1 ; RUN: llc < %s -march=cellspu > %t1.s
2 ; RUN: grep cbd %t1.s | count 5
3 ; RUN: grep chd %t1.s | count 5
4 ; RUN: grep cwd %t1.s | count 11
5 ; RUN: grep -w il %t1.s | count 5
6 ; RUN: grep -w ilh %t1.s | count 6
7 ; RUN: grep iohl %t1.s | count 1
8 ; RUN: grep ilhu %t1.s | count 4
9 ; RUN: grep shufb %t1.s | count 27
10 ; RUN: grep 17219 %t1.s | count 1
11 ; RUN: grep 22598 %t1.s | count 1
12 ; RUN: grep -- -39 %t1.s | count 1
13 ; RUN: grep 24 %t1.s | count 1
14 ; RUN: grep 1159 %t1.s | count 1
15 ; RUN: FileCheck %s < %t1.s
17 ; ModuleID = 'vecinsert.bc'
18 target datalayout = "E-p:32:32:128-f64:64:128-f32:32:128-i64:32:128-i32:32:128-i16:16:128-i8:8:128-i1:8:128-a0:0:128-v128:128:128"
19 target triple = "spu-unknown-elf"
21 ; 67 -> 0x43, as 8-bit vector constant load = 0x4343 (17219)0x4343
22 define <16 x i8> @test_v16i8(<16 x i8> %P, i8 %x) {
24 %tmp1 = insertelement <16 x i8> %P, i8 %x, i32 10
25 %tmp1.1 = insertelement <16 x i8> %tmp1, i8 67, i32 7
26 %tmp1.2 = insertelement <16 x i8> %tmp1.1, i8 %x, i32 15
31 define <8 x i16> @test_v8i16(<8 x i16> %P, i16 %x) {
33 %tmp1 = insertelement <8 x i16> %P, i16 %x, i32 5
34 %tmp1.1 = insertelement <8 x i16> %tmp1, i16 22598, i32 7
35 %tmp1.2 = insertelement <8 x i16> %tmp1.1, i16 %x, i32 2
39 ; 1574023 -> 0x180487 (ILHU 24/IOHL 1159)
40 define <4 x i32> @test_v4i32_1(<4 x i32> %P, i32 %x) {
42 %tmp1 = insertelement <4 x i32> %P, i32 %x, i32 2
43 %tmp1.1 = insertelement <4 x i32> %tmp1, i32 1574023, i32 1
44 %tmp1.2 = insertelement <4 x i32> %tmp1.1, i32 %x, i32 3
48 ; Should generate IL for the load
49 define <4 x i32> @test_v4i32_2(<4 x i32> %P, i32 %x) {
51 %tmp1 = insertelement <4 x i32> %P, i32 %x, i32 2
52 %tmp1.1 = insertelement <4 x i32> %tmp1, i32 -39, i32 1
53 %tmp1.2 = insertelement <4 x i32> %tmp1.1, i32 %x, i32 3
57 define void @variable_v16i8_1(<16 x i8>* %a, i32 %i) nounwind {
59 %arrayidx = getelementptr <16 x i8>* %a, i32 %i
60 %tmp2 = load <16 x i8>* %arrayidx
61 %tmp3 = insertelement <16 x i8> %tmp2, i8 1, i32 1
62 %tmp8 = insertelement <16 x i8> %tmp3, i8 2, i32 11
63 store <16 x i8> %tmp8, <16 x i8>* %arrayidx
67 define void @variable_v8i16_1(<8 x i16>* %a, i32 %i) nounwind {
69 %arrayidx = getelementptr <8 x i16>* %a, i32 %i
70 %tmp2 = load <8 x i16>* %arrayidx
71 %tmp3 = insertelement <8 x i16> %tmp2, i16 1, i32 1
72 %tmp8 = insertelement <8 x i16> %tmp3, i16 2, i32 6
73 store <8 x i16> %tmp8, <8 x i16>* %arrayidx
77 define void @variable_v4i32_1(<4 x i32>* %a, i32 %i) nounwind {
79 %arrayidx = getelementptr <4 x i32>* %a, i32 %i
80 %tmp2 = load <4 x i32>* %arrayidx
81 %tmp3 = insertelement <4 x i32> %tmp2, i32 1, i32 1
82 %tmp8 = insertelement <4 x i32> %tmp3, i32 2, i32 2
83 store <4 x i32> %tmp8, <4 x i32>* %arrayidx
87 define void @variable_v4f32_1(<4 x float>* %a, i32 %i) nounwind {
89 %arrayidx = getelementptr <4 x float>* %a, i32 %i
90 %tmp2 = load <4 x float>* %arrayidx
91 %tmp3 = insertelement <4 x float> %tmp2, float 1.000000e+00, i32 1
92 %tmp8 = insertelement <4 x float> %tmp3, float 2.000000e+00, i32 2
93 store <4 x float> %tmp8, <4 x float>* %arrayidx
97 define void @variable_v2i64_1(<2 x i64>* %a, i32 %i) nounwind {
99 %arrayidx = getelementptr <2 x i64>* %a, i32 %i
100 %tmp2 = load <2 x i64>* %arrayidx
101 %tmp3 = insertelement <2 x i64> %tmp2, i64 615, i32 0
102 store <2 x i64> %tmp3, <2 x i64>* %arrayidx
106 define void @variable_v2i64_2(<2 x i64>* %a, i32 %i) nounwind {
108 %arrayidx = getelementptr <2 x i64>* %a, i32 %i
109 %tmp2 = load <2 x i64>* %arrayidx
110 %tmp3 = insertelement <2 x i64> %tmp2, i64 615, i32 1
111 store <2 x i64> %tmp3, <2 x i64>* %arrayidx
115 define void @variable_v2f64_1(<2 x double>* %a, i32 %i) nounwind {
117 %arrayidx = getelementptr <2 x double>* %a, i32 %i
118 %tmp2 = load <2 x double>* %arrayidx
119 %tmp3 = insertelement <2 x double> %tmp2, double 1.000000e+00, i32 1
120 store <2 x double> %tmp3, <2 x double>* %arrayidx
124 define <4 x i32> @undef_v4i32( i32 %param ) {
128 %val = insertelement <4 x i32> <i32 1, i32 2, i32 3, i32 4>, i32 %param, i32 undef