llvm-6502/test/CodeGen/ARM/vicmp.ll
Stephen Lin 8b2b8a1835 Mass update to CodeGen tests to use CHECK-LABEL for labels corresponding to function definitions for more informative error messages. No functionality change and all updated tests passed locally.
This update was done with the following bash script:

  find test/CodeGen -name "*.ll" | \
  while read NAME; do
    echo "$NAME"
    if ! grep -q "^; *RUN: *llc.*debug" $NAME; then
      TEMP=`mktemp -t temp`
      cp $NAME $TEMP
      sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \
      while read FUNC; do
        sed -i '' "s/;\(.*\)\([A-Za-z0-9_-]*\):\( *\)$FUNC: *\$/;\1\2-LABEL:\3$FUNC:/g" $TEMP
      done
      sed -i '' "s/;\(.*\)-LABEL-LABEL:/;\1-LABEL:/" $TEMP
      sed -i '' "s/;\(.*\)-NEXT-LABEL:/;\1-NEXT:/" $TEMP
      sed -i '' "s/;\(.*\)-NOT-LABEL:/;\1-NOT:/" $TEMP
      sed -i '' "s/;\(.*\)-DAG-LABEL:/;\1-DAG:/" $TEMP
      mv $TEMP $NAME
    fi
  done


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@186280 91177308-0d34-0410-b5e6-96231b3b80d8
2013-07-14 06:24:09 +00:00

114 lines
3.2 KiB
LLVM

; RUN: llc < %s -march=arm -mattr=+neon | FileCheck %s
; This tests icmp operations that do not map directly to NEON instructions.
; Not-equal (ne) operations are implemented by VCEQ/VMVN. Less-than (lt/ult)
; and less-than-or-equal (le/ule) are implemented by swapping the arguments
; to VCGT and VCGE. Test all the operand types for not-equal but only sample
; the other operations.
define <8 x i8> @vcnei8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
;CHECK-LABEL: vcnei8:
;CHECK: vceq.i8
;CHECK-NEXT: vmvn
%tmp1 = load <8 x i8>* %A
%tmp2 = load <8 x i8>* %B
%tmp3 = icmp ne <8 x i8> %tmp1, %tmp2
%tmp4 = sext <8 x i1> %tmp3 to <8 x i8>
ret <8 x i8> %tmp4
}
define <4 x i16> @vcnei16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
;CHECK-LABEL: vcnei16:
;CHECK: vceq.i16
;CHECK-NEXT: vmvn
%tmp1 = load <4 x i16>* %A
%tmp2 = load <4 x i16>* %B
%tmp3 = icmp ne <4 x i16> %tmp1, %tmp2
%tmp4 = sext <4 x i1> %tmp3 to <4 x i16>
ret <4 x i16> %tmp4
}
define <2 x i32> @vcnei32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
;CHECK-LABEL: vcnei32:
;CHECK: vceq.i32
;CHECK-NEXT: vmvn
%tmp1 = load <2 x i32>* %A
%tmp2 = load <2 x i32>* %B
%tmp3 = icmp ne <2 x i32> %tmp1, %tmp2
%tmp4 = sext <2 x i1> %tmp3 to <2 x i32>
ret <2 x i32> %tmp4
}
define <16 x i8> @vcneQi8(<16 x i8>* %A, <16 x i8>* %B) nounwind {
;CHECK-LABEL: vcneQi8:
;CHECK: vceq.i8
;CHECK-NEXT: vmvn
%tmp1 = load <16 x i8>* %A
%tmp2 = load <16 x i8>* %B
%tmp3 = icmp ne <16 x i8> %tmp1, %tmp2
%tmp4 = sext <16 x i1> %tmp3 to <16 x i8>
ret <16 x i8> %tmp4
}
define <8 x i16> @vcneQi16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
;CHECK-LABEL: vcneQi16:
;CHECK: vceq.i16
;CHECK-NEXT: vmvn
%tmp1 = load <8 x i16>* %A
%tmp2 = load <8 x i16>* %B
%tmp3 = icmp ne <8 x i16> %tmp1, %tmp2
%tmp4 = sext <8 x i1> %tmp3 to <8 x i16>
ret <8 x i16> %tmp4
}
define <4 x i32> @vcneQi32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
;CHECK-LABEL: vcneQi32:
;CHECK: vceq.i32
;CHECK-NEXT: vmvn
%tmp1 = load <4 x i32>* %A
%tmp2 = load <4 x i32>* %B
%tmp3 = icmp ne <4 x i32> %tmp1, %tmp2
%tmp4 = sext <4 x i1> %tmp3 to <4 x i32>
ret <4 x i32> %tmp4
}
define <16 x i8> @vcltQs8(<16 x i8>* %A, <16 x i8>* %B) nounwind {
;CHECK-LABEL: vcltQs8:
;CHECK: vcgt.s8
%tmp1 = load <16 x i8>* %A
%tmp2 = load <16 x i8>* %B
%tmp3 = icmp slt <16 x i8> %tmp1, %tmp2
%tmp4 = sext <16 x i1> %tmp3 to <16 x i8>
ret <16 x i8> %tmp4
}
define <4 x i16> @vcles16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
;CHECK-LABEL: vcles16:
;CHECK: vcge.s16
%tmp1 = load <4 x i16>* %A
%tmp2 = load <4 x i16>* %B
%tmp3 = icmp sle <4 x i16> %tmp1, %tmp2
%tmp4 = sext <4 x i1> %tmp3 to <4 x i16>
ret <4 x i16> %tmp4
}
define <4 x i16> @vcltu16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
;CHECK-LABEL: vcltu16:
;CHECK: vcgt.u16
%tmp1 = load <4 x i16>* %A
%tmp2 = load <4 x i16>* %B
%tmp3 = icmp ult <4 x i16> %tmp1, %tmp2
%tmp4 = sext <4 x i1> %tmp3 to <4 x i16>
ret <4 x i16> %tmp4
}
define <4 x i32> @vcleQu32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
;CHECK-LABEL: vcleQu32:
;CHECK: vcge.u32
%tmp1 = load <4 x i32>* %A
%tmp2 = load <4 x i32>* %B
%tmp3 = icmp ule <4 x i32> %tmp1, %tmp2
%tmp4 = sext <4 x i1> %tmp3 to <4 x i32>
ret <4 x i32> %tmp4
}