[llvm] r225517 - Assumption that "VectorizedValue" will always be an Instruction is not correct.

[Hal Finkel]

----- Original Message -----
> From: "Suyog Sarda" <suyog.sarda at samsung.com>
> To: llvm-commits at cs.uiuc.edu
> Sent: Friday, January 9, 2015 4:23:48 AM
> Subject: [llvm] r225517 - Assumption that "VectorizedValue" will always be an	Instruction is not correct.
> 
> Author: suyog
> Date: Fri Jan  9 04:23:48 2015
> New Revision: 225517
> 
> URL: http://llvm.org/viewvc/llvm-project?rev=225517&view=rev
> Log:
> Assumption that "VectorizedValue" will always be an Instruction is
> not correct.
> It can be a constant or a vector argument.
> 
> ex :
> 
> define i32 @hadd(<4 x i32> %a) #0 {
> entry:
>   %vecext = extractelement <4 x i32> %a, i32 0
>   %vecext1 = extractelement <4 x i32> %a, i32 1
>   %add = add i32 %vecext, %vecext1
>   %vecext2 = extractelement <4 x i32> %a, i32 2
>   %add3 = add i32 %add, %vecext2
>   %vecext4 = extractelement <4 x i32> %a, i32 3
>   %add5 = add i32 %add3, %vecext4
>   ret i32 %add5
> }
> 
> 
> Modified:
>     llvm/trunk/lib/Transforms/Vectorize/SLPVectorizer.cpp

I see from the review thread that Arnold said, "I am fine with this going in without a test case. This code is currently not enabled", however, in the future, when submitting a bug fix without test case, please explain why in the commit message.

Thanks,
Hal

> 
> Modified: llvm/trunk/lib/Transforms/Vectorize/SLPVectorizer.cpp
> URL:
> http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/Vectorize/SLPVectorizer.cpp?rev=225517&r1=225516&r2=225517&view=diff
> ==============================================================================
> --- llvm/trunk/lib/Transforms/Vectorize/SLPVectorizer.cpp (original)
> +++ llvm/trunk/lib/Transforms/Vectorize/SLPVectorizer.cpp Fri Jan  9
> 04:23:48 2015
> @@ -3502,11 +3502,10 @@ private:
>    /// \brief Emit a horizontal reduction of the vectorized value.
>    Value *emitReduction(Value *VectorizedValue, IRBuilder<> &Builder)
>    {
>      assert(VectorizedValue && "Need to have a vectorized tree
>      node");
> -    Instruction *ValToReduce =
> dyn_cast<Instruction>(VectorizedValue);
>      assert(isPowerOf2_32(ReduxWidth) &&
>             "We only handle power-of-two reductions for now");
>  
> -    Value *TmpVec = ValToReduce;
> +    Value *TmpVec = VectorizedValue;
>      for (unsigned i = ReduxWidth / 2; i != 0; i >>= 1) {
>        if (IsPairwiseReduction) {
>          Value *LeftMask =
> 
> 
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
> 

-- 
Hal Finkel
Assistant Computational Scientist
Leadership Computing Facility
Argonne National Laboratory