Issue metadata
Sign in to add a comment
|
cc_perftests failure on android/linux chromium.perf at 393644:393658 |
||||||||||||||||||||
Issue descriptionRevision range first seen: Linux: 393624:393687 N5: 393639:393692 N6: 393616:393688 N7v2: 393644:393695 N9: 393588:393658 Link to failing step log: https://uberchromegw.corp.google.com/i/chromium.perf/builders/Linux%20Perf%20(3)/builds/16351/steps/cc_perftests/logs/stdio [ RUN ] TileManagerPerfTest.EvictionTileQueueConstructAndIterate *RESULT tile_manager_eviction_tile_queue_construct_and_iterate: 2_16= 211433.9375 runs/s *RESULT tile_manager_eviction_tile_queue_construct_and_iterate: 2_32= 122527.0625 runs/s *RESULT tile_manager_eviction_tile_queue_construct_and_iterate: 2_64= 76584.2734375 runs/s ../../cc/tiles/tile_manager_perftest.cc:191: Failure Value of: queue->IsEmpty() Actual: true Expected: false *RESULT tile_manager_eviction_tile_queue_construct_and_iterate: 10_16= 141325.625 runs/s *RESULT tile_manager_eviction_tile_queue_construct_and_iterate: 10_32= 97582.8046875 runs/s *RESULT tile_manager_eviction_tile_queue_construct_and_iterate: 10_64= 61614.96875 runs/s *RESULT tile_manager_eviction_tile_queue_construct_and_iterate: 10_128= 35104.47265625 runs/s *RESULT tile_manager_eviction_tile_queue_construct_and_iterate: 50_16= 48373.62109375 runs/s *RESULT tile_manager_eviction_tile_queue_construct_and_iterate: 50_32= 41088.7890625 runs/s *RESULT tile_manager_eviction_tile_queue_construct_and_iterate: 50_64= 32548.404296875 runs/s *RESULT tile_manager_eviction_tile_queue_construct_and_iterate: 50_128= 22726.23828125 runs/s [ FAILED ] TileManagerPerfTest.EvictionTileQueueConstructAndIterate (22019 ms If the test is disabled, please downgrade to Pri-2. jaydasika, could this be related to https://codereview.chromium.org/1975623002 vmpstr, could it be https://codereview.chromium.org/1974033004
,
May 16 2016
I'll take a look.
,
May 16 2016
It was crrev.com/631c144 that broke this. Working on a fix now, it should be quick.
,
May 17 2016
Hi vmpstr@, How's the fix coming along?
,
May 18 2016
(Talked to vmpstr@ offline: sounds like he thinks he can get the fix in before lunch, which means that it's probably not worth disabling this benchmark.)
,
May 18 2016
Actually, the offending patch was reverted last night, so I'll close this and ensure that all tests work when I reland. Thanks! |
|||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||
Comment 1 by jaydasika@chromium.org
, May 16 2016