1 ; Test the MSA intrinsics that are encoded with the ELM instruction format and
2 ; are element extraction operations.
4 ; RUN: llc -march=mips -mattr=+msa < %s | FileCheck %s
6 @llvm_mips_copy_s_b_ARG1 = global <16 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15>, align 16
7 @llvm_mips_copy_s_b_RES = global i32 0, align 16
9 define void @llvm_mips_copy_s_b_test() nounwind {
11 %0 = load <16 x i8>* @llvm_mips_copy_s_b_ARG1
12 %1 = tail call i32 @llvm.mips.copy.s.b(<16 x i8> %0, i32 1)
13 store i32 %1, i32* @llvm_mips_copy_s_b_RES
17 declare i32 @llvm.mips.copy.s.b(<16 x i8>, i32) nounwind
19 ; CHECK: llvm_mips_copy_s_b_test:
23 ; CHECK: .size llvm_mips_copy_s_b_test
25 @llvm_mips_copy_s_h_ARG1 = global <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>, align 16
26 @llvm_mips_copy_s_h_RES = global i32 0, align 16
28 define void @llvm_mips_copy_s_h_test() nounwind {
30 %0 = load <8 x i16>* @llvm_mips_copy_s_h_ARG1
31 %1 = tail call i32 @llvm.mips.copy.s.h(<8 x i16> %0, i32 1)
32 store i32 %1, i32* @llvm_mips_copy_s_h_RES
36 declare i32 @llvm.mips.copy.s.h(<8 x i16>, i32) nounwind
38 ; CHECK: llvm_mips_copy_s_h_test:
42 ; CHECK: .size llvm_mips_copy_s_h_test
44 @llvm_mips_copy_s_w_ARG1 = global <4 x i32> <i32 0, i32 1, i32 2, i32 3>, align 16
45 @llvm_mips_copy_s_w_RES = global i32 0, align 16
47 define void @llvm_mips_copy_s_w_test() nounwind {
49 %0 = load <4 x i32>* @llvm_mips_copy_s_w_ARG1
50 %1 = tail call i32 @llvm.mips.copy.s.w(<4 x i32> %0, i32 1)
51 store i32 %1, i32* @llvm_mips_copy_s_w_RES
55 declare i32 @llvm.mips.copy.s.w(<4 x i32>, i32) nounwind
57 ; CHECK: llvm_mips_copy_s_w_test:
61 ; CHECK: .size llvm_mips_copy_s_w_test
63 @llvm_mips_copy_u_b_ARG1 = global <16 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15>, align 16
64 @llvm_mips_copy_u_b_RES = global i32 0, align 16
66 define void @llvm_mips_copy_u_b_test() nounwind {
68 %0 = load <16 x i8>* @llvm_mips_copy_u_b_ARG1
69 %1 = tail call i32 @llvm.mips.copy.u.b(<16 x i8> %0, i32 1)
70 store i32 %1, i32* @llvm_mips_copy_u_b_RES
74 declare i32 @llvm.mips.copy.u.b(<16 x i8>, i32) nounwind
76 ; CHECK: llvm_mips_copy_u_b_test:
80 ; CHECK: .size llvm_mips_copy_u_b_test
82 @llvm_mips_copy_u_h_ARG1 = global <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>, align 16
83 @llvm_mips_copy_u_h_RES = global i32 0, align 16
85 define void @llvm_mips_copy_u_h_test() nounwind {
87 %0 = load <8 x i16>* @llvm_mips_copy_u_h_ARG1
88 %1 = tail call i32 @llvm.mips.copy.u.h(<8 x i16> %0, i32 1)
89 store i32 %1, i32* @llvm_mips_copy_u_h_RES
93 declare i32 @llvm.mips.copy.u.h(<8 x i16>, i32) nounwind
95 ; CHECK: llvm_mips_copy_u_h_test:
99 ; CHECK: .size llvm_mips_copy_u_h_test
101 @llvm_mips_copy_u_w_ARG1 = global <4 x i32> <i32 0, i32 1, i32 2, i32 3>, align 16
102 @llvm_mips_copy_u_w_RES = global i32 0, align 16
104 define void @llvm_mips_copy_u_w_test() nounwind {
106 %0 = load <4 x i32>* @llvm_mips_copy_u_w_ARG1
107 %1 = tail call i32 @llvm.mips.copy.u.w(<4 x i32> %0, i32 1)
108 store i32 %1, i32* @llvm_mips_copy_u_w_RES
112 declare i32 @llvm.mips.copy.u.w(<4 x i32>, i32) nounwind
114 ; CHECK: llvm_mips_copy_u_w_test:
118 ; CHECK: .size llvm_mips_copy_u_w_test