Issue metadata
Sign in to add a comment
|
12KB regression in resource_sizes (MonochromePublic.apk) at 466172:466172 |
||||||||||||||||||||
Issue descriptionAssigning to agrieve@ since this is a roll and bisect is currently not working.
,
Apr 25 2017
Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8981385680889989232
,
Apr 25 2017
=== BISECT JOB RESULTS === Perf regression found but unable to narrow commit range Build failures prevented the bisect from narrowing the range further. Bisect Details Configuration: android_nexus7_perf_bisect Benchmark : resource_sizes Metric : MonochromePublic.apk_Specifics/normalized apk size Change : 0.02% | 73581091.0 -> 73597475.0 Suspected Commit Range 5 commits in range https://chromium.googlesource.com/chromium/src/+log/49736bfaaee2f7f1c641daf66516ce540eddcbd2..62eefa5666392672359a92bade4d89bc0dc7de34 Revision Result N chromium@466171 73581091 +- 0.0 6 good chromium@466171,v8@d089276e03 --- --- build failure chromium@466171,v8@f28e487858 --- --- build failure chromium@466171,v8@fa1de6145f --- --- build failure chromium@466171,v8@fe77234eb0 --- --- build failure chromium@466172 73597475 +- 0.0 6 bad To Run This Test src/build/android/resource_sizes.py --chromium-output-directory {CHROMIUM_OUTPUT_DIR} --chartjson {CHROMIUM_OUTPUT_DIR}/apks/MonochromePublic.apk Debug Info https://chromeperf.appspot.com/buildbucket_job_status/8981385680889989232 Is this bisect wrong? https://chromeperf.appspot.com/bad_bisect?try_job_id=5838626799222784 | 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 component Speed>Bisection. Thank you!
,
Apr 25 2017
Biggest jump (of 12kb) is from: https://codereview.chromium.org/2828933002 [console] fast console.assert(true) While 12kb on its own is below our size alert threshold, the commit description makes it sound like maybe this shouldn't have such an impact. Symbol diff from //tools/binary_size attached below. Looks like the entire increase is from: https://cs.chromium.org/chromium/src/v8/src/builtins/builtins.cc?q=CallableFor&sq=package:chromium&l=117 The same happened in recent change bug 714894 . Not sure if there's anything to be done here. 7 symbols added (+), 21 changed (~), 16 removed (-), 317530 unchanged (not shown) 0 object files added, 0 removed ~ 13428 t@0xef26b0 13428 v8/v8_base/builtins.o v8::internal::Builtins::CallableFor + 13520 t@0xef2500 92 v8/v8_base/builtins.o v8::internal::BinaryOpWithVectorDescriptor::BinaryOpWithVectorDescriptor + 13612 t@0xef2448 92 v8/v8_base/builtins.o v8::internal::BuiltinDescriptor::BuiltinDescriptor + 13704 t@0xef24a4 92 v8/v8_base/builtins.o v8::internal::CompareDescriptor::CompareDescriptor - 13612 t@0xef2898 -92 v8/v8_base/builtins.o v8::internal::CreateIterResultObjectDescriptor::CreateIterResultObjectDescriptor - 13520 t@0xef283c -92 v8/v8_base/builtins.o v8::internal::DeletePropertyDescriptor::DeletePropertyDescriptor - 13428 t@0xef2448 -92 v8/v8_base/builtins.o v8::internal::LoadDescriptor::LoadDescriptor - 13336 t@0xef24a4 -92 v8/v8_base/builtins.o v8::internal::LoadFieldDescriptor::LoadFieldDescriptor - 13244 t@0xef2614 -92 v8/v8_base/builtins.o v8::internal::LoadGlobalWithVectorDescriptor::LoadGlobalWithVectorDescriptor - 13152 t@0xef25b8 -92 v8/v8_base/builtins.o v8::internal::LoadICProtoArrayDescriptor::LoadICProtoArrayDescriptor - 13060 t@0xef255c -92 v8/v8_base/builtins.o v8::internal::LoadWithVectorDescriptor::LoadWithVectorDescriptor - 12968 t@0xef27e0 -92 v8/v8_base/builtins.o v8::internal::PerformNativePromiseThenDescriptor::PerformNativePromiseThenDescriptor - 12876 t@0xef2784 -92 v8/v8_base/builtins.o v8::internal::RejectNativePromiseDescriptor::RejectNativePromiseDescriptor - 12784 t@0xef2728 -92 v8/v8_base/builtins.o v8::internal::ResolveNativePromiseDescriptor::ResolveNativePromiseDescriptor - 12692 t@0xef2500 -92 v8/v8_base/builtins.o v8::internal::StoreWithVectorDescriptor::StoreWithVectorDescriptor - 12600 t@0xef2670 -92 v8/v8_base/builtins.o v8::internal::TypeConversionDescriptor::TypeConversionDescriptor - 12508 t@0xef26cc -92 v8/v8_base/builtins.o v8::internal::TypeofDescriptor::TypeofDescriptor + 12600 t@0xef255c 92 v8/v8_base/builtins.o v8::internal::WasmRuntimeCallDescriptor::WasmRuntimeCallDescriptor ~ 12516 t@0x12ee548 -84 v8/src/inspector/inspector/v8-console.o v8_inspector::V8Console::Assert
,
Apr 26 2017
The following revision refers to this bug: https://chromium.googlesource.com/v8/v8.git/+/a1a3090479cf8da10a2376f58ef2f33cc045eedb commit a1a3090479cf8da10a2376f58ef2f33cc045eedb Author: kozyatinskiy <kozyatinskiy@chromium.org> Date: Wed Apr 26 15:40:42 2017 [builtins] Builtins::CallableFor(): generate CPP case for ConsoleAssert only. To reduce size of Builtins::CallableFor function we can add only case which we actually use. BUG= chromium:714893 R=ishell@chromium.org Review-Url: https://codereview.chromium.org/2839933003 Cr-Commit-Position: refs/heads/master@{#44900} [modify] https://crrev.com/a1a3090479cf8da10a2376f58ef2f33cc045eedb/src/builtins/builtins.cc
,
May 2 2017
Assuming fixed.
,
May 9 2017
|
|||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||
Comment 1 by rsch...@chromium.org
, Apr 25 2017