1 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2 ; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve < %s | FileCheck %s
3 ; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sme < %s | FileCheck %s
9 define <vscale x 16 x i8> @rbit_i8(<vscale x 16 x i8> %a, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %b) {
10 ; CHECK-LABEL: rbit_i8:
12 ; CHECK-NEXT: rbit z0.b, p0/m, z1.b
14 %out = call <vscale x 16 x i8> @llvm.aarch64.sve.rbit.nxv16i8(<vscale x 16 x i8> %a,
15 <vscale x 16 x i1> %pg,
16 <vscale x 16 x i8> %b)
17 ret <vscale x 16 x i8> %out
20 define <vscale x 8 x i16> @rbit_i16(<vscale x 8 x i16> %a, <vscale x 8 x i1> %pg, <vscale x 8 x i16> %b) {
21 ; CHECK-LABEL: rbit_i16:
23 ; CHECK-NEXT: rbit z0.h, p0/m, z1.h
25 %out = call <vscale x 8 x i16> @llvm.aarch64.sve.rbit.nxv8i16(<vscale x 8 x i16> %a,
26 <vscale x 8 x i1> %pg,
27 <vscale x 8 x i16> %b)
28 ret <vscale x 8 x i16> %out
31 define <vscale x 4 x i32> @rbit_i32(<vscale x 4 x i32> %a, <vscale x 4 x i1> %pg, <vscale x 4 x i32> %b) {
32 ; CHECK-LABEL: rbit_i32:
34 ; CHECK-NEXT: rbit z0.s, p0/m, z1.s
36 %out = call <vscale x 4 x i32> @llvm.aarch64.sve.rbit.nxv4i32(<vscale x 4 x i32> %a,
37 <vscale x 4 x i1> %pg,
38 <vscale x 4 x i32> %b)
39 ret <vscale x 4 x i32> %out
42 define <vscale x 2 x i64> @rbit_i64(<vscale x 2 x i64> %a, <vscale x 2 x i1> %pg, <vscale x 2 x i64> %b) {
43 ; CHECK-LABEL: rbit_i64:
45 ; CHECK-NEXT: rbit z0.d, p0/m, z1.d
47 %out = call <vscale x 2 x i64> @llvm.aarch64.sve.rbit.nxv2i64(<vscale x 2 x i64> %a,
48 <vscale x 2 x i1> %pg,
49 <vscale x 2 x i64> %b)
50 ret <vscale x 2 x i64> %out
57 define <vscale x 8 x i16> @revb_i16(<vscale x 8 x i16> %a, <vscale x 8 x i1> %pg, <vscale x 8 x i16> %b) {
58 ; CHECK-LABEL: revb_i16:
60 ; CHECK-NEXT: revb z0.h, p0/m, z1.h
62 %out = call <vscale x 8 x i16> @llvm.aarch64.sve.revb.nxv8i16(<vscale x 8 x i16> %a,
63 <vscale x 8 x i1> %pg,
64 <vscale x 8 x i16> %b)
65 ret <vscale x 8 x i16> %out
68 define <vscale x 4 x i32> @revb_i32(<vscale x 4 x i32> %a, <vscale x 4 x i1> %pg, <vscale x 4 x i32> %b) {
69 ; CHECK-LABEL: revb_i32:
71 ; CHECK-NEXT: revb z0.s, p0/m, z1.s
73 %out = call <vscale x 4 x i32> @llvm.aarch64.sve.revb.nxv4i32(<vscale x 4 x i32> %a,
74 <vscale x 4 x i1> %pg,
75 <vscale x 4 x i32> %b)
76 ret <vscale x 4 x i32> %out
79 define <vscale x 2 x i64> @revb_i64(<vscale x 2 x i64> %a, <vscale x 2 x i1> %pg, <vscale x 2 x i64> %b) {
80 ; CHECK-LABEL: revb_i64:
82 ; CHECK-NEXT: revb z0.d, p0/m, z1.d
84 %out = call <vscale x 2 x i64> @llvm.aarch64.sve.revb.nxv2i64(<vscale x 2 x i64> %a,
85 <vscale x 2 x i1> %pg,
86 <vscale x 2 x i64> %b)
87 ret <vscale x 2 x i64> %out
94 define <vscale x 4 x i32> @revh_i32(<vscale x 4 x i32> %a, <vscale x 4 x i1> %pg, <vscale x 4 x i32> %b) {
95 ; CHECK-LABEL: revh_i32:
97 ; CHECK-NEXT: revh z0.s, p0/m, z1.s
99 %out = call <vscale x 4 x i32> @llvm.aarch64.sve.revh.nxv4i32(<vscale x 4 x i32> %a,
100 <vscale x 4 x i1> %pg,
101 <vscale x 4 x i32> %b)
102 ret <vscale x 4 x i32> %out
105 define <vscale x 2 x i64> @revh_i64(<vscale x 2 x i64> %a, <vscale x 2 x i1> %pg, <vscale x 2 x i64> %b) {
106 ; CHECK-LABEL: revh_i64:
108 ; CHECK-NEXT: revh z0.d, p0/m, z1.d
110 %out = call <vscale x 2 x i64> @llvm.aarch64.sve.revh.nxv2i64(<vscale x 2 x i64> %a,
111 <vscale x 2 x i1> %pg,
112 <vscale x 2 x i64> %b)
113 ret <vscale x 2 x i64> %out
120 define <vscale x 2 x i64> @revw_i64(<vscale x 2 x i64> %a, <vscale x 2 x i1> %pg, <vscale x 2 x i64> %b) {
121 ; CHECK-LABEL: revw_i64:
123 ; CHECK-NEXT: revw z0.d, p0/m, z1.d
125 %out = call <vscale x 2 x i64> @llvm.aarch64.sve.revw.nxv2i64(<vscale x 2 x i64> %a,
126 <vscale x 2 x i1> %pg,
127 <vscale x 2 x i64> %b)
128 ret <vscale x 2 x i64> %out
131 declare <vscale x 16 x i8> @llvm.aarch64.sve.rbit.nxv16i8(<vscale x 16 x i8>, <vscale x 16 x i1>, <vscale x 16 x i8>)
132 declare <vscale x 8 x i16> @llvm.aarch64.sve.rbit.nxv8i16(<vscale x 8 x i16>, <vscale x 8 x i1>, <vscale x 8 x i16>)
133 declare <vscale x 4 x i32> @llvm.aarch64.sve.rbit.nxv4i32(<vscale x 4 x i32>, <vscale x 4 x i1>, <vscale x 4 x i32>)
134 declare <vscale x 2 x i64> @llvm.aarch64.sve.rbit.nxv2i64(<vscale x 2 x i64>, <vscale x 2 x i1>, <vscale x 2 x i64>)
136 declare <vscale x 8 x i16> @llvm.aarch64.sve.revb.nxv8i16(<vscale x 8 x i16>, <vscale x 8 x i1>, <vscale x 8 x i16>)
137 declare <vscale x 4 x i32> @llvm.aarch64.sve.revb.nxv4i32(<vscale x 4 x i32>, <vscale x 4 x i1>, <vscale x 4 x i32>)
138 declare <vscale x 2 x i64> @llvm.aarch64.sve.revb.nxv2i64(<vscale x 2 x i64>, <vscale x 2 x i1>, <vscale x 2 x i64>)
140 declare <vscale x 4 x i32> @llvm.aarch64.sve.revh.nxv4i32(<vscale x 4 x i32>, <vscale x 4 x i1>, <vscale x 4 x i32>)
141 declare <vscale x 2 x i64> @llvm.aarch64.sve.revh.nxv2i64(<vscale x 2 x i64>, <vscale x 2 x i1>, <vscale x 2 x i64>)
143 declare <vscale x 2 x i64> @llvm.aarch64.sve.revw.nxv2i64(<vscale x 2 x i64>, <vscale x 2 x i1>, <vscale x 2 x i64>)