[LLVMdev] Should more vector [zs]extloads be legal for X86 SSE4.1?
chandlerc at gmail.com
Thu Dec 4 11:14:19 PST 2014
On Thu, Dec 4, 2014 at 11:02 AM, Ahmed Bougacha <ahmed.bougacha at gmail.com>
> So I gave it a shot, but to avoid an ugly hack in X86, I tried to do
> the right thing and changed the extload legalization logic to have a
> 2D array (memory->result types, not just memory), which just seems
> like the saner thing to do:
I think the lack of this is what caused me to give up, so, awesome!
> The X86 stuff is much simpler:
I'll try to take a look at these soon, but the first one should get eyes
other than mine on it (Tim? Jim? Others?).
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the llvm-dev