[llvm-commits] [llvm] r150060 - in /llvm/trunk: include/llvm/IntrinsicsX86.td lib/Target/X86/X86ISelLowering.cpp lib/Target/X86/X86InstrSSE.td

[Craig Topper]

On Wed, Feb 8, 2012 at 2:09 PM, David A. Greene <dag at cray.com> wrote:

> Eric Christopher <echristo at apple.com> writes:
>
> >> You see no benefit even though I have directly stated that their
> >> existance and use is critically important to us?
> >>
> >
> > I disagree with it. "My customers expect a particular sequence of code
> > generated for _mm_XXX_ZZZ" isn't a compiler problem, it's a frontend
> > problem if you don't want the backend to optimize it.
>
> How do you propose solving the problem without intrinsics?
>
> Remember that other code in the same Module should be optimized right up
> to and before/after each intrinsic.
>
> This is not an uncommon requirement for compilers.a
>


There are already numerous instructions in llvm that have no intrinsics
covering them. For instance, most shufffles (pshud, shufps, shufpd, vector
unpacks, pshuflw, pshufhw), basic vector addition and subtraction and I'm
sure many more. How was this "problem" not notice before now if its so
important?


>
>                            -Dave
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>



-- 
~Craig
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20120208/f708a099/attachment.html>