-
Notifications
You must be signed in to change notification settings - Fork 36
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Support for multisampled textures + resolve?\ #198
Comments
I haven't investigated this again since WebGPU times, but I think the support for MSAA will largely look the same as in wgpu. The sample count needs to be used and respected by the texture creation. Then we'll need extra data in |
I made some experiments in my fork and also found that there was a ClearOp::ResolveTo(TextureView) in the code which I successfully used together with some more info in the texturedesc for sample count and some more msaa state in the renderpipelinedesc to get msaa working for the particles demo. I did have some issues with it though on my intel integrted graphics laptop as well as an intel+nvidia machine, though my Ryzen+AMD gpu seemed to work fine. None of the targets had validation errors. I will investigate more and get back, though on vacation atm |
Yes, I think this is a fine way to expose it in the API. Technically, one could try to both store the result and resolve it in Vulkan, but that isn't very useful. If we provide this through |
#203 looks like a nice addition! I am still on vacation and will be back in December so will not be able to contribute much work until then sadly. So if a release is in order then go ahead! (On a side-note I also started investigating https://docs.vulkan.org/features/latest/features/proposals/VK_KHR_dynamic_rendering_local_read.html on the airplane as well as it might become relevant for our company in the future too, though support in drivers seemed not great at the moment) |
I see there are some indications in the code that
sample_count
was almost added to the texture descriptor but don't see it implemented yet.What would be the challenges in supporting multisampled textures and resolve targets for render passes? I would be happy to help though I have mostly experience with wgpu:s user-facing code
(side note: Such a fresh breath of air to read this code! It's all so simple.. blade-graphics looks very promising and I would like to give blade a try for our rendering engine at my company that's currently using wgpu (for this game https://idno.se/swap/). Seeing your presentation and descirbing the simple binding model in blade really motivated me after having to deal with so much pipeline description and setup of bind groups for our multiple shader permutations... )
The text was updated successfully, but these errors were encountered: