[llvm] r204972 - [X86][Vector Cost Model] Add a comment to explain the workaround

[Eric Christopher]

On Thu, Mar 27, 2014 at 3:27 PM, Quentin Colombet <qcolombet at apple.com> wrote:
> Author: qcolombet
> Date: Thu Mar 27 17:27:41 2014
> New Revision: 204972
>
> URL: http://llvm.org/viewvc/llvm-project?rev=204972&view=rev
> Log:
> [X86][Vector Cost Model] Add a comment to explain the workaround
> in my previous commit (r204884).
>
> <rdar://problem/16381225>

Thanks for following up on this Quentin!

-eric

>
> Modified:
>     llvm/trunk/lib/Target/X86/X86TargetTransformInfo.cpp
>
> Modified: llvm/trunk/lib/Target/X86/X86TargetTransformInfo.cpp
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86TargetTransformInfo.cpp?rev=204972&r1=204971&r2=204972&view=diff
> ==============================================================================
> --- llvm/trunk/lib/Target/X86/X86TargetTransformInfo.cpp (original)
> +++ llvm/trunk/lib/Target/X86/X86TargetTransformInfo.cpp Thu Mar 27 17:27:41 2014
> @@ -512,6 +512,11 @@ unsigned X86TTI::getCastInstrCost(unsign
>      { ISD::UINT_TO_FP,  MVT::v4f64, MVT::v4i8,  2 },
>      { ISD::UINT_TO_FP,  MVT::v4f64, MVT::v4i16, 2 },
>      { ISD::UINT_TO_FP,  MVT::v4f64, MVT::v4i32, 6 },
> +    // The generic code to compute the scalar overhead is currently broken.
> +    // Workaround this limitation by estimating the scalarization overhead
> +    // here. We have roughly 10 instructions per scalar element.
> +    // Multiply that by the vector width.
> +    // FIXME: remove that when PR19268 is fixed.
>      { ISD::UINT_TO_FP,  MVT::v2f64, MVT::v2i64, 2*10 },
>      { ISD::UINT_TO_FP,  MVT::v4f64, MVT::v4i64, 4*10 },
>
>
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits