Dzmitry Malyshau

I don't think there is a downside for us to require both viewport and scissor rects to be limited to the render pass attachments size.

@Richard-Yunchao good point! @jdashg what does WebGL do today to handle resize? Does it care about portable behavior during resize?

This might be a little bit silly, but there is an idea to mess with Metal viewport in order to make sure the draw calls are not rasterized. I.e. if...

Yeah, I wasn't sure if empty viewport is valid on our target APIs. Needs to be investigated, if it hasn't been already.

That awkward moment when Metal is more Vulkan than Vulkan? Seriously though, Vulkan is full of compromises. Having large opaque structs that hide actual low-level bits and knobs is not...

Interestingly, ARM results of [benchmarking Vulkan Sub-passes from 2017](https://www.khronos.org/assets/uploads/developers/library/2017-gdc/GDC_Vulkan-on-Mobile_Vulkan-Multipass-ARM_Mar17.pdf) roughly matches your numbers: they got 30% better FPS and 80% reduction in bandwidth.

> In SPIR-V, you can declare an image with dimension`SubpassData` (as opposed to dimension `2D` or `3D` or whatever). Then, you can do `OpImageRead` to read from any position (!!)...

@litherum thank you for the proposal! Here is some use of sub-passes for Order-Independent transparency: https://matthewwellings.com/blog/depth-peeling-order-independent-transparency-in-vulkan/ Do I understand your proposal correctly that: if the intermediates are provided to the...

@Kangz as far as I can see, the dependencies are still a part of the RP compatibility rules (and it kinda makes sense to me): > Two render passes are...

@floooh thank you for filing this! So, if I understand you correctly, the downside you are worried about is the memory consumption? We are looking at, basically, difference between 2x...