Degerz
Degerz
I'd prefer it if every _extension_ exposed were in **ALL** of the supported APIs and had wide ranging hardware support as well. Exposing shader framebuffer fetch just after the public...
Having the feature alone supported in all of the APIs isn't good enough. Need to be considerate about the hardware support and hardware vendors as well. If we're going to...
> Direct3D12 is very conservative already anyways. The 3 API's being targeted are only for newish hardware anyways (plenty of Direct3D11 class cards will never have Direct3D 12 drivers). The...
I didn't say Fermi was getting Vulkan drivers but it does have D3D12 drivers so it would absolutely be useful if WebGPU can target Fermi even if only for Windows...
This is more a more reasonable request compared to exposing shader framebuffer fetch since AMD and Nvidia hardware can support this extension but let's leave it on the backburner for...
It's probably for the best to minimize the number of extensions in the not too distant future since it's a rather harmful feature in terms of portability ...
I do not believe WebGPU should be the place intended to push for standardizing new hardware functionality among IHVs since it's supposed to serve as a wrapper for D3D12/Vulkan/Metal. I...
> So, it is an extension that all the API's have, so it should be an extension in WebGPU. Not good enough as I mentioned in the other thread. Need...
As far as "high-performance" is concerned, it's even recommended by Nvidia in their DX12 [advices](https://developer.nvidia.com/dx12-dos-and-donts#Features) that the app developer compare using ROVs to _alternative_ approaches ... Not only is support...
After having seen some numbers collected on AMD HW I'm going to have to protest this idea ... Sample: https://software.intel.com/en-us/articles/oit-approximation-with-pixel-synchronization Code: https://github.com/GameTechDev/AOIT-Update AMD RX 5700, ROV off -> on (2-node):...