On Mon, May 20, 2019 at 03:54:49PM -0600, Jeff Law wrote:
> >>> Adjust gcc.target/i386/pr22076.c for 64-bit.
> >>>
> >>> * gcc.target/i386/pr22076.c: Adjusted for 64-bit.
> >> Well, it looks like you're just papering over a code quality regression?
> >> Or am I missing something?
> >
> > We have to load 64bit value from memory to 128 bit XMM register using movq.
> >
> > OTOH, we could just use -mno-sse2 which disables XMM emulation.
> Ah, we can't have a MEM operand on paddb for xmm registers...
We can, but then it will load 128-bits from memory instead of just 64-bits,
it might not be sufficiently aligned, unaligned load might cross page
boundary, ...
I think what we generate is fine.
Jakub
