[llvm] e975382 - [PowerPC] Respect SDNodeFlags in lowering SELECT_CC
Qiu Chaofan via llvm-commits
llvm-commits at lists.llvm.org
Tue May 12 23:05:58 PDT 2020
Author: Qiu Chaofan
Date: 2020-05-13T14:05:47+08:00
New Revision: e9753822b5ad5e43e79f5970b5e86ceb329f8497
URL: https://github.com/llvm/llvm-project/commit/e9753822b5ad5e43e79f5970b5e86ceb329f8497
DIFF: https://github.com/llvm/llvm-project/commit/e9753822b5ad5e43e79f5970b5e86ceb329f8497.diff
LOG: [PowerPC] Respect SDNodeFlags in lowering SELECT_CC
Legalizer should respect both command-line options or SDNode-level
fast-math flags.
Also, this patch propagates other flags during custom simplifying.
Reviewed By: steven.zhang
Differential Revision: https://reviews.llvm.org/D79074
Added:
Modified:
llvm/lib/Target/PowerPC/PPCISelLowering.cpp
llvm/test/CodeGen/PowerPC/scalar-equal.ll
llvm/test/CodeGen/PowerPC/scalar-min-max.ll
llvm/test/CodeGen/PowerPC/scalar_cmp.ll
Removed:
################################################################################
diff --git a/llvm/lib/Target/PowerPC/PPCISelLowering.cpp b/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
index 9d49b3a1e069..bbc08c0bbd8f 100644
--- a/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
+++ b/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
@@ -7992,6 +7992,7 @@ SDValue PPCTargetLowering::LowerSELECT_CC(SDValue Op, SelectionDAG &DAG) const {
SDValue LHS = Op.getOperand(0), RHS = Op.getOperand(1);
SDValue TV = Op.getOperand(2), FV = Op.getOperand(3);
SDLoc dl(Op);
+ SDNodeFlags Flags = Op.getNode()->getFlags();
// We have xsmaxcdp/xsmincdp which are OK to emit even in the
// presence of infinities.
@@ -8012,15 +8013,10 @@ SDValue PPCTargetLowering::LowerSELECT_CC(SDValue Op, SelectionDAG &DAG) const {
// general, fsel-based lowering of select is a finite-math-only optimization.
// For more information, see section F.3 of the 2.06 ISA specification.
// With ISA 3.0
- if (!DAG.getTarget().Options.NoInfsFPMath ||
- !DAG.getTarget().Options.NoNaNsFPMath)
+ if ((!DAG.getTarget().Options.NoInfsFPMath && !Flags.hasNoInfs()) ||
+ (!DAG.getTarget().Options.NoNaNsFPMath && !Flags.hasNoNaNs()))
return Op;
- // TODO: Propagate flags from the select rather than global settings.
- SDNodeFlags Flags;
- Flags.setNoInfs(true);
- Flags.setNoNaNs(true);
-
// If the RHS of the comparison is a 0.0, we don't need to do the
// subtraction at all.
SDValue Sel1;
diff --git a/llvm/test/CodeGen/PowerPC/scalar-equal.ll b/llvm/test/CodeGen/PowerPC/scalar-equal.ll
index 3cd570b712dd..41dabb9c8e16 100644
--- a/llvm/test/CodeGen/PowerPC/scalar-equal.ll
+++ b/llvm/test/CodeGen/PowerPC/scalar-equal.ll
@@ -51,6 +51,44 @@ define double @testoeq(double %a, double %b, double %c, double %d) {
; NO-FAST-P8-NEXT: # %bb.1: # %entry
; NO-FAST-P8-NEXT: fmr f1, f4
; NO-FAST-P8-NEXT: blr
+entry:
+ %cmp = fcmp oeq double %a, %b
+ %cond = select i1 %cmp, double %c, double %d
+ ret double %cond
+}
+
+define double @testoeq_fast(double %a, double %b, double %c, double %d) {
+; FAST-P8-LABEL: testoeq_fast:
+; FAST-P8: # %bb.0: # %entry
+; FAST-P8-NEXT: xssubdp f0, f1, f2
+; FAST-P8-NEXT: xsnegdp f1, f0
+; FAST-P8-NEXT: fsel f0, f0, f3, f4
+; FAST-P8-NEXT: fsel f1, f1, f0, f4
+; FAST-P8-NEXT: blr
+;
+; FAST-P9-LABEL: testoeq_fast:
+; FAST-P9: # %bb.0: # %entry
+; FAST-P9-NEXT: xssubdp f0, f1, f2
+; FAST-P9-NEXT: fsel f1, f0, f3, f4
+; FAST-P9-NEXT: xsnegdp f0, f0
+; FAST-P9-NEXT: fsel f1, f0, f1, f4
+; FAST-P9-NEXT: blr
+;
+; NO-FAST-P9-LABEL: testoeq_fast:
+; NO-FAST-P9: # %bb.0: # %entry
+; NO-FAST-P9-NEXT: xssubdp f0, f1, f2
+; NO-FAST-P9-NEXT: fsel f1, f0, f3, f4
+; NO-FAST-P9-NEXT: xsnegdp f0, f0
+; NO-FAST-P9-NEXT: fsel f1, f0, f1, f4
+; NO-FAST-P9-NEXT: blr
+;
+; NO-FAST-P8-LABEL: testoeq_fast:
+; NO-FAST-P8: # %bb.0: # %entry
+; NO-FAST-P8-NEXT: xssubdp f0, f1, f2
+; NO-FAST-P8-NEXT: xsnegdp f1, f0
+; NO-FAST-P8-NEXT: fsel f0, f0, f3, f4
+; NO-FAST-P8-NEXT: fsel f1, f1, f0, f4
+; NO-FAST-P8-NEXT: blr
entry:
%cmp = fcmp fast oeq double %a, %b
%cond = select fast i1 %cmp, double %c, double %d
diff --git a/llvm/test/CodeGen/PowerPC/scalar-min-max.ll b/llvm/test/CodeGen/PowerPC/scalar-min-max.ll
index a7e95228b74d..d5e4a59277cc 100644
--- a/llvm/test/CodeGen/PowerPC/scalar-min-max.ll
+++ b/llvm/test/CodeGen/PowerPC/scalar-min-max.ll
@@ -122,10 +122,8 @@ define dso_local float @testfmax_fast(float %a, float %b) local_unnamed_addr {
;
; NO-FAST-P8-LABEL: testfmax_fast:
; NO-FAST-P8: # %bb.0: # %entry
-; NO-FAST-P8-NEXT: fcmpu cr0, f1, f2
-; NO-FAST-P8-NEXT: bgtlr cr0
-; NO-FAST-P8-NEXT: # %bb.1: # %entry
-; NO-FAST-P8-NEXT: fmr f1, f2
+; NO-FAST-P8-NEXT: xssubsp f0, f2, f1
+; NO-FAST-P8-NEXT: fsel f1, f0, f2, f1
; NO-FAST-P8-NEXT: blr
entry:
%cmp = fcmp fast ogt float %a, %b
@@ -145,10 +143,8 @@ define dso_local double @testdmax_fast(double %a, double %b) local_unnamed_addr
;
; NO-FAST-P8-LABEL: testdmax_fast:
; NO-FAST-P8: # %bb.0: # %entry
-; NO-FAST-P8-NEXT: xscmpudp cr0, f1, f2
-; NO-FAST-P8-NEXT: bgtlr cr0
-; NO-FAST-P8-NEXT: # %bb.1: # %entry
-; NO-FAST-P8-NEXT: fmr f1, f2
+; NO-FAST-P8-NEXT: xssubdp f0, f2, f1
+; NO-FAST-P8-NEXT: fsel f1, f0, f2, f1
; NO-FAST-P8-NEXT: blr
entry:
%cmp = fcmp fast ogt double %a, %b
@@ -168,10 +164,8 @@ define dso_local float @testfmin_fast(float %a, float %b) local_unnamed_addr {
;
; NO-FAST-P8-LABEL: testfmin_fast:
; NO-FAST-P8: # %bb.0: # %entry
-; NO-FAST-P8-NEXT: fcmpu cr0, f1, f2
-; NO-FAST-P8-NEXT: bltlr cr0
-; NO-FAST-P8-NEXT: # %bb.1: # %entry
-; NO-FAST-P8-NEXT: fmr f1, f2
+; NO-FAST-P8-NEXT: xssubsp f0, f1, f2
+; NO-FAST-P8-NEXT: fsel f1, f0, f2, f1
; NO-FAST-P8-NEXT: blr
entry:
%cmp = fcmp fast olt float %a, %b
@@ -191,10 +185,8 @@ define dso_local double @testdmin_fast(double %a, double %b) local_unnamed_addr
;
; NO-FAST-P8-LABEL: testdmin_fast:
; NO-FAST-P8: # %bb.0: # %entry
-; NO-FAST-P8-NEXT: xscmpudp cr0, f1, f2
-; NO-FAST-P8-NEXT: bltlr cr0
-; NO-FAST-P8-NEXT: # %bb.1: # %entry
-; NO-FAST-P8-NEXT: fmr f1, f2
+; NO-FAST-P8-NEXT: xssubdp f0, f1, f2
+; NO-FAST-P8-NEXT: fsel f1, f0, f2, f1
; NO-FAST-P8-NEXT: blr
entry:
%cmp = fcmp fast olt double %a, %b
diff --git a/llvm/test/CodeGen/PowerPC/scalar_cmp.ll b/llvm/test/CodeGen/PowerPC/scalar_cmp.ll
index 365e0e4e55d6..1fe0ec758587 100644
--- a/llvm/test/CodeGen/PowerPC/scalar_cmp.ll
+++ b/llvm/test/CodeGen/PowerPC/scalar_cmp.ll
@@ -118,21 +118,18 @@ define float @select_fast_oeq_float(float %a, float %b, float %c, float %d) {
;
; NO-FAST-P8-LABEL: select_fast_oeq_float:
; NO-FAST-P8: # %bb.0: # %entry
-; NO-FAST-P8-NEXT: fcmpu cr0, f1, f2
-; NO-FAST-P8-NEXT: fmr f1, f3
-; NO-FAST-P8-NEXT: beqlr cr0
-; NO-FAST-P8-NEXT: # %bb.1: # %entry
-; NO-FAST-P8-NEXT: fmr f1, f4
+; NO-FAST-P8-NEXT: xssubsp f0, f1, f2
+; NO-FAST-P8-NEXT: xssubsp f1, f2, f1
+; NO-FAST-P8-NEXT: fsel f0, f0, f3, f4
+; NO-FAST-P8-NEXT: fsel f1, f1, f0, f4
; NO-FAST-P8-NEXT: blr
;
; NO-FAST-P9-LABEL: select_fast_oeq_float:
; NO-FAST-P9: # %bb.0: # %entry
-; NO-FAST-P9-NEXT: fcmpu cr0, f1, f2
-; NO-FAST-P9-NEXT: beq cr0, .LBB2_2
-; NO-FAST-P9-NEXT: # %bb.1: # %entry
-; NO-FAST-P9-NEXT: fmr f3, f4
-; NO-FAST-P9-NEXT: .LBB2_2: # %entry
-; NO-FAST-P9-NEXT: fmr f1, f3
+; NO-FAST-P9-NEXT: xssubsp f0, f2, f1
+; NO-FAST-P9-NEXT: xssubsp f1, f1, f2
+; NO-FAST-P9-NEXT: fsel f1, f1, f3, f4
+; NO-FAST-P9-NEXT: fsel f1, f0, f1, f4
; NO-FAST-P9-NEXT: blr
entry:
%cmp = fcmp fast oeq float %a, %b
@@ -159,21 +156,18 @@ define double @select_fast_oeq_double(double %a, double %b, double %c, double %d
;
; NO-FAST-P8-LABEL: select_fast_oeq_double:
; NO-FAST-P8: # %bb.0: # %entry
-; NO-FAST-P8-NEXT: xscmpudp cr0, f1, f2
-; NO-FAST-P8-NEXT: fmr f1, f3
-; NO-FAST-P8-NEXT: beqlr cr0
-; NO-FAST-P8-NEXT: # %bb.1: # %entry
-; NO-FAST-P8-NEXT: fmr f1, f4
+; NO-FAST-P8-NEXT: xssubdp f0, f1, f2
+; NO-FAST-P8-NEXT: xsnegdp f1, f0
+; NO-FAST-P8-NEXT: fsel f0, f0, f3, f4
+; NO-FAST-P8-NEXT: fsel f1, f1, f0, f4
; NO-FAST-P8-NEXT: blr
;
; NO-FAST-P9-LABEL: select_fast_oeq_double:
; NO-FAST-P9: # %bb.0: # %entry
-; NO-FAST-P9-NEXT: xscmpudp cr0, f1, f2
-; NO-FAST-P9-NEXT: beq cr0, .LBB3_2
-; NO-FAST-P9-NEXT: # %bb.1: # %entry
-; NO-FAST-P9-NEXT: fmr f3, f4
-; NO-FAST-P9-NEXT: .LBB3_2: # %entry
-; NO-FAST-P9-NEXT: fmr f1, f3
+; NO-FAST-P9-NEXT: xssubdp f0, f1, f2
+; NO-FAST-P9-NEXT: fsel f1, f0, f3, f4
+; NO-FAST-P9-NEXT: xsnegdp f0, f0
+; NO-FAST-P9-NEXT: fsel f1, f0, f1, f4
; NO-FAST-P9-NEXT: blr
entry:
%cmp = fcmp fast oeq double %a, %b
@@ -288,21 +282,18 @@ define float @select_fast_one_float(float %a, float %b, float %c, float %d) {
;
; NO-FAST-P8-LABEL: select_fast_one_float:
; NO-FAST-P8: # %bb.0: # %entry
-; NO-FAST-P8-NEXT: fcmpu cr0, f1, f2
-; NO-FAST-P8-NEXT: fmr f1, f3
-; NO-FAST-P8-NEXT: bnelr cr0
-; NO-FAST-P8-NEXT: # %bb.1: # %entry
-; NO-FAST-P8-NEXT: fmr f1, f4
+; NO-FAST-P8-NEXT: xssubsp f0, f1, f2
+; NO-FAST-P8-NEXT: xssubsp f1, f2, f1
+; NO-FAST-P8-NEXT: fsel f0, f0, f4, f3
+; NO-FAST-P8-NEXT: fsel f1, f1, f0, f3
; NO-FAST-P8-NEXT: blr
;
; NO-FAST-P9-LABEL: select_fast_one_float:
; NO-FAST-P9: # %bb.0: # %entry
-; NO-FAST-P9-NEXT: fcmpu cr0, f1, f2
-; NO-FAST-P9-NEXT: bne cr0, .LBB6_2
-; NO-FAST-P9-NEXT: # %bb.1: # %entry
-; NO-FAST-P9-NEXT: fmr f3, f4
-; NO-FAST-P9-NEXT: .LBB6_2: # %entry
-; NO-FAST-P9-NEXT: fmr f1, f3
+; NO-FAST-P9-NEXT: xssubsp f0, f2, f1
+; NO-FAST-P9-NEXT: xssubsp f1, f1, f2
+; NO-FAST-P9-NEXT: fsel f1, f1, f4, f3
+; NO-FAST-P9-NEXT: fsel f1, f0, f1, f3
; NO-FAST-P9-NEXT: blr
entry:
%cmp = fcmp fast one float %a, %b
@@ -329,21 +320,18 @@ define double @select_fast_one_double(double %a, double %b, double %c, double %d
;
; NO-FAST-P8-LABEL: select_fast_one_double:
; NO-FAST-P8: # %bb.0: # %entry
-; NO-FAST-P8-NEXT: xscmpudp cr0, f1, f2
-; NO-FAST-P8-NEXT: fmr f1, f3
-; NO-FAST-P8-NEXT: bnelr cr0
-; NO-FAST-P8-NEXT: # %bb.1: # %entry
-; NO-FAST-P8-NEXT: fmr f1, f4
+; NO-FAST-P8-NEXT: xssubdp f0, f1, f2
+; NO-FAST-P8-NEXT: xsnegdp f1, f0
+; NO-FAST-P8-NEXT: fsel f0, f0, f4, f3
+; NO-FAST-P8-NEXT: fsel f1, f1, f0, f3
; NO-FAST-P8-NEXT: blr
;
; NO-FAST-P9-LABEL: select_fast_one_double:
; NO-FAST-P9: # %bb.0: # %entry
-; NO-FAST-P9-NEXT: xscmpudp cr0, f1, f2
-; NO-FAST-P9-NEXT: bne cr0, .LBB7_2
-; NO-FAST-P9-NEXT: # %bb.1: # %entry
-; NO-FAST-P9-NEXT: fmr f3, f4
-; NO-FAST-P9-NEXT: .LBB7_2: # %entry
-; NO-FAST-P9-NEXT: fmr f1, f3
+; NO-FAST-P9-NEXT: xssubdp f0, f1, f2
+; NO-FAST-P9-NEXT: fsel f1, f0, f4, f3
+; NO-FAST-P9-NEXT: xsnegdp f0, f0
+; NO-FAST-P9-NEXT: fsel f1, f0, f1, f3
; NO-FAST-P9-NEXT: blr
entry:
%cmp = fcmp fast one double %a, %b
@@ -446,21 +434,14 @@ define float @select_fast_oge_float(float %a, float %b, float %c, float %d) {
;
; NO-FAST-P8-LABEL: select_fast_oge_float:
; NO-FAST-P8: # %bb.0: # %entry
-; NO-FAST-P8-NEXT: fcmpu cr0, f1, f2
-; NO-FAST-P8-NEXT: fmr f1, f3
-; NO-FAST-P8-NEXT: bgelr cr0
-; NO-FAST-P8-NEXT: # %bb.1: # %entry
-; NO-FAST-P8-NEXT: fmr f1, f4
+; NO-FAST-P8-NEXT: xssubsp f0, f1, f2
+; NO-FAST-P8-NEXT: fsel f1, f0, f3, f4
; NO-FAST-P8-NEXT: blr
;
; NO-FAST-P9-LABEL: select_fast_oge_float:
; NO-FAST-P9: # %bb.0: # %entry
-; NO-FAST-P9-NEXT: fcmpu cr0, f1, f2
-; NO-FAST-P9-NEXT: bge cr0, .LBB10_2
-; NO-FAST-P9-NEXT: # %bb.1: # %entry
-; NO-FAST-P9-NEXT: fmr f3, f4
-; NO-FAST-P9-NEXT: .LBB10_2: # %entry
-; NO-FAST-P9-NEXT: fmr f1, f3
+; NO-FAST-P9-NEXT: xssubsp f0, f1, f2
+; NO-FAST-P9-NEXT: fsel f1, f0, f3, f4
; NO-FAST-P9-NEXT: blr
entry:
%cmp = fcmp fast oge float %a, %b
@@ -483,21 +464,14 @@ define double @select_fast_oge_double(double %a, double %b, double %c, double %d
;
; NO-FAST-P8-LABEL: select_fast_oge_double:
; NO-FAST-P8: # %bb.0: # %entry
-; NO-FAST-P8-NEXT: xscmpudp cr0, f1, f2
-; NO-FAST-P8-NEXT: fmr f1, f3
-; NO-FAST-P8-NEXT: bgelr cr0
-; NO-FAST-P8-NEXT: # %bb.1: # %entry
-; NO-FAST-P8-NEXT: fmr f1, f4
+; NO-FAST-P8-NEXT: xssubdp f0, f1, f2
+; NO-FAST-P8-NEXT: fsel f1, f0, f3, f4
; NO-FAST-P8-NEXT: blr
;
; NO-FAST-P9-LABEL: select_fast_oge_double:
; NO-FAST-P9: # %bb.0: # %entry
-; NO-FAST-P9-NEXT: xscmpudp cr0, f1, f2
-; NO-FAST-P9-NEXT: bge cr0, .LBB11_2
-; NO-FAST-P9-NEXT: # %bb.1: # %entry
-; NO-FAST-P9-NEXT: fmr f3, f4
-; NO-FAST-P9-NEXT: .LBB11_2: # %entry
-; NO-FAST-P9-NEXT: fmr f1, f3
+; NO-FAST-P9-NEXT: xssubdp f0, f1, f2
+; NO-FAST-P9-NEXT: fsel f1, f0, f3, f4
; NO-FAST-P9-NEXT: blr
entry:
%cmp = fcmp fast oge double %a, %b
@@ -596,21 +570,14 @@ define float @select_fast_olt_float(float %a, float %b, float %c, float %d) {
;
; NO-FAST-P8-LABEL: select_fast_olt_float:
; NO-FAST-P8: # %bb.0: # %entry
-; NO-FAST-P8-NEXT: fcmpu cr0, f1, f2
-; NO-FAST-P8-NEXT: fmr f1, f3
-; NO-FAST-P8-NEXT: bltlr cr0
-; NO-FAST-P8-NEXT: # %bb.1: # %entry
-; NO-FAST-P8-NEXT: fmr f1, f4
+; NO-FAST-P8-NEXT: xssubsp f0, f1, f2
+; NO-FAST-P8-NEXT: fsel f1, f0, f4, f3
; NO-FAST-P8-NEXT: blr
;
; NO-FAST-P9-LABEL: select_fast_olt_float:
; NO-FAST-P9: # %bb.0: # %entry
-; NO-FAST-P9-NEXT: fcmpu cr0, f1, f2
-; NO-FAST-P9-NEXT: blt cr0, .LBB14_2
-; NO-FAST-P9-NEXT: # %bb.1: # %entry
-; NO-FAST-P9-NEXT: fmr f3, f4
-; NO-FAST-P9-NEXT: .LBB14_2: # %entry
-; NO-FAST-P9-NEXT: fmr f1, f3
+; NO-FAST-P9-NEXT: xssubsp f0, f1, f2
+; NO-FAST-P9-NEXT: fsel f1, f0, f4, f3
; NO-FAST-P9-NEXT: blr
entry:
%cmp = fcmp fast olt float %a, %b
@@ -633,21 +600,14 @@ define double @select_fast_olt_double(double %a, double %b, double %c, double %d
;
; NO-FAST-P8-LABEL: select_fast_olt_double:
; NO-FAST-P8: # %bb.0: # %entry
-; NO-FAST-P8-NEXT: xscmpudp cr0, f1, f2
-; NO-FAST-P8-NEXT: fmr f1, f3
-; NO-FAST-P8-NEXT: bltlr cr0
-; NO-FAST-P8-NEXT: # %bb.1: # %entry
-; NO-FAST-P8-NEXT: fmr f1, f4
+; NO-FAST-P8-NEXT: xssubdp f0, f1, f2
+; NO-FAST-P8-NEXT: fsel f1, f0, f4, f3
; NO-FAST-P8-NEXT: blr
;
; NO-FAST-P9-LABEL: select_fast_olt_double:
; NO-FAST-P9: # %bb.0: # %entry
-; NO-FAST-P9-NEXT: xscmpudp cr0, f1, f2
-; NO-FAST-P9-NEXT: blt cr0, .LBB15_2
-; NO-FAST-P9-NEXT: # %bb.1: # %entry
-; NO-FAST-P9-NEXT: fmr f3, f4
-; NO-FAST-P9-NEXT: .LBB15_2: # %entry
-; NO-FAST-P9-NEXT: fmr f1, f3
+; NO-FAST-P9-NEXT: xssubdp f0, f1, f2
+; NO-FAST-P9-NEXT: fsel f1, f0, f4, f3
; NO-FAST-P9-NEXT: blr
entry:
%cmp = fcmp fast olt double %a, %b
@@ -746,21 +706,14 @@ define float @select_fast_ogt_float(float %a, float %b, float %c, float %d) {
;
; NO-FAST-P8-LABEL: select_fast_ogt_float:
; NO-FAST-P8: # %bb.0: # %entry
-; NO-FAST-P8-NEXT: fcmpu cr0, f1, f2
-; NO-FAST-P8-NEXT: fmr f1, f3
-; NO-FAST-P8-NEXT: bgtlr cr0
-; NO-FAST-P8-NEXT: # %bb.1: # %entry
-; NO-FAST-P8-NEXT: fmr f1, f4
+; NO-FAST-P8-NEXT: xssubsp f0, f2, f1
+; NO-FAST-P8-NEXT: fsel f1, f0, f4, f3
; NO-FAST-P8-NEXT: blr
;
; NO-FAST-P9-LABEL: select_fast_ogt_float:
; NO-FAST-P9: # %bb.0: # %entry
-; NO-FAST-P9-NEXT: fcmpu cr0, f1, f2
-; NO-FAST-P9-NEXT: bgt cr0, .LBB18_2
-; NO-FAST-P9-NEXT: # %bb.1: # %entry
-; NO-FAST-P9-NEXT: fmr f3, f4
-; NO-FAST-P9-NEXT: .LBB18_2: # %entry
-; NO-FAST-P9-NEXT: fmr f1, f3
+; NO-FAST-P9-NEXT: xssubsp f0, f2, f1
+; NO-FAST-P9-NEXT: fsel f1, f0, f4, f3
; NO-FAST-P9-NEXT: blr
entry:
%cmp = fcmp fast ogt float %a, %b
@@ -783,21 +736,14 @@ define double @select_fast_ogt_double(double %a, double %b, double %c, double %d
;
; NO-FAST-P8-LABEL: select_fast_ogt_double:
; NO-FAST-P8: # %bb.0: # %entry
-; NO-FAST-P8-NEXT: xscmpudp cr0, f1, f2
-; NO-FAST-P8-NEXT: fmr f1, f3
-; NO-FAST-P8-NEXT: bgtlr cr0
-; NO-FAST-P8-NEXT: # %bb.1: # %entry
-; NO-FAST-P8-NEXT: fmr f1, f4
+; NO-FAST-P8-NEXT: xssubdp f0, f2, f1
+; NO-FAST-P8-NEXT: fsel f1, f0, f4, f3
; NO-FAST-P8-NEXT: blr
;
; NO-FAST-P9-LABEL: select_fast_ogt_double:
; NO-FAST-P9: # %bb.0: # %entry
-; NO-FAST-P9-NEXT: xscmpudp cr0, f1, f2
-; NO-FAST-P9-NEXT: bgt cr0, .LBB19_2
-; NO-FAST-P9-NEXT: # %bb.1: # %entry
-; NO-FAST-P9-NEXT: fmr f3, f4
-; NO-FAST-P9-NEXT: .LBB19_2: # %entry
-; NO-FAST-P9-NEXT: fmr f1, f3
+; NO-FAST-P9-NEXT: xssubdp f0, f2, f1
+; NO-FAST-P9-NEXT: fsel f1, f0, f4, f3
; NO-FAST-P9-NEXT: blr
entry:
%cmp = fcmp fast ogt double %a, %b
@@ -900,21 +846,14 @@ define float @select_fast_ole_float(float %a, float %b, float %c, float %d) {
;
; NO-FAST-P8-LABEL: select_fast_ole_float:
; NO-FAST-P8: # %bb.0: # %entry
-; NO-FAST-P8-NEXT: fcmpu cr0, f1, f2
-; NO-FAST-P8-NEXT: fmr f1, f3
-; NO-FAST-P8-NEXT: blelr cr0
-; NO-FAST-P8-NEXT: # %bb.1: # %entry
-; NO-FAST-P8-NEXT: fmr f1, f4
+; NO-FAST-P8-NEXT: xssubsp f0, f2, f1
+; NO-FAST-P8-NEXT: fsel f1, f0, f3, f4
; NO-FAST-P8-NEXT: blr
;
; NO-FAST-P9-LABEL: select_fast_ole_float:
; NO-FAST-P9: # %bb.0: # %entry
-; NO-FAST-P9-NEXT: fcmpu cr0, f1, f2
-; NO-FAST-P9-NEXT: ble cr0, .LBB22_2
-; NO-FAST-P9-NEXT: # %bb.1: # %entry
-; NO-FAST-P9-NEXT: fmr f3, f4
-; NO-FAST-P9-NEXT: .LBB22_2: # %entry
-; NO-FAST-P9-NEXT: fmr f1, f3
+; NO-FAST-P9-NEXT: xssubsp f0, f2, f1
+; NO-FAST-P9-NEXT: fsel f1, f0, f3, f4
; NO-FAST-P9-NEXT: blr
entry:
%cmp = fcmp fast ole float %a, %b
@@ -937,21 +876,14 @@ define double @select_fast_ole_double(double %a, double %b, double %c, double %d
;
; NO-FAST-P8-LABEL: select_fast_ole_double:
; NO-FAST-P8: # %bb.0: # %entry
-; NO-FAST-P8-NEXT: xscmpudp cr0, f1, f2
-; NO-FAST-P8-NEXT: fmr f1, f3
-; NO-FAST-P8-NEXT: blelr cr0
-; NO-FAST-P8-NEXT: # %bb.1: # %entry
-; NO-FAST-P8-NEXT: fmr f1, f4
+; NO-FAST-P8-NEXT: xssubdp f0, f2, f1
+; NO-FAST-P8-NEXT: fsel f1, f0, f3, f4
; NO-FAST-P8-NEXT: blr
;
; NO-FAST-P9-LABEL: select_fast_ole_double:
; NO-FAST-P9: # %bb.0: # %entry
-; NO-FAST-P9-NEXT: xscmpudp cr0, f1, f2
-; NO-FAST-P9-NEXT: ble cr0, .LBB23_2
-; NO-FAST-P9-NEXT: # %bb.1: # %entry
-; NO-FAST-P9-NEXT: fmr f3, f4
-; NO-FAST-P9-NEXT: .LBB23_2: # %entry
-; NO-FAST-P9-NEXT: fmr f1, f3
+; NO-FAST-P9-NEXT: xssubdp f0, f2, f1
+; NO-FAST-P9-NEXT: fsel f1, f0, f3, f4
; NO-FAST-P9-NEXT: blr
entry:
%cmp = fcmp fast ole double %a, %b
More information about the llvm-commits
mailing list