Chrome uses 75% more CPU on airma.sh html game than Safari (and 18% more than Firefox) |
|||||||||||||||||
Issue descriptionGoogle Chrome 65.0.3302.0 (Official Build) canary (64-bit) Revision ed06b93000c86b85c098703b0763862e0ff9a699-refs/heads/master@{#526108} OS Mac OS X 10.12.8 JavaScript V8 6.5.102 This is on a 2017 15" Touchbar Macbook Pro. 1. Load https://airma.sh 1. Wait for the loading screen to show the spinning planes. The below numbers are from watching the spinning planes on the login screen. You can also show the difference by playing the game. == Chrome 65.0.3302.0 (Official Build) canary (64-bit) Renderer 14.5% Browser 10.7% GPU Process 9.5% WindowServer 10.6% 14.5+10.7+9.5+10.6 => 45.3% == Safari 11.0.2 Renderer 10.4% Browser 0.0% WindowServer 15.5% 10.4+15.5 => 25.9% == Firefox 57.0 (Quantum) Renderer 16.0% Browser 10.3% WindowServer 12.2% 16.0+10.3+12.2 => 38.5% This game is a really amazing HTML5 experience and runs great in all browsers, unfortunately it also seems to show Chrome uses considerably more CPU and power (as reported by energy monitor) than other browsers for gaming.
,
Jan 9 2018
,
Jan 17 2018
Attaching an Instruments 8.3.3 trace. The attached screenshots show one of the usual suspects: CPU in the main thread used to unmarshall mojo parameters, destroy mojo messages, etc.; 27% of CPU consumed by the IOThread, which is only servicing mojo messages (i.e. no mojo messages, no IOThread CPU consumption). > This may be inherent in Chrome's multi-process and GPU sandboxing architecture. Strictly more work is done when remoting graphics commands from the renderer process to the GPU process than in single-process browsers, but the security properties of Chrome's architecture are better. kbr@ - have you seen this doc: https://docs.google.com/document/d/1Ij2U31v3jq7X-4wx6Wtjan6c5GgNJUsZE6GRJZvISK0/edit?usp=sharing
,
Jan 18 2018
shrike@: no, hadn't seen it. Good and interesting writeup.
,
Jan 18
(4 days ago)
This issue has been Available for over a year. If it's no longer important or seems unlikely to be fixed, please consider closing it out. If it is important, please re-triage the issue. Sorry for the inconvenience if the bug really should have been left as Available. For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Jan 18
(4 days ago)
|
|||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||
Comment 1 by kbr@chromium.org
, Dec 24 2017Components: Internals>GPU>Internals