[llvm-commits] PATCH: remove VICmp and VFCmp.

[Chris Lattner]

On Jul 9, 2009, at 11:47 AM, David Greene wrote:

> On Thursday 09 July 2009 13:24, Chris Lattner wrote:
>
>>> So the processor really operates with such bit-packed vectors,
>>> producing
>>> them as results for comparisons?
>>
>> If a processor has that, it can use intrinsics.  It's not as if two
>> chips that support <4 x i1> natively in hardware would agree on a
>> memory format anyway :)
>
> Well, this is what I've thought about for Larrabee but we'd be  
> forced to
> use intrinsics because the LLVM IR doesn't currently support a mask
> concept.  Unfortunately, that will SEVERELY limit optimization  
> opportunity.

I don't know enough about Larrabee to know what the "right thing" is  
for it.  However, I'm most concerned with nailing down the semantics  
we have *today*.  When someone cares enough about larrabee to do the  
right thing, we can extend the IR and/or revisit prior decisions as  
needed.

-Chris