1 ; RUN: llc -mtriple=thumbv8.1m.main-arm-none-eabi -mattr=+mve.fp %s -o - | FileCheck %s
3 define arm_aapcs_vfpcc <4 x i32> @vmlau32(<4 x i32> %A, <4 x i32> %B, i32 %X) nounwind {
4 ; CHECK-LABEL: vmlau32:
5 ; CHECK: @ %bb.0: @ %entry
6 ; CHECK-NEXT: vmla.u32 q0, q1, r0
9 %0 = insertelement <4 x i32> undef, i32 %X, i32 0
10 %1 = shufflevector <4 x i32> %0, <4 x i32> undef, <4 x i32> zeroinitializer
11 %2 = mul nsw <4 x i32> %B, %1
12 %3 = add nsw <4 x i32> %A, %2
16 define arm_aapcs_vfpcc <4 x i32> @vmlau32b(<4 x i32> %A, <4 x i32> %B, i32 %X) nounwind {
17 ; CHECK-LABEL: vmlau32b:
18 ; CHECK: @ %bb.0: @ %entry
19 ; CHECK-NEXT: vmla.u32 q0, q1, r0
22 %0 = insertelement <4 x i32> undef, i32 %X, i32 0
23 %1 = shufflevector <4 x i32> %0, <4 x i32> undef, <4 x i32> zeroinitializer
24 %2 = mul nsw <4 x i32> %1, %B
25 %3 = add nsw <4 x i32> %2, %A
29 define arm_aapcs_vfpcc <8 x i16> @vmlau16(<8 x i16> %A, <8 x i16> %B, i16 %X) nounwind {
30 ; CHECK-LABEL: vmlau16:
31 ; CHECK: @ %bb.0: @ %entry
32 ; CHECK-NEXT: vmla.u16 q0, q1, r0
35 %0 = insertelement <8 x i16> undef, i16 %X, i32 0
36 %1 = shufflevector <8 x i16> %0, <8 x i16> undef, <8 x i32> zeroinitializer
37 %2 = mul nsw <8 x i16> %B, %1
38 %3 = add nsw <8 x i16> %A, %2
42 define arm_aapcs_vfpcc <8 x i16> @vmlau16b(<8 x i16> %A, <8 x i16> %B, i16 %X) nounwind {
43 ; CHECK-LABEL: vmlau16b:
44 ; CHECK: @ %bb.0: @ %entry
45 ; CHECK-NEXT: vmla.u16 q0, q1, r0
48 %0 = insertelement <8 x i16> undef, i16 %X, i32 0
49 %1 = shufflevector <8 x i16> %0, <8 x i16> undef, <8 x i32> zeroinitializer
50 %2 = mul nsw <8 x i16> %1, %B
51 %3 = add nsw <8 x i16> %2, %A
55 define arm_aapcs_vfpcc <16 x i8> @vmlau8(<16 x i8> %A, <16 x i8> %B, i8 %X) nounwind {
56 ; CHECK-LABEL: vmlau8:
57 ; CHECK: @ %bb.0: @ %entry
58 ; CHECK-NEXT: vmla.u8 q0, q1, r0
61 %0 = insertelement <16 x i8> undef, i8 %X, i32 0
62 %1 = shufflevector <16 x i8> %0, <16 x i8> undef, <16 x i32> zeroinitializer
63 %2 = mul nsw <16 x i8> %B, %1
64 %3 = add nsw <16 x i8> %A, %2
68 define arm_aapcs_vfpcc <16 x i8> @vmlau8b(<16 x i8> %A, <16 x i8> %B, i8 %X) nounwind {
69 ; CHECK-LABEL: vmlau8b:
70 ; CHECK: @ %bb.0: @ %entry
71 ; CHECK-NEXT: vmla.u8 q0, q1, r0
74 %0 = insertelement <16 x i8> undef, i8 %X, i32 0
75 %1 = shufflevector <16 x i8> %0, <16 x i8> undef, <16 x i32> zeroinitializer
76 %2 = mul nsw <16 x i8> %1, %B
77 %3 = add nsw <16 x i8> %2, %A