[PATCH][AArch64] Implemented Neon scalar vdup_lane intrinsics

[Ana Pazos]

Thanks Jiangning,

I will removed the redundant v1i64 pattern and submit the patch.

Thanks,

Ana.

 

From: Jiangning Liu [mailto:liujiangning1 at gmail.com] 
Sent: Tuesday, November 19, 2013 11:46 PM
To: Ana Pazos
Cc: cfe-commits at cs.uiuc.edu; llvm-commits at cs.uiuc.edu for LLVM; Tim
Northover
Subject: Re: [PATCH][AArch64] Implemented Neon scalar vdup_lane intrinsics

 

Hi Ana,

 

+def : Pat<(v1i64 (extract_subvector (v2i64 VPR128:$Rn),
neon_uimm1_bare:$Imm)),

+          (v1i64 (DUPdv_D VPR128:$Rn, neon_uimm1_bare:$Imm))>

 

This rule should be a redundant one, because we already have the following
patterns to cover all cases around DUPdv_D.

 

multiclass NeonI_SDUP<PatFrag GetLow, PatFrag GetHigh,

                 ValueType ResTy, ValueType OpTy> {

  def : Pat<(ResTy (GetLow VPR128:$Rn)),

            (ResTy (DUPdv_D (OpTy VPR128:$Rn), 0))>;

  def : Pat<(ResTy (GetHigh VPR128:$Rn)),

            (ResTy (DUPdv_D (OpTy VPR128:$Rn), 1))>;

}

 

However, the name NeonI_SDUP is inconsistent with the one you are adding
like NeonI_Scalar_DUP_Vec_pattern, so maybe you can change it to be
NeonI_Scalar_DUP_Vec_pattern_D or something else like that.

 

All others LGTM.

 

Thanks,

-Jiangning

 

2013/11/20 Ana Pazos <apazos at codeaurora.org>

Hi Tim and reviewers,

 

This patch supports scalar vdup_lane intrinsics, a fix for scalar dup alias
and test cases.

 

This completes the scalar intrinsics implementation.

 

I would like to push these patches (and the ones I posted earlier today for
scalar by element intrinsics) to the release 3.4 branch as well.

 

Please review them and let me know if ok to check them in mainline and
release 3.4.

 

Thanks,

Ana.


_______________________________________________
cfe-commits mailing list
cfe-commits at cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits





 

-- 

Thanks,

-Jiangning

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20131120/6b5e2516/attachment.html>