1; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
2; RUN: opt -lower-matrix-intrinsics -S < %s | FileCheck %s
3; RUN: opt -passes='lower-matrix-intrinsics' -S < %s | FileCheck %s
4
5define void @strided_store_3x2(<6 x float> %in, float* %out) {
6; CHECK-LABEL: @strided_store_3x2(
7; CHECK-NEXT:    [[SPLIT:%.*]] = shufflevector <6 x float> [[IN:%.*]], <6 x float> undef, <3 x i32> <i32 0, i32 1, i32 2>
8; CHECK-NEXT:    [[SPLIT1:%.*]] = shufflevector <6 x float> [[IN]], <6 x float> undef, <3 x i32> <i32 3, i32 4, i32 5>
9; CHECK-NEXT:    [[VEC_CAST:%.*]] = bitcast float* [[OUT:%.*]] to <3 x float>*
10; CHECK-NEXT:    store <3 x float> [[SPLIT]], <3 x float>* [[VEC_CAST]], align 4
11; CHECK-NEXT:    [[VEC_GEP:%.*]] = getelementptr float, float* [[OUT]], i64 5
12; CHECK-NEXT:    [[VEC_CAST2:%.*]] = bitcast float* [[VEC_GEP]] to <3 x float>*
13; CHECK-NEXT:    store <3 x float> [[SPLIT1]], <3 x float>* [[VEC_CAST2]], align 4
14; CHECK-NEXT:    ret void
15;
16  call void @llvm.matrix.column.major.store(<6 x float> %in, float* %out, i64 5, i1 false, i32 3, i32 2)
17  ret void
18}
19
20define void @strided_store_3x2_nonconst_stride(<6 x float> %in, i64 %stride, float* %out) {
21; CHECK-LABEL: @strided_store_3x2_nonconst_stride(
22; CHECK-NEXT:    [[SPLIT:%.*]] = shufflevector <6 x float> [[IN:%.*]], <6 x float> undef, <3 x i32> <i32 0, i32 1, i32 2>
23; CHECK-NEXT:    [[SPLIT1:%.*]] = shufflevector <6 x float> [[IN]], <6 x float> undef, <3 x i32> <i32 3, i32 4, i32 5>
24; CHECK-NEXT:    [[VEC_START:%.*]] = mul i64 0, [[STRIDE:%.*]]
25; CHECK-NEXT:    [[VEC_GEP:%.*]] = getelementptr float, float* [[OUT:%.*]], i64 [[VEC_START]]
26; CHECK-NEXT:    [[VEC_CAST:%.*]] = bitcast float* [[VEC_GEP]] to <3 x float>*
27; CHECK-NEXT:    store <3 x float> [[SPLIT]], <3 x float>* [[VEC_CAST]], align 4
28; CHECK-NEXT:    [[VEC_START2:%.*]] = mul i64 1, [[STRIDE]]
29; CHECK-NEXT:    [[VEC_GEP3:%.*]] = getelementptr float, float* [[OUT]], i64 [[VEC_START2]]
30; CHECK-NEXT:    [[VEC_CAST4:%.*]] = bitcast float* [[VEC_GEP3]] to <3 x float>*
31; CHECK-NEXT:    store <3 x float> [[SPLIT1]], <3 x float>* [[VEC_CAST4]], align 4
32; CHECK-NEXT:    ret void
33;
34  call void @llvm.matrix.column.major.store(<6 x float> %in, float* %out, i64 %stride, i1 false, i32 3, i32 2)
35  ret void
36}
37
38
39declare void @llvm.matrix.column.major.store(<6 x float>, float*, i64, i1, i32, i32)
40
41define void @strided_store_2x3(<10 x float> %in, float* %out) {
42; CHECK-LABEL: @strided_store_2x3(
43; CHECK-NEXT:    [[SPLIT:%.*]] = shufflevector <10 x float> [[IN:%.*]], <10 x float> undef, <2 x i32> <i32 0, i32 1>
44; CHECK-NEXT:    [[SPLIT1:%.*]] = shufflevector <10 x float> [[IN]], <10 x float> undef, <2 x i32> <i32 2, i32 3>
45; CHECK-NEXT:    [[SPLIT2:%.*]] = shufflevector <10 x float> [[IN]], <10 x float> undef, <2 x i32> <i32 4, i32 5>
46; CHECK-NEXT:    [[SPLIT3:%.*]] = shufflevector <10 x float> [[IN]], <10 x float> undef, <2 x i32> <i32 6, i32 7>
47; CHECK-NEXT:    [[SPLIT4:%.*]] = shufflevector <10 x float> [[IN]], <10 x float> undef, <2 x i32> <i32 8, i32 9>
48; CHECK-NEXT:    [[VEC_CAST:%.*]] = bitcast float* [[OUT:%.*]] to <2 x float>*
49; CHECK-NEXT:    store <2 x float> [[SPLIT]], <2 x float>* [[VEC_CAST]], align 4
50; CHECK-NEXT:    [[VEC_GEP:%.*]] = getelementptr float, float* [[OUT]], i64 4
51; CHECK-NEXT:    [[VEC_CAST5:%.*]] = bitcast float* [[VEC_GEP]] to <2 x float>*
52; CHECK-NEXT:    store <2 x float> [[SPLIT1]], <2 x float>* [[VEC_CAST5]], align 4
53; CHECK-NEXT:    [[VEC_GEP6:%.*]] = getelementptr float, float* [[OUT]], i64 8
54; CHECK-NEXT:    [[VEC_CAST7:%.*]] = bitcast float* [[VEC_GEP6]] to <2 x float>*
55; CHECK-NEXT:    store <2 x float> [[SPLIT2]], <2 x float>* [[VEC_CAST7]], align 4
56; CHECK-NEXT:    [[VEC_GEP8:%.*]] = getelementptr float, float* [[OUT]], i64 12
57; CHECK-NEXT:    [[VEC_CAST9:%.*]] = bitcast float* [[VEC_GEP8]] to <2 x float>*
58; CHECK-NEXT:    store <2 x float> [[SPLIT3]], <2 x float>* [[VEC_CAST9]], align 4
59; CHECK-NEXT:    [[VEC_GEP10:%.*]] = getelementptr float, float* [[OUT]], i64 16
60; CHECK-NEXT:    [[VEC_CAST11:%.*]] = bitcast float* [[VEC_GEP10]] to <2 x float>*
61; CHECK-NEXT:    store <2 x float> [[SPLIT4]], <2 x float>* [[VEC_CAST11]], align 4
62; CHECK-NEXT:    ret void
63;
64  call void @llvm.matrix.column.major.store.v10f32(<10 x float> %in, float* %out, i64 4, i1 false, i32 2, i32 5)
65  ret void
66}
67
68declare void @llvm.matrix.column.major.store.v10f32(<10 x float>, float*, i64, i1, i32, i32)
69