\ development could have looked like if it was a backwards compatible hardware architecture instead of an API compatible driver world.
-
-
-
though this seems somewhat analogous to how HW vendors handle DX byte code, itself like a backwards compatible ISA
- Show replies
New conversation -
-
-
IMO part of the problem here is that a huge hole has been dug: too many features, all conflicting
-
*and* at the same time, IHVs are responsible for continual ugly perf hacks to make X latest game be Y% faster
- Show replies
New conversation -
-
-
and the fact that performance on PC is, for many applications (not VR!) "solved", trades get different values.
-
Don't think NVIDIA/AMD/Intel would ever settle on common GPU instruction set.
End of conversation
New conversation -
-
-
x86 changed instruction set and embedded internal registers all the time. The instruction set is a minor part of GPUs.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
we have N conflicting standards. let's solve the problem by adding another standard,
-
Ouch. I'm alternate-history-ing here, not making an actual engineering suggestion. :-)
- Show replies
New conversation -
-
-
With Moore's Law dying will get a clean slate CPU design that tries optimizing for gains instead of shrinking the die?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
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.