diff options
author | Sami Kyostila <skyostil@google.com> | 2015-03-19 12:30:32 +0000 |
---|---|---|
committer | Sami Kyostila <skyostil@google.com> | 2015-03-19 12:32:48 +0000 |
commit | b0a90f4b6c81fc65dd0c57ebc00b0ba6fe784b22 (patch) | |
tree | 78a78483e593363bdab7425adaf54db3ccd1e3a5 /content/browser/frame_host/navigation_controller_impl_browsertest.cc | |
parent | 23ca53245b79c816a5bc0e969a5b3393107fffdd (diff) | |
download | chromium_src-b0a90f4b6c81fc65dd0c57ebc00b0ba6fe784b22.zip chromium_src-b0a90f4b6c81fc65dd0c57ebc00b0ba6fe784b22.tar.gz chromium_src-b0a90f4b6c81fc65dd0c57ebc00b0ba6fe784b22.tar.bz2 |
Add second batch JS benchmark and code generator
This patch adds a "Second batch JS" benchmark and a code generator for
creating synthetic Javascript content for the different variants of the
benchmark.
The benchmark flow is:
1. Load an initial page quickly. We also start a continuous animation
which gauges the responsiveness of the main thread.
2. Start loading Javascript asynchronously.
3. When the loading completes, wait 1 second and click on a button
which causes about 5% of the Javascript to execute.
The aim is to keep the page responsive during loading and minimize the
latency of the click handler. There are three variants with a small (15K),
medium (150K) and large (1.2M) amount of Javascript.
The generated Javascript consists of half top-level closures, half top-level
functions, each of which calls a varying number of inner functions to
perform arithmetic.
Note that currently the benchmark only measures the responsiveness of the
click handler, but eventually it will be able to track this over the entire
interaction flow (crbug.com/460206).
BUG=428350
R=rmcilroy@chromium.org
Review URL: https://codereview.chromium.org/1000203002
Cr-Commit-Position: refs/heads/master@{#321340}
Diffstat (limited to 'content/browser/frame_host/navigation_controller_impl_browsertest.cc')
0 files changed, 0 insertions, 0 deletions