[PATCH] D30296: v2f32 ops

[Jonas Paulsson via Phabricator via llvm-commits]

jonpa created this revision.
Herald added a subscriber: aemerson.

A few days ago I described this issue on llvm-commits along with patch:

> Hi,
> 
> I found that on SystemZ, for v2f32, four and not two scalar operations are emitted. This is because the v2f32 type is widened, which is good in cases of memory-only operations for instance. There is however no fp32 vector support on z13, so these will always be scalarized. If this is done after type-legalization, four and not two operations are produced, which is particularly bad in case of fp32 divide inside a vectorized loop.
> 
> In order to fix this, my patch unrolls these operations before type legalization (with a target DAG combine). They must also have the operation action of 'Expand', since otherwise other DAGCombiner methods may re-vectorized them again. This happened in reduceBuildVecConvertToConvertBuildVec(), where I also needed to add a call to TLI.isOperationLegalOrCustom(Opcode, VT), to check on the *result* type, which would in this case be v2f32. It would not work to
>  mark all the operand VT's of SINT_TO_FP as 'Expand', because this is only true for the v2f32 result case.
> 
> I do get some failing regression tests, which I am not sure about:
> 
> Failing Tests (3):
> 
>   LLVM :: CodeGen/ARM/vdup.ll
>   LLVM :: CodeGen/X86/2009-02-26-MachineLICMBug.ll
>   LLVM :: CodeGen/X86/cvtv2f32.ll
>    
> 
> Is it ok to add the check for the result type per what I did?
> 
> /Jonas

In short, this is something I did for the SystemZ target, but because I had to add an extra check in reduceBuildVecConvertToConvertBuildVec(), three tests failed.

I have now tried to regenerate the tests, but only one of them was successfully regenerated. I include that test diff, plus the two output diffs of the two remaining test cases.

I hope that either the tests have been improved as expected (which I can't tell for myself for sure), or that someone gives me a pointer on how to change the patch.

Thanks

/Jonas


https://reviews.llvm.org/D30296

Files:
  lib/CodeGen/SelectionDAG/DAGCombiner.cpp
  lib/Target/SystemZ/SystemZISelLowering.cpp
  test/CodeGen/SystemZ/fp32-vec-conv.ll
  test/CodeGen/SystemZ/fp32-vec-ops.ll
  test/CodeGen/X86/cvtv2f32.ll
  test_CodeGen_ARM_vdup.diff
  test_CodeGen_X86_MLICMbug.diff

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D30296.89502.patch
Type: text/x-patch
Size: 12574 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20170223/729574e7/attachment.bin>