Unlike other types of render tasks, pictures can have hundreds of dependencies. The dependency vector is re-built every frame, leading to a lot of vector re-allocations in some pages.
Depends on D60151
Details
Details
- Reviewers
gw - Commits
- rMOZILLACENTRALdfe4fa983d2b: Bug 1607836 - Preallocate picture task dependency vectors. r=gw
- Bugzilla Bug ID
- 1607836
Diff Detail
Diff Detail
- Repository
- rMOZILLACENTRAL mozilla-central
- Lint
Automatic diff as part of commit; lint not applicable. - Unit
Automatic diff as part of commit; unit tests not applicable.
Event Timeline
Comment Actions
I wonder if there's a way to make such optimizations less custom each time.
For instance, I wonder if it might make sense to have a struct that holds "recyclable" vecs and at the end of the frame, and used vecs can be added to that from any subsystem, and then on the next frame, systems can claim vecs from that struct. Not exactly sure how it would work, but it might be worth pondering / discussing this in future?