Following up from 1650403, it would be worth exploring spatial heuristics to break clusters (especially large ones) in order to take advantage of culling at the cluster level and skipping more primitives during the visibility pass. There is also this comment from Glenn in https://phabricator.services.mozilla.com/D82192 to consider: [...] we could in theory remove the code at [1] which clears out the prim instance info for prims that are in an invisible cluster. Although, it's probably best / safer to keep this, unless we have a way to validate / assert that we don't access stale indices by mistake. [1] https://searchfox.org/mozilla-central/rev/91d82d7cbf05a71954dfa49d0e43824c7c973e62/gfx/wr/webrender/src/visibility.rs#270
Bug 1651018 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.
Following up from bug 1650403, it would be worth exploring spatial heuristics to break clusters (especially large ones) in order to take advantage of culling at the cluster level and skipping more primitives during the visibility pass. There is also this comment from Glenn in https://phabricator.services.mozilla.com/D82192 to consider: [...] we could in theory remove the code at [1] which clears out the prim instance info for prims that are in an invisible cluster. Although, it's probably best / safer to keep this, unless we have a way to validate / assert that we don't access stale indices by mistake. [1] https://searchfox.org/mozilla-central/rev/91d82d7cbf05a71954dfa49d0e43824c7c973e62/gfx/wr/webrender/src/visibility.rs#270