So given that the WebGPU developers don't have any influence over OS GPU abstractions, do you think they should have given up and not made anything, or write compilers that can accept all major shading languages on any OS, or are you saying SPIR-V?
-
-
Cool thanks. This sounds like the nascent SPIR-V ecosystem and Mesa Vulkan drivers, except I'm guessing you want a lower level abstraction like CUDA PTX, which would come at perf cost when up-converting to high-level APIs.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
The representation could be an already-existing one (e.g. SPIR-V if you think that is okay), or a modification of that, if judged to be good enough. But the point is to push the ball in the right direction instead of the wrong direction.
-
I am suspicious of SPIR-V (look at all the weird special hardcoded stuff that's in there already), but have not used it so won't comment.
- Show replies
New conversation -
-
-
Intel are doing close to this. It's hard to see through the fog of uncertainty, but OneAPI is groping toward that goal. It's a bunch of tools and languages and so on with interop as the focus. Of course it's Yet More Standards in the short term. We'll see where it leads.
-
An obvious tensions is between what Jon wants which is a super low level "assembly" that works on every GPU, and Intel's 50 years of experience doing this on CPUs where it's a gigantic pain in the arse.
- Show replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.