ARM: optimization for sub+abs.
This patch will optimize abs(x-y)
FROM
sub, movs, rsbmi
TO
subs, rsbmi
For abs, we will use cmp instead of movs. This is necessary because we already
have an existing peephole pass which optimizes away cmp following sub.
rdar: 11633193
llvm-svn: 158551
diff --git a/llvm/test/CodeGen/ARM/thumb2-it-block.ll b/llvm/test/CodeGen/ARM/thumb2-it-block.ll
index 28fd469..a25352c 100644
--- a/llvm/test/CodeGen/ARM/thumb2-it-block.ll
+++ b/llvm/test/CodeGen/ARM/thumb2-it-block.ll
@@ -3,10 +3,10 @@
define i32 @test(i32 %a, i32 %b) {
entry:
-; CHECK: movs.w
+; CHECK: cmp
; CHECK-NEXT: it mi
; CHECK-NEXT: rsbmi
-; CHECK-NEXT: movs.w
+; CHECK-NEXT: cmp
; CHECK-NEXT: it mi
; CHECK-NEXT: rsbmi
%cmp1 = icmp slt i32 %a, 0