If someone wonders why the G-buffer only uses 18 MB of ESRAM, the answer is: we are doing lighting, post processing, VT page generation and particle rendering using async compute in the background. Only culling + g-buffer draws + shadow draws are in the render pipeline. Everything else is async compute. This gives us huge perf gains. Unfortunately I had only a 25 minute slot, meaning that I had to remove all my async compute slides.
If it's not too much to ask Sebbbi, is it possible to post those cut async compute slides as well? I guess things should be more open to discuss now about DX12, I was speaking with @Max McMullen about Multi-Adaptor, and how iGPU mixed with dGPU, DX12 can have the dGPU to signal the iGPU to do work when it's done. But not mentioned in those slides is that during our chat he mentioned that the GPU if the developer chose to, could signal the CPU for the next step in the pipeline.
I'm wondering if it's possible using async compute to have a core listening for those signals from the GPU to make an async call back to the gpu? This is a round about way of having a GPU issue its own draw calls?