1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
|
// DEFINE: %{option} = enable-runtime-library=true
// DEFINE: %{compile} = mlir-opt %s --sparse-compiler=%{option}
// DEFINE: %{run} = mlir-cpu-runner \
// DEFINE: -e entry -entry-point-result=void \
// DEFINE: -shared-libs=%mlir_c_runner_utils | \
// DEFINE: FileCheck %s
//
// RUN: %{compile} | %{run}
//
// Do the same run, but now with direct IR generation.
// REDEFINE: %{option} = enable-runtime-library=false
// RUN: %{compile} | %{run}
// Do the same run, but now with direct IR generation and, if available, VLA
// vectorization.
// REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
// REDEFINE: --entry-function=entry_lli \
// REDEFINE: --extra-module=%S/Inputs/main_for_lli.ll \
// REDEFINE: %VLA_ARCH_ATTR_OPTIONS \
// REDEFINE: --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
// REDEFINE: FileCheck %s
// RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
#CSR = #sparse_tensor.encoding<{ lvlTypes = [ "dense", "compressed" ] }>
#trait_scale = {
indexing_maps = [
affine_map<(i,j) -> (i,j)> // X (out)
],
iterator_types = ["parallel", "parallel"],
doc = "X(i,j) = X(i,j) * 2"
}
//
// Integration test that lowers a kernel annotated as sparse to actual sparse
// code, initializes a matching sparse storage scheme from a dense tensor,
// and runs the resulting code with the JIT compiler.
//
module {
//
// A kernel that scales a sparse matrix A by a factor of 2.0.
//
func.func @sparse_scale(%argx: tensor<8x8xf32, #CSR>) -> tensor<8x8xf32, #CSR> {
%c = arith.constant 2.0 : f32
%0 = linalg.generic #trait_scale
outs(%argx: tensor<8x8xf32, #CSR>) {
^bb(%x: f32):
%1 = arith.mulf %x, %c : f32
linalg.yield %1 : f32
} -> tensor<8x8xf32, #CSR>
return %0 : tensor<8x8xf32, #CSR>
}
//
// Main driver that converts a dense tensor into a sparse tensor
// and then calls the sparse scaling kernel with the sparse tensor
// as input argument.
//
func.func @entry() {
%c0 = arith.constant 0 : index
%f0 = arith.constant 0.0 : f32
// Initialize a dense tensor.
%0 = arith.constant dense<[
[1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 1.0],
[0.0, 2.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
[0.0, 0.0, 3.0, 0.0, 0.0, 0.0, 0.0, 0.0],
[0.0, 0.0, 0.0, 4.0, 0.0, 0.0, 0.0, 0.0],
[0.0, 1.0, 0.0, 0.0, 5.0, 0.0, 0.0, 0.0],
[0.0, 1.0, 1.0, 0.0, 0.0, 6.0, 0.0, 0.0],
[0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 7.0, 1.0],
[0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 8.0]
]> : tensor<8x8xf32>
// Convert dense tensor to sparse tensor and call sparse kernel.
%1 = sparse_tensor.convert %0 : tensor<8x8xf32> to tensor<8x8xf32, #CSR>
%2 = call @sparse_scale(%1)
: (tensor<8x8xf32, #CSR>) -> tensor<8x8xf32, #CSR>
// Print the resulting compacted values for verification.
//
// CHECK: ( 2, 2, 2, 4, 6, 8, 2, 10, 2, 2, 12, 2, 14, 2, 2, 16 )
//
%m = sparse_tensor.values %2 : tensor<8x8xf32, #CSR> to memref<?xf32>
%v = vector.transfer_read %m[%c0], %f0: memref<?xf32>, vector<16xf32>
vector.print %v : vector<16xf32>
// Release the resources.
bufferization.dealloc_tensor %1 : tensor<8x8xf32, #CSR>
return
}
}
|
