Now this is decoupled, and shadow tracing work for many bounces can be accumulated in a queue. Previously, continuing to the next bounce would have to wait for all light and shadows to be resolved at the previous bounce. Now we detect low GPU occupancy and adaptively increase the number of samples to batch together, which then increases occupancy.Īnother part of the solution was to change the shadow kernel scheduling. Previously we’d end up rendering 1 sample at a time. This was improved by making the algorithm to estimate the number of samples to render in one batch smarter. One issue we found is that in GPU rendering, if only a small subset of the whole image is slow to render (like a character’s hair) then GPU occupancy would be low. While most benchmark scenes were rendering faster with Cycles X, a few involving many layers of transparent hair were showing performance regressions compared to 2.93. Since then we have restored volume rendering, and found that GPU rendering performance improved 3-5x in various volume scenes. For details, see the GPU kernel documentation and GPU performance development tasks.Īt the time of the initial announcement there was no volume rendering support. This is an accumulation of many incremental changes. Render time on an NVIDIA Quadro RTX A6000 with OptiX GPU rendering performance has been further improved. Well give a quick overview of recent developments.
Since the announcement, developers have been working to complete and stabilize the code, as well add new features and improve performance. The rendering improvements from the Cycles X project will be in the upcoming Blender 3.0 release.
0 kommentar(er)
