewed-by: Fergal Daly <fergal@chromium.org>
Auto-Submit: Majid Valipour <majidvp@chromium.org>
Cr-Commit-Position: refs/heads/master@{#628822}
This commit is contained in:
Majid Valipour
2019-02-04 19:17:06 +00:00
committed by Commit Bot
parent 8dbb714587
commit f9a0698211
1 changed files with 17 additions and 0 deletions

17
docs/speed/microbenchmark_regressions.md

@ -2,8 +2,25 @@
[TOC]
## Microbenchmarks
A microbenchmark attempts to measure the performance of a "small" bit of code.
These tests are typically in the sub-millisecond range. The code being tested
usually performs no I/O, or else is a test of some single, specific I/O task.
[1]
We maintain and measure a set of such microbenchmarks for example
[blink_perf](./benchmark/harnesses/blink_perf.md), dromaeo, gpu_perftests, and
rasterize_and_record_micro.
[1] https://github.com/google/caliper/wiki/JavaMicrobenchmarks#what-is-a-microbenchmark
## Interesting sources of false positives
Due to their nature microbenchmarks are much more sensitive to changes in the
underlying abstraction levels which can result in a large number of false
positive alerts. Below are some of such changes relevant to Chromium.
### AFDO rolls.
Automatic Feedback Directed Optimization is a process that produces