...I figured I could skip the post tess vertex stage alltogether and just feed the tess data directly into a kernelhttps://twitter.com/warrenm/status/989170503662292994 …
Gotcha. I’m not aware of a way to do precisely that (feeding tessellated vertices to another kernel without an intervening render pipeline), but it sounds like you have a solution that works. Are you finding performance to be acceptable?
-
-
Hi Warren, just a follow up, while the performance was acceptable using a vertex shader as an intervening pipeline, I ended up not using tessellation at all. Instead doing pseudo tess by custom impl. in a kernel, then feeding that 2 another kernel. More involved but faster.
-
Thanks for the follow-up. Glad you found a better solution. If possible, I’d encourage you to file an enhancement request Radar documenting your use case, as I think it might be useful to others.
-
Sure, I will do that. Thanks for taking the time to help.
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.