Experimenting with different threading models like cull/draw gives significant boosts in speed. I’m wondering whether one can truly shift between threading and non-threading models at runtime. This has been answered here, but I don’t notice any difference when I use it:
But this means the existing window vanishes for a few seconds, before a new one is opened, something that is obviously jarring. So does anyone know of how to truly seamlessly transition between threaded and non-threaded models, where app, draw and cull are separated and fused alternatively at runtime? I’m thinking of maybe using a decoy window during transitions, but would that be the only way?
Hmm, I thought it was supposed to work, but it does seem there’s some preparation work that happens only at window open time. I’m not sure there’s a good reason for that - feel free to file a feature request on GitHub for this.