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
|
// RUN: mlir-opt %s -convert-vector-to-scf -lower-affine -convert-scf-to-cf -convert-vector-to-llvm="enable-amx" -finalize-memref-to-llvm -convert-func-to-llvm -reconcile-unrealized-casts | \
// RUN: mlir-translate -mlir-to-llvmir | \
// RUN: %lli --entry-function=entry --mattr="+amx-tile,+amx-int8,+amx-bf16" --dlopen=%mlir_c_runner_utils | \
// RUN: FileCheck %s
// Note: To run this test, your CPU must support AMX.
func.func @print(%arg0: memref<4x32xf32>) {
%fu = arith.constant -1.0: f32
%c0 = arith.constant 0: index
%c1 = arith.constant 1: index
%c4 = arith.constant 4: index
scf.for %i = %c0 to %c4 step %c1 {
%0 = vector.transfer_read %arg0[%i, %c0], %fu: memref<4x32xf32>, vector<32xf32>
vector.print %0 : vector<32xf32>
}
return
}
func.func @kernel(%arg0: memref<4x32xf32>) {
%c0 = arith.constant 0: index
%c2 = arith.constant 2 : index
%c4 = arith.constant 4 : index
%c16 = arith.constant 16 : index
%c32 = arith.constant 32 : index
scf.for %i = %c0 to %c4 step %c2 {
scf.for %j = %c0 to %c32 step %c16 {
%0 = amx.tile_zero : vector<2x16xf32>
amx.tile_store %arg0[%i, %j], %0 : memref<4x32xf32>, vector<2x16xf32>
func.call @print(%arg0) : (memref<4x32xf32>) -> ()
}
}
return
}
func.func @entry() -> i32 {
%f1 = arith.constant 1.0: f32
%c0 = arith.constant 0: index
%c1 = arith.constant 1: index
%c4 = arith.constant 4 : index
%c32 = arith.constant 32 : index
// Set up memory.
%a = memref.alloc() : memref<4x32xf32>
scf.for %i = %c0 to %c4 step %c1 {
scf.for %j = %c0 to %c32 step %c1 {
memref.store %f1, %a[%i, %j] : memref<4x32xf32>
}
}
// Call kernel.
func.call @kernel(%a) : (memref<4x32xf32>) -> ()
// Verify progress of blocked tilezero.
//
// CHECK: ( 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 )
// CHECK-NEXT: ( 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 )
// CHECK-NEXT: ( 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 )
// CHECK-NEXT: ( 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 )
//
// CHECK-NEXT: ( 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 )
// CHECK-NEXT: ( 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 )
// CHECK-NEXT: ( 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 )
// CHECK-NEXT: ( 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 )
//
// CHECK-NEXT: ( 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 )
// CHECK-NEXT: ( 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 )
// CHECK-NEXT: ( 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 )
// CHECK-NEXT: ( 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 )
//
// CHECK-NEXT: ( 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 )
// CHECK-NEXT: ( 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 )
// CHECK-NEXT: ( 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 )
// CHECK-NEXT: ( 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 )
//
// Release resources.
memref.dealloc %a : memref<4x32xf32>
%i0 = arith.constant 0 : i32
return %i0 : i32
}
|
