关于CUDA分配线程块的问题

#include <stdio.h>

__global__ void computeArray(int* array, int size)
{
    int i = blockIdx.x * blockDim.x + threadIdx.x;
    int j = blockIdx.y * blockDim.y + threadIdx.y;
    int k = blockIdx.z * blockDim.z + threadIdx.z;

    // Check if the thread is within the array boundaries
    if (i < size && j < size && k < size)
    {
        // Perform computation on the array element
        int index = i + j*size + k*size*size;
        array[index] = array[index] + 1;
    }
}

int main()
{
    int size = 10;
    int arraySize = size * size * size;
    int* array;

    // Allocate memory for the array on the GPU
    cudaMalloc(&array, arraySize * sizeof(int));

    // Set the array values to 0
    cudaMemset(array, 0, arraySize * sizeof(int));

    // Define the thread block size
    dim3 blockSize(8, 8, 8);

    // Calculate the number of thread blocks in each dimension
    dim3 gridSize((size + blockSize.x - 1) / blockSize.x,
                  (size + blockSize.y - 1) / blockSize.y,
                  (size + blockSize.z - 1) / blockSize.z);

    // Launch the kernel
    computeArray<<<gridSize, blockSize>>>(array, size);

    // Wait for the kernel to finish
    cudaDeviceSynchronize();

    // Print the results
    for (int i = 0; i < size; i++) {
        for (int j = 0; j < size; j++) {
            for (int k = 0; k < size; k++) {
                int index = i + j*size + k*size*size;
                printf("%d ", array[index]);
            }
            printf("\n");
        }
        printf("\n");
    }

    // Free the memory on the GPU
    cudaFree(array);

    return 0;
}