Issue metadata
Sign in to add a comment
|
VR text rendering regression in range 538070 - 538120 |
||||||||||||||||||||||||
Issue descriptionRegression: 0.951 ms (4.1%) Metric: TextPerfTest.render_time_avg Story: render_lorem_ipsum_700_chars_no_test_change Benchmark: xr.vr.common_perftests Affected graphs: https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICQ_Mnp5QoM The culprit CL is the Skia roll https://chromium.googlesource.com/chromium/src/+/f2fa8cfcdab5f221f451cc7d8a75103c23b51143, which only contains one CL (https://skia.googlesource.com/skia.git/+/aefcccb5d84fb0e0cb8161d4021e8fbd724e2d9e). It looks like some of the other text rendering tests also regressed slightly, but not dramatically enough to trigger an alert.
,
Mar 2 2018
You should be able to just build and run the vr_common_perftests target. The gn args used by the bot are: is_chrome_branded = true is_component_build = false is_debug = false is_official_build = true proprietary_codecs = true strip_absolute_paths_from_debug_symbols = true symbol_level = 1 target_cpu = "arm" target_os = "android" use_goma = true
,
Mar 9 2018
I checked out my CL (f2fa8cfcdab5f221f451cc7d8a75103c23b51143) and compared it with its parent (0ed911addbe5d8988b6de0907becdaaf0ba0cf73). The vr_common_perftests target doesn't need to be recomipled between those two so I don't think that my CL is causing a performance regression.
,
Jun 13 2018
Closing this since the affected metrics seem to have returned to normal by now.
,
Jul 4
|
|||||||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||||||
Comment 1 by liyuqian@google.com
, Mar 2 2018