1 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 2
2 ; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sme2 -force-streaming < %s | FileCheck %s
4 ; lookup table expand one register
6 define {<vscale x 16 x i8>, <vscale x 16 x i8>} @luti4_i8(<vscale x 16 x i8> %x) {
7 ; CHECK-LABEL: luti4_i8:
9 ; CHECK-NEXT: luti4 { z0.b, z1.b }, zt0, z0[3]
11 %res = call {<vscale x 16 x i8>, <vscale x 16 x i8>} @llvm.aarch64.sme.luti4.lane.zt.x2.nxv16i8(i32 0, <vscale x 16 x i8> %x, i32 3)
12 ret {<vscale x 16 x i8>, <vscale x 16 x i8>} %res
15 define {<vscale x 8 x i16>, <vscale x 8 x i16>} @luti4_i16(<vscale x 16 x i8> %x) {
16 ; CHECK-LABEL: luti4_i16:
18 ; CHECK-NEXT: luti4 { z0.h, z1.h }, zt0, z0[3]
20 %res = call {<vscale x 8 x i16>, <vscale x 8 x i16>} @llvm.aarch64.sme.luti4.lane.zt.x2.nxv8i16(i32 0, <vscale x 16 x i8> %x, i32 3)
21 ret {<vscale x 8 x i16>, <vscale x 8 x i16>} %res
24 define {<vscale x 4 x i32>, <vscale x 4 x i32>} @luti4_i32(<vscale x 16 x i8> %x) {
25 ; CHECK-LABEL: luti4_i32:
27 ; CHECK-NEXT: luti4 { z0.s, z1.s }, zt0, z0[3]
29 %res = call {<vscale x 4 x i32>, <vscale x 4 x i32>} @llvm.aarch64.sme.luti4.lane.zt.x2.nxv4i32(i32 0, <vscale x 16 x i8> %x, i32 3)
30 ret {<vscale x 4 x i32>, <vscale x 4 x i32>} %res
33 define {<vscale x 8 x half>, <vscale x 8 x half>} @luti4_f16(<vscale x 16 x i8> %x) {
34 ; CHECK-LABEL: luti4_f16:
36 ; CHECK-NEXT: luti4 { z0.h, z1.h }, zt0, z0[3]
38 %res = call {<vscale x 8 x half>, <vscale x 8 x half>} @llvm.aarch64.sme.luti4.lane.zt.x2.nxv8f16(i32 0, <vscale x 16 x i8> %x, i32 3)
39 ret {<vscale x 8 x half>, <vscale x 8 x half>} %res
42 define {<vscale x 8 x bfloat>, <vscale x 8 x bfloat>} @luti4_bf16(<vscale x 16 x i8> %x) {
43 ; CHECK-LABEL: luti4_bf16:
45 ; CHECK-NEXT: luti4 { z0.h, z1.h }, zt0, z0[3]
47 %res = call {<vscale x 8 x bfloat>, <vscale x 8 x bfloat>} @llvm.aarch64.sme.luti4.lane.zt.x2.nxv8bf16(i32 0, <vscale x 16 x i8> %x, i32 3)
48 ret {<vscale x 8 x bfloat>, <vscale x 8 x bfloat>} %res
51 define {<vscale x 4 x float>, <vscale x 4 x float>} @luti4_f32(<vscale x 16 x i8> %x) {
52 ; CHECK-LABEL: luti4_f32:
54 ; CHECK-NEXT: luti4 { z0.s, z1.s }, zt0, z0[3]
56 %res = call {<vscale x 4 x float>, <vscale x 4 x float>} @llvm.aarch64.sme.luti4.lane.zt.x2.nxv4f32(i32 0, <vscale x 16 x i8> %x, i32 3)
57 ret {<vscale x 4 x float>, <vscale x 4 x float>} %res
60 declare {<vscale x 16 x i8>, <vscale x 16 x i8>} @llvm.aarch64.sme.luti4.lane.zt.x2.nxv16i8(i32, <vscale x 16 x i8>, i32)
61 declare {<vscale x 8 x i16>, <vscale x 8 x i16>} @llvm.aarch64.sme.luti4.lane.zt.x2.nxv8i16(i32, <vscale x 16 x i8>, i32)
62 declare {<vscale x 4 x i32>, <vscale x 4 x i32>} @llvm.aarch64.sme.luti4.lane.zt.x2.nxv4i32(i32, <vscale x 16 x i8>, i32)
63 declare {<vscale x 8 x half>, <vscale x 8 x half>} @llvm.aarch64.sme.luti4.lane.zt.x2.nxv8f16(i32, <vscale x 16 x i8>, i32)
64 declare {<vscale x 8 x bfloat>, <vscale x 8 x bfloat>} @llvm.aarch64.sme.luti4.lane.zt.x2.nxv8bf16(i32, <vscale x 16 x i8>, i32)
65 declare {<vscale x 4 x float>, <vscale x 4 x float>} @llvm.aarch64.sme.luti4.lane.zt.x2.nxv4f32(i32, <vscale x 16 x i8>, i32)