Issue metadata
Sign in to add a comment
|
44.5%-69% regression in blink_perf.bindings at 378997:379010 |
||||||||||||||||||||
Issue descriptionSee the link to graphs below.
,
Mar 4 2016
=== Auto-CCing suspected CL author littledan@chromium.org === Hi littledan@chromium.org, the bisect results pointed to your CL below as possibly causing a regression. Please have a look at this info and see whether your CL be related. ===== BISECT JOB RESULTS ===== Status: completed ===== SUSPECTED CL(s) ===== Subject : Implement TypedArray(typedarray) constructor Author : littledan Commit description: The ES2016 draft spec defines a sort of fast path for constructing a TypedArray based on another TypedArray. This patch implements that alternative path in TypedArray construction. It is verified by test262 tests, which now pass. This patch also has a slight cleanup of TypedArray code by using a macro for TypedArray type checks, as is done for other types. This patch includes a minor spec violation: In the same-type case, the spec indicates that the underlying ArrayBuffer should be copied until the end, and this is fixed up by making the [[ArrayLength]] shorter. This is observable with the buffer getter. This patch just copies the used part of the underlying ArrayBuffer. R=adamk BUG= v8:4726 LOG=Y Review URL: https://codereview.chromium.org/1754593003 Cr-Commit-Position: refs/heads/master@{#34443} Commit : 2fa1c88442d83865f611efb890642d1eb643f320 Date : Wed Mar 02 18:06:29 2016 ===== TESTED REVISIONS ===== Revision Mean Value Std. Dev. Num Values Good? chromium@379003 311.303656 3.027309 5 good chromium@379003,v8@b7a4351404310.499517 3.42541 5 good chromium@379003,v8@56eca6d315311.788766 1.622158 5 good chromium@379003,v8@2d090b11d0310.489075 2.586841 5 good chromium@379003,v8@2fa1c88442122.001943 0.392664 5 bad chromium@379003,v8@017375f328121.436855 0.710167 5 bad chromium@379004 136.510823 30.685762 5 bad Bisect job ran on: android_nexus5_perf_bisect Bug ID: 592007 Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --also-run-disabled-tests blink_perf.bindings Test Metric: typed-array-construct-from-same-type/typed-array-construct-from-same-type Relative Change: 56.15% Score: 99.9 Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_nexus5_perf_bisect/builds/3469 Job details: https://chromeperf.appspot.com/buildbucket_job_status/9019106562529293920 | O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq | X | for more information addressing perf regression bugs. For feedback, | / \ | file a bug with label Cr-Tests-AutoBisect. Thank you!
,
Mar 4 2016
Definitely looks like my patch caused the issue--the tests that got slower are getting at exactly what the patch changed. I'm working on reproducing locally.
,
Mar 8 2016
The following revision refers to this bug: https://chromium.googlesource.com/v8/v8.git/+/ab6e48de4815a03cb6523fbce209fd6fb6df9696 commit ab6e48de4815a03cb6523fbce209fd6fb6df9696 Author: littledan <littledan@chromium.org> Date: Tue Mar 08 20:08:24 2016 Optimize new TypedArray(typedArray) constructor A previous spec compliance fix for TypedArrays caused a ~4x performance regression. This patch removes the regression by calling out to a path within the runtime which implements array copying more efficiently. BUG= chromium:592007 R=adamk LOG=Y Review URL: https://codereview.chromium.org/1767893002 Cr-Commit-Position: refs/heads/master@{#34601} [modify] https://crrev.com/ab6e48de4815a03cb6523fbce209fd6fb6df9696/src/js/typedarray.js
,
Mar 10 2016
|
|||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||
Comment 1 by lanwei@google.com
, Mar 4 2016