summaryrefslogtreecommitdiffstats
path: root/test/CodeGen/ARM/vpadd.ll
diff options
context:
space:
mode:
Diffstat (limited to 'test/CodeGen/ARM/vpadd.ll')
-rw-r--r--test/CodeGen/ARM/vpadd.ll142
1 files changed, 129 insertions, 13 deletions
diff --git a/test/CodeGen/ARM/vpadd.ll b/test/CodeGen/ARM/vpadd.ll
index baff492..2125573 100644
--- a/test/CodeGen/ARM/vpadd.ll
+++ b/test/CodeGen/ARM/vpadd.ll
@@ -1,39 +1,155 @@
-; RUN: llvm-as < %s | llc -march=arm -mattr=+neon > %t
-; RUN: grep {vpadd\\.i8} %t | count 1
-; RUN: grep {vpadd\\.i16} %t | count 1
-; RUN: grep {vpadd\\.i32} %t | count 1
-; RUN: grep {vpadd\\.f32} %t | count 1
+; RUN: llc < %s -march=arm -mattr=+neon | FileCheck %s
define <8 x i8> @vpaddi8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
+;CHECK: vpaddi8:
+;CHECK: vpadd.i8
%tmp1 = load <8 x i8>* %A
%tmp2 = load <8 x i8>* %B
- %tmp3 = call <8 x i8> @llvm.arm.neon.vpaddi.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2)
+ %tmp3 = call <8 x i8> @llvm.arm.neon.vpadd.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2)
ret <8 x i8> %tmp3
}
define <4 x i16> @vpaddi16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
+;CHECK: vpaddi16:
+;CHECK: vpadd.i16
%tmp1 = load <4 x i16>* %A
%tmp2 = load <4 x i16>* %B
- %tmp3 = call <4 x i16> @llvm.arm.neon.vpaddi.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2)
+ %tmp3 = call <4 x i16> @llvm.arm.neon.vpadd.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2)
ret <4 x i16> %tmp3
}
define <2 x i32> @vpaddi32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
+;CHECK: vpaddi32:
+;CHECK: vpadd.i32
%tmp1 = load <2 x i32>* %A
%tmp2 = load <2 x i32>* %B
- %tmp3 = call <2 x i32> @llvm.arm.neon.vpaddi.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
+ %tmp3 = call <2 x i32> @llvm.arm.neon.vpadd.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
ret <2 x i32> %tmp3
}
define <2 x float> @vpaddf32(<2 x float>* %A, <2 x float>* %B) nounwind {
+;CHECK: vpaddf32:
+;CHECK: vpadd.f32
%tmp1 = load <2 x float>* %A
%tmp2 = load <2 x float>* %B
- %tmp3 = call <2 x float> @llvm.arm.neon.vpaddf.v2f32(<2 x float> %tmp1, <2 x float> %tmp2)
+ %tmp3 = call <2 x float> @llvm.arm.neon.vpadd.v2f32(<2 x float> %tmp1, <2 x float> %tmp2)
ret <2 x float> %tmp3
}
-declare <8 x i8> @llvm.arm.neon.vpaddi.v8i8(<8 x i8>, <8 x i8>) nounwind readnone
-declare <4 x i16> @llvm.arm.neon.vpaddi.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
-declare <2 x i32> @llvm.arm.neon.vpaddi.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
+declare <8 x i8> @llvm.arm.neon.vpadd.v8i8(<8 x i8>, <8 x i8>) nounwind readnone
+declare <4 x i16> @llvm.arm.neon.vpadd.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
+declare <2 x i32> @llvm.arm.neon.vpadd.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
-declare <2 x float> @llvm.arm.neon.vpaddf.v2f32(<2 x float>, <2 x float>) nounwind readnone
+declare <2 x float> @llvm.arm.neon.vpadd.v2f32(<2 x float>, <2 x float>) nounwind readnone
+
+define <4 x i16> @vpaddls8(<8 x i8>* %A) nounwind {
+;CHECK: vpaddls8:
+;CHECK: vpaddl.s8
+ %tmp1 = load <8 x i8>* %A
+ %tmp2 = call <4 x i16> @llvm.arm.neon.vpaddls.v4i16.v8i8(<8 x i8> %tmp1)
+ ret <4 x i16> %tmp2
+}
+
+define <2 x i32> @vpaddls16(<4 x i16>* %A) nounwind {
+;CHECK: vpaddls16:
+;CHECK: vpaddl.s16
+ %tmp1 = load <4 x i16>* %A
+ %tmp2 = call <2 x i32> @llvm.arm.neon.vpaddls.v2i32.v4i16(<4 x i16> %tmp1)
+ ret <2 x i32> %tmp2
+}
+
+define <1 x i64> @vpaddls32(<2 x i32>* %A) nounwind {
+;CHECK: vpaddls32:
+;CHECK: vpaddl.s32
+ %tmp1 = load <2 x i32>* %A
+ %tmp2 = call <1 x i64> @llvm.arm.neon.vpaddls.v1i64.v2i32(<2 x i32> %tmp1)
+ ret <1 x i64> %tmp2
+}
+
+define <4 x i16> @vpaddlu8(<8 x i8>* %A) nounwind {
+;CHECK: vpaddlu8:
+;CHECK: vpaddl.u8
+ %tmp1 = load <8 x i8>* %A
+ %tmp2 = call <4 x i16> @llvm.arm.neon.vpaddlu.v4i16.v8i8(<8 x i8> %tmp1)
+ ret <4 x i16> %tmp2
+}
+
+define <2 x i32> @vpaddlu16(<4 x i16>* %A) nounwind {
+;CHECK: vpaddlu16:
+;CHECK: vpaddl.u16
+ %tmp1 = load <4 x i16>* %A
+ %tmp2 = call <2 x i32> @llvm.arm.neon.vpaddlu.v2i32.v4i16(<4 x i16> %tmp1)
+ ret <2 x i32> %tmp2
+}
+
+define <1 x i64> @vpaddlu32(<2 x i32>* %A) nounwind {
+;CHECK: vpaddlu32:
+;CHECK: vpaddl.u32
+ %tmp1 = load <2 x i32>* %A
+ %tmp2 = call <1 x i64> @llvm.arm.neon.vpaddlu.v1i64.v2i32(<2 x i32> %tmp1)
+ ret <1 x i64> %tmp2
+}
+
+define <8 x i16> @vpaddlQs8(<16 x i8>* %A) nounwind {
+;CHECK: vpaddlQs8:
+;CHECK: vpaddl.s8
+ %tmp1 = load <16 x i8>* %A
+ %tmp2 = call <8 x i16> @llvm.arm.neon.vpaddls.v8i16.v16i8(<16 x i8> %tmp1)
+ ret <8 x i16> %tmp2
+}
+
+define <4 x i32> @vpaddlQs16(<8 x i16>* %A) nounwind {
+;CHECK: vpaddlQs16:
+;CHECK: vpaddl.s16
+ %tmp1 = load <8 x i16>* %A
+ %tmp2 = call <4 x i32> @llvm.arm.neon.vpaddls.v4i32.v8i16(<8 x i16> %tmp1)
+ ret <4 x i32> %tmp2
+}
+
+define <2 x i64> @vpaddlQs32(<4 x i32>* %A) nounwind {
+;CHECK: vpaddlQs32:
+;CHECK: vpaddl.s32
+ %tmp1 = load <4 x i32>* %A
+ %tmp2 = call <2 x i64> @llvm.arm.neon.vpaddls.v2i64.v4i32(<4 x i32> %tmp1)
+ ret <2 x i64> %tmp2
+}
+
+define <8 x i16> @vpaddlQu8(<16 x i8>* %A) nounwind {
+;CHECK: vpaddlQu8:
+;CHECK: vpaddl.u8
+ %tmp1 = load <16 x i8>* %A
+ %tmp2 = call <8 x i16> @llvm.arm.neon.vpaddlu.v8i16.v16i8(<16 x i8> %tmp1)
+ ret <8 x i16> %tmp2
+}
+
+define <4 x i32> @vpaddlQu16(<8 x i16>* %A) nounwind {
+;CHECK: vpaddlQu16:
+;CHECK: vpaddl.u16
+ %tmp1 = load <8 x i16>* %A
+ %tmp2 = call <4 x i32> @llvm.arm.neon.vpaddlu.v4i32.v8i16(<8 x i16> %tmp1)
+ ret <4 x i32> %tmp2
+}
+
+define <2 x i64> @vpaddlQu32(<4 x i32>* %A) nounwind {
+;CHECK: vpaddlQu32:
+;CHECK: vpaddl.u32
+ %tmp1 = load <4 x i32>* %A
+ %tmp2 = call <2 x i64> @llvm.arm.neon.vpaddlu.v2i64.v4i32(<4 x i32> %tmp1)
+ ret <2 x i64> %tmp2
+}
+
+declare <4 x i16> @llvm.arm.neon.vpaddls.v4i16.v8i8(<8 x i8>) nounwind readnone
+declare <2 x i32> @llvm.arm.neon.vpaddls.v2i32.v4i16(<4 x i16>) nounwind readnone
+declare <1 x i64> @llvm.arm.neon.vpaddls.v1i64.v2i32(<2 x i32>) nounwind readnone
+
+declare <4 x i16> @llvm.arm.neon.vpaddlu.v4i16.v8i8(<8 x i8>) nounwind readnone
+declare <2 x i32> @llvm.arm.neon.vpaddlu.v2i32.v4i16(<4 x i16>) nounwind readnone
+declare <1 x i64> @llvm.arm.neon.vpaddlu.v1i64.v2i32(<2 x i32>) nounwind readnone
+
+declare <8 x i16> @llvm.arm.neon.vpaddls.v8i16.v16i8(<16 x i8>) nounwind readnone
+declare <4 x i32> @llvm.arm.neon.vpaddls.v4i32.v8i16(<8 x i16>) nounwind readnone
+declare <2 x i64> @llvm.arm.neon.vpaddls.v2i64.v4i32(<4 x i32>) nounwind readnone
+
+declare <8 x i16> @llvm.arm.neon.vpaddlu.v8i16.v16i8(<16 x i8>) nounwind readnone
+declare <4 x i32> @llvm.arm.neon.vpaddlu.v4i32.v8i16(<8 x i16>) nounwind readnone
+declare <2 x i64> @llvm.arm.neon.vpaddlu.v2i64.v4i32(<4 x i32>) nounwind readnone
OpenPOWER on IntegriCloud