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
|
// RUN: mlir-opt %s \
// RUN: -gpu-lower-to-nvvm-pipeline="cubin-chip=sm_90a cubin-features=+ptx80 opt-level=3" \
// RUN: | mlir-cpu-runner \
// RUN: --shared-libs=%mlir_cuda_runtime \
// RUN: --shared-libs=%mlir_runner_utils \
// RUN: --shared-libs=%mlir_c_runner_utils \
// RUN: --entry-point-result=void \
// RUN: | FileCheck %s
// CHECK: clusterIdx: (1, 1, 0) in Cluster Dimension: (2, 2, 1) blockIdx: (3, 3, 0)
module attributes {gpu.container_module} {
func.func @main() {
%c1 = arith.constant 1 : index
%c2 = arith.constant 2 : index
%c4 = arith.constant 4 : index
gpu.launch_func @gpumodule::@kernel_cluster clusters in(%c2,%c2,%c1) blocks in (%c4, %c4, %c1) threads in (%c1, %c1, %c1)
return
}
gpu.module @gpumodule {
gpu.func @kernel_cluster() kernel attributes {gpu.known_block_size = array<i32: 1, 1, 1>, gpu.known_grid_size = array<i32: 2, 2, 1>} {
%cidX = gpu.cluster_id x
%cidY = gpu.cluster_id y
%cidZ = gpu.cluster_id z
%cdimX = gpu.cluster_dim_blocks x
%cdimY = gpu.cluster_dim_blocks y
%cdimZ = gpu.cluster_dim_blocks z
%bidX = gpu.block_id x
%bidY = gpu.block_id y
%bidZ = gpu.block_id z
%cidX_i32 = index.casts %cidX : index to i32
%cidY_i32 = index.casts %cidY : index to i32
%cidZ_i32 = index.casts %cidZ : index to i32
%cdimX_i32 = index.casts %cdimX : index to i32
%cdimY_i32 = index.casts %cdimY : index to i32
%cdimZ_i32 = index.casts %cdimZ : index to i32
%bidX_i32 = index.casts %bidX : index to i32
%bidY_i32 = index.casts %bidY : index to i32
%bidZ_i32 = index.casts %bidZ : index to i32
%c3 = arith.constant 3 : index
%cnd1 = arith.cmpi eq, %bidX, %c3 : index
%cnd2 = arith.cmpi eq, %bidY, %c3 : index
scf.if %cnd1 {
scf.if %cnd2 {
gpu.printf "clusterIdx: (%d, %d, %d) in Cluster Dimension: (%d, %d, %d) blockIdx: (%d, %d, %d) \n"
%cidX_i32,
%cidY_i32,
%cidZ_i32,
%cdimX_i32,
%cdimY_i32,
%cdimZ_i32,
%bidX_i32,
%bidY_i32,
%bidZ_i32
:
i32, i32, i32, i32, i32, i32, i32, i32, i32
}
}
gpu.return
}
}
}
|
