Some background for this is in https://bugzilla.mozilla.org/show_bug.cgi?id=1805644#c16 AWSY doesn't show too much difference, possibly because of the 3rd patch on the patch queue I'm about to upload https://treeherder.mozilla.org/perfherder/compare?originalProject=mozilla-central&originalRevision=6424e727b6adc253c38b7d7d3c861c9360de74da&newProject=try&newRevision=5f6ae029066462d2c4e4c1ceb8d10aa77c41ebae&page=1&framework=4 BrowserTime really likes this https://treeherder.mozilla.org/perfherder/compare?originalProject=try&originalRevision=b3dfdc2535422cf85769eef6bb762befaeb9ffb6&newProject=try&newRevision=336b8407412e68151f1afa447bea363288f4b861&page=1&framework=13&showOnlyImportant=1 especially Speedometer* part of those tests https://treeherder.mozilla.org/perfherder/comparesubtest?originalProject=try&newProject=try&newRevision=336b8407412e68151f1afa447bea363288f4b861&originalSignature=4586009&newSignature=4586009&framework=13&originalRevision=b3dfdc2535422cf85769eef6bb762befaeb9ffb6&page=1&showOnlyImportant=1 The tweaks to the cache sizes in the 2nd patch are based on experimenting with different numbers. The non-default arenas we have really need to be larger than they currently are when running busy code, like speedometer. It is probably worth to experiment with the limits some more, but it is a bit time consuming to get all the performance and awsy numbers. Not sure what would be the right component for this, but since this is all about performance, Core: Performance it is :) I could have also split this to several bugs, but the patches are small and go all together.
Bug 1815069 Comment 0 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
Some background for this is in https://bugzilla.mozilla.org/show_bug.cgi?id=1805644#c16 AWSY doesn't show too much difference, possibly because of the 3rd patch on the patch queue I'm about to upload https://treeherder.mozilla.org/perfherder/compare?originalProject=mozilla-central&originalRevision=6424e727b6adc253c38b7d7d3c861c9360de74da&newProject=try&newRevision=5f6ae029066462d2c4e4c1ceb8d10aa77c41ebae&page=1&framework=4 BrowserTime really likes this https://treeherder.mozilla.org/perfherder/compare?originalProject=try&originalRevision=b3dfdc2535422cf85769eef6bb762befaeb9ffb6&newProject=try&newRevision=336b8407412e68151f1afa447bea363288f4b861&page=1&framework=13&showOnlyImportant=1 especially Speedometer* part of those tests https://treeherder.mozilla.org/perfherder/comparesubtest?originalProject=try&newProject=try&newRevision=336b8407412e68151f1afa447bea363288f4b861&originalSignature=4586009&newSignature=4586009&framework=13&originalRevision=b3dfdc2535422cf85769eef6bb762befaeb9ffb6&page=1&showOnlyImportant=1 The tweaks to the cache sizes in the 2nd patch are based on experimenting with different numbers. The non-default arenas we have really need to be larger than they currently are when running busy code, like speedometer. It is probably worth to experiment with the limits some more, but it is a bit time consuming to get all the performance and awsy numbers. Not sure what would be the right component for this, but since this is all about performance, Core: Performance it is :) I could have also split this to several bugs, but the patches are small and go all together. I need to still run this on tryserver to ensure I didn't break any checks on debug builds. I wouldn't be surprised if I missed something.