Hal Finkel 87c1e42be7 Fix PPC optimizeCompareInstr swapped-sub argument handling
When matching a compare with a subtract where the arguments of the compare are
swapped w.r.t. the arguments of the subtract, we need to negate the predicates
(or CR bit indices) of the users. This, however, is not the same as inverting
the predicate (negating LT -> GT, but inverting LT -> GE, for example). The ARM
backend seems to do this correctly, but when I adapted the code for the PPC
backend, I introduced an error in this logic.

Comparison optimization is now enabled again by default.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179899 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-19 22:08:38 +00:00

144 lines
3.6 KiB
LLVM

; RUN: llc < %s -mtriple=powerpc64-unknown-linux-gnu -mcpu=a2 -disable-ppc-cmp-opt=0 | FileCheck %s
target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-f128:128:128-v128:128:128-n32:64"
target triple = "powerpc64-unknown-linux-gnu"
define signext i32 @foo(i32 signext %a, i32 signext %b, i32* nocapture %c) #0 {
entry:
%sub = sub nsw i32 %a, %b
store i32 %sub, i32* %c, align 4, !tbaa !0
%cmp = icmp sgt i32 %a, %b
%cond = select i1 %cmp, i32 %a, i32 %b
ret i32 %cond
; CHECK: @foo
; CHECK-NOT: subf.
}
define signext i32 @foo2(i32 signext %a, i32 signext %b, i32* nocapture %c) #0 {
entry:
%shl = shl i32 %a, %b
store i32 %shl, i32* %c, align 4, !tbaa !0
%cmp = icmp sgt i32 %shl, 0
%conv = zext i1 %cmp to i32
ret i32 %conv
; CHECK: @foo2
; CHECK-NOT: slw.
}
define i64 @fool(i64 %a, i64 %b, i64* nocapture %c) #0 {
entry:
%sub = sub nsw i64 %a, %b
store i64 %sub, i64* %c, align 8, !tbaa !3
%cmp = icmp sgt i64 %a, %b
%cond = select i1 %cmp, i64 %a, i64 %b
ret i64 %cond
; CHECK: @fool
; CHECK: subf. [[REG:[0-9]+]], 4, 3
; CHECK: isel 3, 3, 4, 1
; CHECK: std [[REG]], 0(5)
}
define i64 @foolb(i64 %a, i64 %b, i64* nocapture %c) #0 {
entry:
%sub = sub nsw i64 %a, %b
store i64 %sub, i64* %c, align 8, !tbaa !3
%cmp = icmp sle i64 %a, %b
%cond = select i1 %cmp, i64 %a, i64 %b
ret i64 %cond
; CHECK: @foolb
; CHECK: subf. [[REG:[0-9]+]], 4, 3
; CHECK: isel 3, 4, 3, 1
; CHECK: std [[REG]], 0(5)
}
define i64 @foolc(i64 %a, i64 %b, i64* nocapture %c) #0 {
entry:
%sub = sub nsw i64 %b, %a
store i64 %sub, i64* %c, align 8, !tbaa !3
%cmp = icmp sgt i64 %a, %b
%cond = select i1 %cmp, i64 %a, i64 %b
ret i64 %cond
; CHECK: @foolc
; CHECK: subf. [[REG:[0-9]+]], 3, 4
; CHECK: isel 3, 3, 4, 0
; CHECK: std [[REG]], 0(5)
}
define i64 @foold(i64 %a, i64 %b, i64* nocapture %c) #0 {
entry:
%sub = sub nsw i64 %b, %a
store i64 %sub, i64* %c, align 8, !tbaa !3
%cmp = icmp eq i64 %a, %b
%cond = select i1 %cmp, i64 %a, i64 %b
ret i64 %cond
; CHECK: @foold
; CHECK: subf. [[REG:[0-9]+]], 3, 4
; CHECK: isel 3, 3, 4, 2
; CHECK: std [[REG]], 0(5)
}
define i64 @foold2(i64 %a, i64 %b, i64* nocapture %c) #0 {
entry:
%sub = sub nsw i64 %a, %b
store i64 %sub, i64* %c, align 8, !tbaa !3
%cmp = icmp eq i64 %a, %b
%cond = select i1 %cmp, i64 %a, i64 %b
ret i64 %cond
; CHECK: @foold2
; CHECK: subf. [[REG:[0-9]+]], 4, 3
; CHECK: isel 3, 3, 4, 2
; CHECK: std [[REG]], 0(5)
}
define i64 @foo2l(i64 %a, i64 %b, i64* nocapture %c) #0 {
entry:
%shl = shl i64 %a, %b
store i64 %shl, i64* %c, align 8, !tbaa !3
%cmp = icmp sgt i64 %shl, 0
%conv1 = zext i1 %cmp to i64
ret i64 %conv1
; CHECK: @foo2l
; CHECK: sld. 4, 3, 4
; CHECK: std 4, 0(5)
}
define double @food(double %a, double %b, double* nocapture %c) #0 {
entry:
%sub = fsub double %a, %b
store double %sub, double* %c, align 8, !tbaa !3
%cmp = fcmp ogt double %a, %b
%cond = select i1 %cmp, double %a, double %b
ret double %cond
; CHECK: @food
; CHECK: fsub. 0, 1, 2
; CHECK: stfd 0, 0(5)
}
define float @foof(float %a, float %b, float* nocapture %c) #0 {
entry:
%sub = fsub float %a, %b
store float %sub, float* %c, align 4, !tbaa !3
%cmp = fcmp ogt float %a, %b
%cond = select i1 %cmp, float %a, float %b
ret float %cond
; CHECK: @foof
; CHECK: fsubs. 0, 1, 2
; CHECK: stfs 0, 0(5)
}
!0 = metadata !{metadata !"int", metadata !1}
!1 = metadata !{metadata !"omnipotent char", metadata !2}
!2 = metadata !{metadata !"Simple C/C++ TBAA"}
!3 = metadata !{metadata !"long", metadata !1}
!4 = metadata !{metadata !"any pointer", metadata !1}