It appears that after the whole 970 VRAM fiasco, there is a new story that haunts NVIDIA. As we’ve already reported, NVIDIA’s GPUs are really weak in Async Compute scenarios. And according to some Oculus employees, Preemption for context switches is possibly catastrophic for NVIDIA’s GPUs.
But first things first. What is Preemption? PreEmption allows tasks with a higher priority (which can be set manually or automatically) to go and be processed first, and less time sensitive tasks will be forced to wait until that work is completed. The graphics cards handle this by use of Context Switching, however in DX11 Context Switching can cause a lot of idle time (due to the API’s serial thinking). That’s where Async Compute/Shading comes to save the day.
As it’s been pointed out by AMD, Async Compute/Shading can counter the latency introduced by Preemption Context Switching by interleaving these tasks across multiple threads to shorten overall render time.
Now as Microprocessor analyst, David Kanter, said, some Oculus employees told him that the best Preemption support for context switches was with AMD by far, Intel was pretty good and NVIDIA was possibly catastrophic (1:21:00 in the video below).
David Kanter added:
“NVIDIA is very – to their credit – open and honest about this and how you tune for Oculus Rift is that you have to be super careful because you can miss a frame boundary because the preemption is not particularly well latency. And again, this is, it’s not like it’s a bad decision on the part of NVIDIA. It’s, you know, that’s just what made sense. Preemption was not something that was super important when the chip was designed and the API support was… there wasn’t much bang for your buck.”
David concluded that now NVIDIA will – most probably – improve this thing in Pascal.
Still, we have to give credit to AMD for offering more future-proof graphics cards to its customers, even though NVIDIA’s GPUs are, without a doubt, faster under DX11.