1; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
2; RUN: opt -passes='lower-matrix-intrinsics' -S < %s | FileCheck %s
3
4; Currently we only lower stores with shape information, but need to embed the
5; matrix in a flat vector for function calls and returns.
6define <8 x double> @strided_load_4x4(<8 x double> %in, <8 x double>* %Ptr) {
7; CHECK-LABEL: @strided_load_4x4(
8; CHECK-NEXT:    [[SPLIT:%.*]] = shufflevector <8 x double> [[IN:%.*]], <8 x double> poison, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
9; CHECK-NEXT:    [[SPLIT1:%.*]] = shufflevector <8 x double> [[IN]], <8 x double> poison, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
10; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x double> [[SPLIT]], i64 0
11; CHECK-NEXT:    [[TMP2:%.*]] = insertelement <2 x double> poison, double [[TMP1]], i64 0
12; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <4 x double> [[SPLIT1]], i64 0
13; CHECK-NEXT:    [[TMP4:%.*]] = insertelement <2 x double> [[TMP2]], double [[TMP3]], i64 1
14; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <4 x double> [[SPLIT]], i64 1
15; CHECK-NEXT:    [[TMP6:%.*]] = insertelement <2 x double> poison, double [[TMP5]], i64 0
16; CHECK-NEXT:    [[TMP7:%.*]] = extractelement <4 x double> [[SPLIT1]], i64 1
17; CHECK-NEXT:    [[TMP8:%.*]] = insertelement <2 x double> [[TMP6]], double [[TMP7]], i64 1
18; CHECK-NEXT:    [[TMP9:%.*]] = extractelement <4 x double> [[SPLIT]], i64 2
19; CHECK-NEXT:    [[TMP10:%.*]] = insertelement <2 x double> poison, double [[TMP9]], i64 0
20; CHECK-NEXT:    [[TMP11:%.*]] = extractelement <4 x double> [[SPLIT1]], i64 2
21; CHECK-NEXT:    [[TMP12:%.*]] = insertelement <2 x double> [[TMP10]], double [[TMP11]], i64 1
22; CHECK-NEXT:    [[TMP13:%.*]] = extractelement <4 x double> [[SPLIT]], i64 3
23; CHECK-NEXT:    [[TMP14:%.*]] = insertelement <2 x double> poison, double [[TMP13]], i64 0
24; CHECK-NEXT:    [[TMP15:%.*]] = extractelement <4 x double> [[SPLIT1]], i64 3
25; CHECK-NEXT:    [[TMP16:%.*]] = insertelement <2 x double> [[TMP14]], double [[TMP15]], i64 1
26; CHECK-NEXT:    [[TMP17:%.*]] = shufflevector <2 x double> [[TMP4]], <2 x double> [[TMP8]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
27; CHECK-NEXT:    [[TMP18:%.*]] = shufflevector <2 x double> [[TMP12]], <2 x double> [[TMP16]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
28; CHECK-NEXT:    [[TMP19:%.*]] = shufflevector <4 x double> [[TMP17]], <4 x double> [[TMP18]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
29; CHECK-NEXT:    [[TMP20:%.*]] = bitcast <8 x double>* [[PTR:%.*]] to double*
30; CHECK-NEXT:    [[VEC_CAST:%.*]] = bitcast double* [[TMP20]] to <2 x double>*
31; CHECK-NEXT:    store <2 x double> [[TMP4]], <2 x double>* [[VEC_CAST]], align 8
32; CHECK-NEXT:    [[VEC_GEP:%.*]] = getelementptr double, double* [[TMP20]], i64 2
33; CHECK-NEXT:    [[VEC_CAST2:%.*]] = bitcast double* [[VEC_GEP]] to <2 x double>*
34; CHECK-NEXT:    store <2 x double> [[TMP8]], <2 x double>* [[VEC_CAST2]], align 8
35; CHECK-NEXT:    [[VEC_GEP3:%.*]] = getelementptr double, double* [[TMP20]], i64 4
36; CHECK-NEXT:    [[VEC_CAST4:%.*]] = bitcast double* [[VEC_GEP3]] to <2 x double>*
37; CHECK-NEXT:    store <2 x double> [[TMP12]], <2 x double>* [[VEC_CAST4]], align 8
38; CHECK-NEXT:    [[VEC_GEP5:%.*]] = getelementptr double, double* [[TMP20]], i64 6
39; CHECK-NEXT:    [[VEC_CAST6:%.*]] = bitcast double* [[VEC_GEP5]] to <2 x double>*
40; CHECK-NEXT:    store <2 x double> [[TMP16]], <2 x double>* [[VEC_CAST6]], align 8
41; CHECK-NEXT:    call void @foo(<8 x double> [[TMP19]])
42; CHECK-NEXT:    ret <8 x double> [[TMP19]]
43;
44  %transposed = call <8 x double> @llvm.matrix.transpose(<8 x double> %in, i32 4, i32 2)
45  store <8 x double> %transposed, <8 x double>* %Ptr, align 8
46  call void @foo(<8 x double> %transposed)
47  ret <8 x double> %transposed
48}
49
50declare <8 x double> @llvm.matrix.transpose(<8 x double>, i32, i32)
51
52declare void @foo(<8 x double>)
53