1 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2 ; RUN: llc -mtriple=thumbv8.1m.main -mattr=+mve.fp -verify-machineinstrs -o - %s | FileCheck %s
4 define arm_aapcs_vfpcc <4 x i32> @test_vaddq_u32(<4 x i32> %a, <4 x i32> %b) local_unnamed_addr #0 {
5 ; CHECK-LABEL: test_vaddq_u32:
6 ; CHECK: @ %bb.0: @ %entry
7 ; CHECK-NEXT: vadd.i32 q0, q1, q0
10 %0 = add <4 x i32> %b, %a
14 define arm_aapcs_vfpcc <8 x half> @test_vaddq_f16(<8 x half> %a, <8 x half> %b) local_unnamed_addr #0 {
15 ; CHECK-LABEL: test_vaddq_f16:
16 ; CHECK: @ %bb.0: @ %entry
17 ; CHECK-NEXT: vadd.f16 q0, q0, q1
20 %0 = fadd <8 x half> %a, %b
24 define arm_aapcs_vfpcc <16 x i8> @test_vaddq_m_s8(<16 x i8> %inactive, <16 x i8> %a, <16 x i8> %b, i16 zeroext %p) local_unnamed_addr #1 {
25 ; CHECK-LABEL: test_vaddq_m_s8:
26 ; CHECK: @ %bb.0: @ %entry
27 ; CHECK-NEXT: vmsr p0, r0
29 ; CHECK-NEXT: vaddt.i8 q0, q1, q2
32 %0 = zext i16 %p to i32
33 %1 = tail call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 %0)
34 %2 = tail call <16 x i8> @llvm.arm.mve.add.predicated.v16i8.v16i1(<16 x i8> %a, <16 x i8> %b, <16 x i1> %1, <16 x i8> %inactive)
38 declare <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32) #2
40 declare <16 x i8> @llvm.arm.mve.add.predicated.v16i8.v16i1(<16 x i8>, <16 x i8>, <16 x i1>, <16 x i8>) #2
42 define arm_aapcs_vfpcc <4 x float> @test_vaddq_m_f32(<4 x float> %inactive, <4 x float> %a, <4 x float> %b, i16 zeroext %p) local_unnamed_addr #1 {
43 ; CHECK-LABEL: test_vaddq_m_f32:
44 ; CHECK: @ %bb.0: @ %entry
45 ; CHECK-NEXT: vmsr p0, r0
47 ; CHECK-NEXT: vaddt.f32 q0, q1, q2
50 %0 = zext i16 %p to i32
51 %1 = tail call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 %0)
52 %2 = tail call <4 x float> @llvm.arm.mve.add.predicated.v4f32.v4i1(<4 x float> %a, <4 x float> %b, <4 x i1> %1, <4 x float> %inactive)
56 declare <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32) #2
58 declare <4 x float> @llvm.arm.mve.add.predicated.v4f32.v4i1(<4 x float>, <4 x float>, <4 x i1>, <4 x float>) #2
60 define arm_aapcs_vfpcc <8 x i16> @test_vaddq_x_u16(<8 x i16> %a, <8 x i16> %b, i16 zeroext %p) local_unnamed_addr #1 {
61 ; CHECK-LABEL: test_vaddq_x_u16:
62 ; CHECK: @ %bb.0: @ %entry
63 ; CHECK-NEXT: vmsr p0, r0
65 ; CHECK-NEXT: vaddt.i16 q0, q0, q1
68 %0 = zext i16 %p to i32
69 %1 = tail call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 %0)
70 %2 = tail call <8 x i16> @llvm.arm.mve.add.predicated.v8i16.v8i1(<8 x i16> %a, <8 x i16> %b, <8 x i1> %1, <8 x i16> undef)
74 declare <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32) #2
76 declare <8 x i16> @llvm.arm.mve.add.predicated.v8i16.v8i1(<8 x i16>, <8 x i16>, <8 x i1>, <8 x i16>) #2
78 define arm_aapcs_vfpcc <8 x half> @test_vaddq_x_f16(<8 x half> %a, <8 x half> %b, i16 zeroext %p) local_unnamed_addr #1 {
79 ; CHECK-LABEL: test_vaddq_x_f16:
80 ; CHECK: @ %bb.0: @ %entry
81 ; CHECK-NEXT: vmsr p0, r0
83 ; CHECK-NEXT: vaddt.f16 q0, q0, q1
86 %0 = zext i16 %p to i32
87 %1 = tail call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 %0)
88 %2 = tail call <8 x half> @llvm.arm.mve.add.predicated.v8f16.v8i1(<8 x half> %a, <8 x half> %b, <8 x i1> %1, <8 x half> undef)
92 declare <8 x half> @llvm.arm.mve.add.predicated.v8f16.v8i1(<8 x half>, <8 x half>, <8 x i1>, <8 x half>) #2
94 define arm_aapcs_vfpcc <4 x i32> @test_vaddq_n_u32(<4 x i32> %a, i32 %b) {
95 ; CHECK-LABEL: test_vaddq_n_u32:
96 ; CHECK: @ %bb.0: @ %entry
97 ; CHECK-NEXT: vadd.i32 q0, q0, r0
100 %.splatinsert = insertelement <4 x i32> undef, i32 %b, i32 0
101 %.splat = shufflevector <4 x i32> %.splatinsert, <4 x i32> undef, <4 x i32> zeroinitializer
102 %0 = add <4 x i32> %.splat, %a
106 define arm_aapcs_vfpcc <8 x half> @test_vaddq_n_f16(<8 x half> %a, float %b.coerce) {
107 ; CHECK-LABEL: test_vaddq_n_f16:
108 ; CHECK: @ %bb.0: @ %entry
109 ; CHECK-NEXT: vmov r0, s4
110 ; CHECK-NEXT: vadd.f16 q0, q0, r0
113 %0 = bitcast float %b.coerce to i32
114 %tmp.0.extract.trunc = trunc i32 %0 to i16
115 %1 = bitcast i16 %tmp.0.extract.trunc to half
116 %.splatinsert = insertelement <8 x half> undef, half %1, i32 0
117 %.splat = shufflevector <8 x half> %.splatinsert, <8 x half> undef, <8 x i32> zeroinitializer
118 %2 = fadd <8 x half> %.splat, %a
122 define arm_aapcs_vfpcc <16 x i8> @test_vaddq_m_n_s8(<16 x i8> %inactive, <16 x i8> %a, i8 signext %b, i16 zeroext %p) {
123 ; CHECK-LABEL: test_vaddq_m_n_s8:
124 ; CHECK: @ %bb.0: @ %entry
125 ; CHECK-NEXT: vmsr p0, r1
127 ; CHECK-NEXT: vaddt.i8 q0, q1, r0
130 %.splatinsert = insertelement <16 x i8> undef, i8 %b, i32 0
131 %.splat = shufflevector <16 x i8> %.splatinsert, <16 x i8> undef, <16 x i32> zeroinitializer
132 %0 = zext i16 %p to i32
133 %1 = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 %0)
134 %2 = call <16 x i8> @llvm.arm.mve.add.predicated.v16i8.v16i1(<16 x i8> %a, <16 x i8> %.splat, <16 x i1> %1, <16 x i8> %inactive)
138 define arm_aapcs_vfpcc <4 x float> @test_vaddq_m_n_f32(<4 x float> %inactive, <4 x float> %a, float %b, i16 zeroext %p) {
139 ; CHECK-LABEL: test_vaddq_m_n_f32:
140 ; CHECK: @ %bb.0: @ %entry
141 ; CHECK-NEXT: vmov r1, s8
142 ; CHECK-NEXT: vmsr p0, r0
144 ; CHECK-NEXT: vaddt.f32 q0, q1, r1
147 %.splatinsert = insertelement <4 x float> undef, float %b, i32 0
148 %.splat = shufflevector <4 x float> %.splatinsert, <4 x float> undef, <4 x i32> zeroinitializer
149 %0 = zext i16 %p to i32
150 %1 = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 %0)
151 %2 = call <4 x float> @llvm.arm.mve.add.predicated.v4f32.v4i1(<4 x float> %a, <4 x float> %.splat, <4 x i1> %1, <4 x float> %inactive)
155 define arm_aapcs_vfpcc <8 x i16> @test_vaddq_x_n_u16(<8 x i16> %a, i16 zeroext %b, i16 zeroext %p) {
156 ; CHECK-LABEL: test_vaddq_x_n_u16:
157 ; CHECK: @ %bb.0: @ %entry
158 ; CHECK-NEXT: vmsr p0, r1
160 ; CHECK-NEXT: vaddt.i16 q0, q0, r0
163 %.splatinsert = insertelement <8 x i16> undef, i16 %b, i32 0
164 %.splat = shufflevector <8 x i16> %.splatinsert, <8 x i16> undef, <8 x i32> zeroinitializer
165 %0 = zext i16 %p to i32
166 %1 = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 %0)
167 %2 = call <8 x i16> @llvm.arm.mve.add.predicated.v8i16.v8i1(<8 x i16> %a, <8 x i16> %.splat, <8 x i1> %1, <8 x i16> undef)
171 define arm_aapcs_vfpcc <8 x half> @test_vaddq_x_n_f16(<8 x half> %a, float %b.coerce, i16 zeroext %p) {
172 ; CHECK-LABEL: test_vaddq_x_n_f16:
173 ; CHECK: @ %bb.0: @ %entry
174 ; CHECK-NEXT: vmov r1, s4
175 ; CHECK-NEXT: vmsr p0, r0
177 ; CHECK-NEXT: vaddt.f16 q0, q0, r1
180 %0 = bitcast float %b.coerce to i32
181 %tmp.0.extract.trunc = trunc i32 %0 to i16
182 %1 = bitcast i16 %tmp.0.extract.trunc to half
183 %.splatinsert = insertelement <8 x half> undef, half %1, i32 0
184 %.splat = shufflevector <8 x half> %.splatinsert, <8 x half> undef, <8 x i32> zeroinitializer
185 %2 = zext i16 %p to i32
186 %3 = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 %2)
187 %4 = call <8 x half> @llvm.arm.mve.add.predicated.v8f16.v8i1(<8 x half> %a, <8 x half> %.splat, <8 x i1> %3, <8 x half> undef)