[llvm] 2516376 - [AArch64] Enhance last active true vector combine

[via llvm-commits]

Author: zhongyunde
Date: 2022-04-06T09:54:28+08:00
New Revision: 251637690ab486bd13d8408453314aa933501dd0

URL: https://github.com/llvm/llvm-project/commit/251637690ab486bd13d8408453314aa933501dd0
DIFF: https://github.com/llvm/llvm-project/commit/251637690ab486bd13d8408453314aa933501dd0.diff

LOG: [AArch64] Enhance last active true vector combine

Last active extracting will output LASTB + WHILELS, and the WHILELS itself
is a flag-setting operation, so perform it preferly.

Reviewed By: paulwalker-arm, sdesmalen

Differential Revision: https://reviews.llvm.org/D122551

Added: 
    

Modified: 
    llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
    llvm/test/CodeGen/AArch64/sve-extract-element.ll

Removed: 
    


################################################################################
diff  --git a/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp b/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
index f6e1c6140147e..68c8d73fbcb6f 100644
--- a/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
+++ b/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
@@ -14551,15 +14551,15 @@ performLastTrueTestVectorCombine(SDNode *N,
   if (!Subtarget->hasSVE() || DCI.isBeforeLegalize())
     return SDValue();
 
-  SDValue SetCC = N->getOperand(0);
-  EVT OpVT = SetCC.getValueType();
+  SDValue N0 = N->getOperand(0);
+  EVT OpVT = N0.getValueType();
 
   if (!OpVT.isScalableVector() || OpVT.getVectorElementType() != MVT::i1)
     return SDValue();
 
   // Idx == (add (mul vscale, NumEls), -1)
   SDValue Idx = N->getOperand(1);
-  if (Idx.getOpcode() != ISD::ADD)
+  if (Idx.getOpcode() != ISD::ADD || !isAllOnesConstant(Idx.getOperand(1)))
     return SDValue();
 
   SDValue VS = Idx.getOperand(0);
@@ -14570,16 +14570,10 @@ performLastTrueTestVectorCombine(SDNode *N,
   if (VS.getConstantOperandVal(0) != NumEls)
     return SDValue();
 
-  // Restricted the DAG combine to only cases where we're extracting from a
-  // flag-setting operation
-  auto *CI = dyn_cast<ConstantSDNode>(Idx.getOperand(1));
-  if (!CI || !CI->isAllOnes() || SetCC.getOpcode() != ISD::SETCC)
-    return SDValue();
-
   // Extracts of lane EC-1 for SVE can be expressed as PTEST(Op, LAST) ? 1 : 0
   SelectionDAG &DAG = DCI.DAG;
   SDValue Pg = getPTrue(DAG, SDLoc(N), OpVT, AArch64SVEPredPattern::all);
-  return getPTest(DAG, N->getValueType(0), Pg, SetCC, AArch64CC::LAST_ACTIVE);
+  return getPTest(DAG, N->getValueType(0), Pg, N0, AArch64CC::LAST_ACTIVE);
 }
 
 static SDValue

diff  --git a/llvm/test/CodeGen/AArch64/sve-extract-element.ll b/llvm/test/CodeGen/AArch64/sve-extract-element.ll
index 393f894e4b659..6e3da13ae77ba 100644
--- a/llvm/test/CodeGen/AArch64/sve-extract-element.ll
+++ b/llvm/test/CodeGen/AArch64/sve-extract-element.ll
@@ -504,12 +504,9 @@ define i1 @test_lane9_8xi1(<vscale x 8 x i1> %a) #0 {
 define i1 @test_last_8xi1(<vscale x 8 x i1> %a) #0 {
 ; CHECK-LABEL: test_last_8xi1:
 ; CHECK:       // %bb.0:
-; CHECK-NEXT:    cnth x8
-; CHECK-NEXT:    mov z0.h, p0/z, #1 // =0x1
-; CHECK-NEXT:    sub x8, x8, #1
-; CHECK-NEXT:    whilels p0.h, xzr, x8
-; CHECK-NEXT:    lastb w8, p0, z0.h
-; CHECK-NEXT:    and w0, w8, #0x1
+; CHECK-NEXT:    ptrue p1.h
+; CHECK-NEXT:    ptest p1, p0.b
+; CHECK-NEXT:    cset w0, lo
 ; CHECK-NEXT:    ret
   %vscale = call i64 @llvm.vscale.i64()
   %shl = shl nuw nsw i64 %vscale, 3