1// RUN: mlir-cuda-runner %s --shared-libs=%cuda_wrapper_library_dir/libcuda-runtime-wrappers%shlibext,%linalg_test_lib_dir/libmlir_runner_utils%shlibext --entry-point-result=void | FileCheck %s
2
3// CHECK: [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12]
4func @main() {
5  %arg = alloc() : memref<13xi32>
6  %dst = memref_cast %arg : memref<13xi32> to memref<?xi32>
7  %one = constant 1 : index
8  %c0 = constant 0 : index
9  %sx = dim %dst, %c0 : memref<?xi32>
10  %cast_dst = memref_cast %dst : memref<?xi32> to memref<*xi32>
11  gpu.host_register %cast_dst : memref<*xi32>
12  gpu.launch blocks(%bx, %by, %bz) in (%grid_x = %one, %grid_y = %one, %grid_z = %one)
13             threads(%tx, %ty, %tz) in (%block_x = %sx, %block_y = %one, %block_z = %one) {
14    %t0 = index_cast %tx : index to i32
15    store %t0, %dst[%tx] : memref<?xi32>
16    gpu.terminator
17  }
18  gpu.launch blocks(%bx, %by, %bz) in (%grid_x = %one, %grid_y = %one, %grid_z = %one)
19             threads(%tx, %ty, %tz) in (%block_x = %sx, %block_y = %one, %block_z = %one) {
20    %t0 = index_cast %tx : index to i32
21    store %t0, %dst[%tx] : memref<?xi32>
22    gpu.terminator
23  }
24  call @print_memref_i32(%cast_dst) : (memref<*xi32>) -> ()
25  return
26}
27
28func private @print_memref_i32(%memref : memref<*xi32>)
29