diff options
author | hclam@chromium.org <hclam@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98> | 2013-07-26 01:19:53 +0000 |
---|---|---|
committer | hclam@chromium.org <hclam@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98> | 2013-07-26 01:19:53 +0000 |
commit | 0d2ab2595f8c26da2d293f52bd5ec4a148c317ac (patch) | |
tree | 39bec7303a2bf2e39ae7fed2c513b2acff4969ad /tools | |
parent | 783069cd120ccd06ccc528e07ca92eca2b8b566c (diff) | |
download | chromium_src-0d2ab2595f8c26da2d293f52bd5ec4a148c317ac.zip chromium_src-0d2ab2595f8c26da2d293f52bd5ec4a148c317ac.tar.gz chromium_src-0d2ab2595f8c26da2d293f52bd5ec4a148c317ac.tar.bz2 |
Make image_decoding_measurement more accurate
The test sequence is now changed to:
1. Load an image but not visible. Wait until fully loaded.
2. Make the image visible. Wait for rAF.
3. Make the invisible and go to (1).
Waiting for the image to be fully loaded ensures there is no progressive
decode. Timing of the decoding operation is better controlled by the
visibility of the image.
This change also makes the test runs faster by using only 5 iterations
and less warm up time.
BUG=260052
R=nduca@chromium.org
Review URL: https://codereview.chromium.org/20113003
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@213738 0039d316-1c4b-4281-b951-d872f2087c98
Diffstat (limited to 'tools')
-rw-r--r-- | tools/perf/measurements/image_decoding.py | 2 |
1 files changed, 2 insertions, 0 deletions
diff --git a/tools/perf/measurements/image_decoding.py b/tools/perf/measurements/image_decoding.py index 1baf345..3f67ae4 100644 --- a/tools/perf/measurements/image_decoding.py +++ b/tools/perf/measurements/image_decoding.py @@ -49,3 +49,5 @@ class ImageDecoding(page_measurement.PageMeasurement): return image_decoding_avg = sum(durations) / len(durations) results.Add('ImageDecoding_avg', 'ms', image_decoding_avg) + results.Add('ImageLoading_avg', 'ms', + tab.EvaluateJavaScript('averageLoadingTimeMs()')) |