I know that Epic is currently working on parallel rendering and you got a task “Baseline rendering performance optimizations” that’s in the backlog, but are there any single-thread rendering optimizations coming? Are you accepting more pull requests for that?
Related to that I got two pull requests which are still in the queue since beginning of May and I haven’t got any comment on them. The two requests are these, and they are both related to the render thread:
- Optimize BuildTileVertexBuffer() block throughput from 13.45 cycles to 12.90 cycles on Sandy Bridge.
- Optimize BuildParticleVertexBuffer() block throughput from 31.55 cycles to 23.40 cycles on Sandy Bridge.
Should I invest more time optimizing the work happening on the rendering thread or should I wait until the work on parallel rendering is done?