[LLVMdev] GCCBuiltin and Intrinsic Mapping

Eric Christopher echristo at apple.com
Mon Sep 13 13:42:27 PDT 2010


On Sep 13, 2010, at 1:24 PM, Bruno Cardoso Lopes wrote:

> On Mon, Sep 13, 2010 at 8:27 AM, David A. Greene <dag at cray.com> wrote:
>> Eli Friedman <eli.friedman at gmail.com> writes:
>> 
>>> int_x86_avx_vhadd_pd_xmm doesn't exist on trunk.  Why does it exist on
>>> your branch if the semantics are exactly equivalent to
>>> int_x86_sse3_hadd_pd?  The register allocator can handle converting to
>>> three-address form if the target provides the appropriate hooks.
>> 
>> Because in some cases users may want to explicitly use non-VEX encoded
>> instructions.  So we need to differentiate.
> 
> I don't see why one would like to emit 256-bit wide reg instructions and at the
> same time non-VEX encoded 128-bit ones. For cases like this one can compile
> the former alone and then link with regular sse code, right?
> 

Yep.  I don't see any reason either.

-eric



More information about the llvm-dev mailing list