On Dec 7, 2007, at 8:13 AM, Chris Lattner wrote:

> Hey Evan,
>
>> +++ llvm/trunk/lib/Target/X86/X86ISelLowering.cpp Fri Dec  7
>> 02:07:39 2007
>> @@ -2754,10 +2754,33 @@
>>   }
>>
>>   std::swap(V1, V2);
>> -  Mask = DAG.getNode(ISD::BUILD_VECTOR, MaskVT, &MaskVec[0],
>> MaskVec.size());
>> +  Mask = DAG.getNode(ISD::BUILD_VECTOR, MaskVT, &MaskVec[0],
>> NumElems);
>>   return DAG.getNode(ISD::VECTOR_SHUFFLE, VT, V1, V2, Mask);
>> }
>>
>> +static
>> +SDOperand CommuteVectorShuffleMask(SDOperand Mask, SelectionDAG
>> &DAG) {
>
> Please add a comment over this that gives an example of what it
> does.  I kept misreading the name as "compute..." and couldn't figure
> out what it did :)

Ok.

>
>
>> +static
>> +SDOperand LowerVECTOR_SHUFFLEv8i16(SDOperand V1, SDOperand V2,
>> +                                   SDOperand PermMask,
>> SelectionDAG &DAG,
>> +                                   TargetLowering &TLI) {
>
> Does this cause any pessimizations for v8i16 shuffles that could be
> done as v4i32?

No. Because the lowering before this patch is even worse. :-) I plan  
on improving this further.

Evan

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

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

Reply via email to