1; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
2; RUN: opt -passes='lower-matrix-intrinsics' -S < %s | FileCheck %s
3
4define <9 x float> @strided_load_3x3(float* %in, i64 %stride) {
5; CHECK-LABEL: @strided_load_3x3(
6; CHECK-NEXT:  entry:
7; CHECK-NEXT:    [[VEC_START:%.*]] = mul i64 0, [[STRIDE:%.*]]
8; CHECK-NEXT:    [[VEC_GEP:%.*]] = getelementptr float, float* [[IN:%.*]], i64 [[VEC_START]]
9; CHECK-NEXT:    [[VEC_CAST:%.*]] = bitcast float* [[VEC_GEP]] to <3 x float>*
10; CHECK-NEXT:    [[COL_LOAD:%.*]] = load <3 x float>, <3 x float>* [[VEC_CAST]], align 4
11; CHECK-NEXT:    [[VEC_START1:%.*]] = mul i64 1, [[STRIDE]]
12; CHECK-NEXT:    [[VEC_GEP2:%.*]] = getelementptr float, float* [[IN]], i64 [[VEC_START1]]
13; CHECK-NEXT:    [[VEC_CAST3:%.*]] = bitcast float* [[VEC_GEP2]] to <3 x float>*
14; CHECK-NEXT:    [[COL_LOAD4:%.*]] = load <3 x float>, <3 x float>* [[VEC_CAST3]], align 4
15; CHECK-NEXT:    [[VEC_START5:%.*]] = mul i64 2, [[STRIDE]]
16; CHECK-NEXT:    [[VEC_GEP6:%.*]] = getelementptr float, float* [[IN]], i64 [[VEC_START5]]
17; CHECK-NEXT:    [[VEC_CAST7:%.*]] = bitcast float* [[VEC_GEP6]] to <3 x float>*
18; CHECK-NEXT:    [[COL_LOAD8:%.*]] = load <3 x float>, <3 x float>* [[VEC_CAST7]], align 4
19; CHECK-NEXT:    [[TMP0:%.*]] = shufflevector <3 x float> [[COL_LOAD]], <3 x float> [[COL_LOAD4]], <6 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5>
20; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <3 x float> [[COL_LOAD8]], <3 x float> poison, <6 x i32> <i32 0, i32 1, i32 2, i32 undef, i32 undef, i32 undef>
21; CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <6 x float> [[TMP0]], <6 x float> [[TMP1]], <9 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8>
22; CHECK-NEXT:    ret <9 x float> [[TMP2]]
23;
24entry:
25  %load = call <9 x float> @llvm.matrix.column.major.load(float* %in, i64 %stride, i1 false, i32 3, i32 3)
26  ret <9 x float> %load
27}
28
29declare <9 x float> @llvm.matrix.column.major.load(float*, i64, i1, i32, i32)
30
31define <9 x float> @strided_load_9x1(float* %in, i64 %stride) {
32; CHECK-LABEL: @strided_load_9x1(
33; CHECK-NEXT:  entry:
34; CHECK-NEXT:    [[VEC_START:%.*]] = mul i64 0, [[STRIDE:%.*]]
35; CHECK-NEXT:    [[VEC_GEP:%.*]] = getelementptr float, float* [[IN:%.*]], i64 [[VEC_START]]
36; CHECK-NEXT:    [[VEC_CAST:%.*]] = bitcast float* [[VEC_GEP]] to <9 x float>*
37; CHECK-NEXT:    [[COL_LOAD:%.*]] = load <9 x float>, <9 x float>* [[VEC_CAST]], align 4
38; CHECK-NEXT:    ret <9 x float> [[COL_LOAD]]
39;
40entry:
41  %load = call <9 x float> @llvm.matrix.column.major.load(float* %in, i64 %stride, i1 false, i32 9, i32 1)
42  ret <9 x float> %load
43}
44
45declare <8 x float> @llvm.matrix.column.major.load.v8f32(float*, i64, i1, i32, i32)
46
47define <8 x float> @strided_load_4x2(float* %in, i64 %stride) {
48; CHECK-LABEL: @strided_load_4x2(
49; CHECK-NEXT:  entry:
50; CHECK-NEXT:    [[VEC_START:%.*]] = mul i64 0, [[STRIDE:%.*]]
51; CHECK-NEXT:    [[VEC_GEP:%.*]] = getelementptr float, float* [[IN:%.*]], i64 [[VEC_START]]
52; CHECK-NEXT:    [[VEC_CAST:%.*]] = bitcast float* [[VEC_GEP]] to <4 x float>*
53; CHECK-NEXT:    [[COL_LOAD:%.*]] = load <4 x float>, <4 x float>* [[VEC_CAST]], align 4
54; CHECK-NEXT:    [[VEC_START1:%.*]] = mul i64 1, [[STRIDE]]
55; CHECK-NEXT:    [[VEC_GEP2:%.*]] = getelementptr float, float* [[IN]], i64 [[VEC_START1]]
56; CHECK-NEXT:    [[VEC_CAST3:%.*]] = bitcast float* [[VEC_GEP2]] to <4 x float>*
57; CHECK-NEXT:    [[COL_LOAD4:%.*]] = load <4 x float>, <4 x float>* [[VEC_CAST3]], align 4
58; CHECK-NEXT:    [[TMP0:%.*]] = shufflevector <4 x float> [[COL_LOAD]], <4 x float> [[COL_LOAD4]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
59; CHECK-NEXT:    ret <8 x float> [[TMP0]]
60;
61entry:
62  %load = call <8 x float> @llvm.matrix.column.major.load.v8f32(float* %in, i64 %stride, i1 false, i32 4, i32 2)
63  ret <8 x float> %load
64}
65