- From: Joe Pea via GitHub <sysbot+gh@w3.org>
- Date: Wed, 05 Aug 2020 22:23:07 +0000
- To: public-houdini-archive@w3.org
I made two tests. For each test, - let it run without opening devtools for 10 seconds or so (to ensure no deoptimization from devtools), - and make sure the window size for the drawing area is the same for both in case that makes any difference. These are the tests: - Test 1: https://codepen.io/trusktr/pen/f3194e541c2088d46a7eab72f3e58cbe?editors=1010 - This test does not read the style attribute value (see line 84), but instead creates a dummy string and pushes it into an array of strings. - Test 2: https://codepen.io/trusktr/pen/9e2f452d26eb035055922dad91e9f436?editors=1010 - This test actually reads the style attribute value, and pushes the result into an array of strings exactly the same way. After you wait for a while (f.e. 10 seconds), open devtools for each test, and you'll see the average results outputted for every 100 animation frames. What I found is that the average speed for the first one is almost 4 times faster. I suppose this means that in the second test the reading of the style attribute value does in fact de-optimize (as expected), so this means Chrome seems to have the above optimization in place. I'm in Linux, so I can't test in Safari. Firefox doesn't have `attributeStyleMap` yet, so I didn't test there. -- GitHub Notification of comment by trusktr Please view or discuss this issue at https://github.com/w3c/css-houdini-drafts/issues/997#issuecomment-669551661 using your GitHub account -- Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config
Received on Wednesday, 5 August 2020 22:23:09 UTC