maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity

[CodeGen][SVE] Avoid scalarizing zero splat stores on scalable vectors. 

Summary: Implemented in replaceZeroVectorStore(). Fixes several warnings in AArch64 SVE unit tests.

Reviewers: sdesmalen, kmclaughlin, dancgr, efriedma, each, andwar, rengolin

Reviewed By: sdesmalen

Subscribers: tschuett, kristof.beyls, hiraditya, rkruppe, psnobl, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D80824
This commit is contained in:
Henry Kao 2020-06-09 12:33:47 -04:00 committed by Danilo C. Grael
parent 8233439fdb
commit 4dcc0d1958
1 changed files with 4 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
View File

@ -12069,6 +12069,10 @@ static SDValue replaceZeroVectorStore(SelectionDAG &DAG, StoreSDNode &St) {
SDValue StVal = St.getValue();
EVT VT = StVal.getValueType();
// Avoid scalarizing zero splat stores for scalable vectors.
if (VT.isScalableVector())
return SDValue();
// It is beneficial to scalarize a zero splat store for 2 or 3 i64 elements or
// 2, 3 or 4 i32 elements.
int NumVecElts = VT.getVectorNumElements();