source: CIVL/examples/cuda/cuda-omp.cu@ 83af34d

//http://www.arc.vt.edu/resources/software/cuda/docs/cuda-omp.cu

#include <omp.h>
#include <cuda.h>
#include <stdio.h>      
#include <stdlib.h>      

// A kernel that increments each array element by the value b

__global__ void kernelAddConstant(int *g_a, const int b)
{
        int idx = blockIdx.x * blockDim.x + threadIdx.x;
        g_a[idx] += b;
}

// Check whether each element was incremented by the value b
int correctResult(int *data, const int n, const int b)
{
        for(int i = 0; i < n; i++)
                if(data[i] != i + b)
                        return 0;
        return 1;
}

int main(int argc, char *argv[])
{

        // Variable which holds number of GPUs
        int num_gpus = 0;   

        // Determine the number of CUDA capable GPUs
        cudaGetDeviceCount(&num_gpus);
        if(num_gpus < 1)
        {
                printf("No CUDA Capable GPU(s) Detected \n");
                return 1;
        }

        // Display the CPU and GPU processor specification         
        int num_procs = omp_get_num_procs();
        printf("number of host CPUs:\t%d\n", num_procs);
        printf("number of CUDA devices:\t%d\n", num_gpus);
        for(int i = 0; i < num_gpus; i++)
        {
                _kernelStatus tmp;
                cudaDeviceProp dprop;
                cudaGetDeviceProperties(&dprop, i);
                printf("\t Device %d is a %s\n", i, dprop.name);
        }


        // Initialize the variables 
        unsigned int n = num_gpus * 8192;
        unsigned int nbytes = n * sizeof(int);
        int *a = 0;             // pointer to data on the CPU
        int b = 3;              // value by which each array array element will be incremented
        a = (int*)malloc(nbytes);
        
        if(0 == a)
        {
                printf("couldn't allocate CPU memory\n");
                return 1;
        }
        
        for(unsigned int i = 0; i < n; i++)
                a[i] = i;
    
        // Set the number of threads to the number of GPUs on the system
        omp_set_num_threads(num_gpus);

        #pragma omp parallel
        {
                unsigned int cpu_thread_id = omp_get_thread_num();
                unsigned int num_cpu_threads = omp_get_num_threads();

                // Assign and check the GPU device for each thread
                int gpu_id = -1;
                cudaSetDevice(cpu_thread_id % num_gpus);        
                cudaGetDevice(&gpu_id);

                printf("CPU thread %d (of %d) uses CUDA device %d\n", cpu_thread_id, num_cpu_threads, gpu_id);

                // Variable on the device associated with this CPU thread
                int *d_a = 0; 

                // Variable for the CPU
                int *sub_a = a + cpu_thread_id * n / num_cpu_threads;
   
                unsigned int nbytes_per_kernel = nbytes / num_cpu_threads;
                dim3 gpu_threads = {128, 1, 1};  // 128 threads per block
                dim3 gpu_blocks = {(n / (gpu_threads.x * num_cpu_threads)), 1, 1};

                //Allocate memory on the device
                cudaMalloc((void**)&d_a, nbytes_per_kernel);

                //Initialize the array on the device with zeros
                cudaMemset(d_a, 0, nbytes_per_kernel);

                //Copy data from host to device
                cudaMemcpy(d_a, sub_a, nbytes_per_kernel, cudaMemcpyHostToDevice);
        
                //Launch the kernel
                kernelAddConstant<<<gpu_blocks, gpu_threads>>>(d_a, b);

                //Copy the result  from the device to the host
                cudaMemcpy(sub_a, d_a, nbytes_per_kernel, cudaMemcpyDeviceToHost);
          
                //Deallocate the memory on the device
                cudaFree(d_a);

        }
        

        if(cudaSuccess != cudaGetLastError()) {
                int err_num = cudaGetLastError();
                const char * err_str = cudaGetErrorString(err_num);
                printf("%s\n", err_str);
        }


        //Check for correctness of the result
        if(correctResult(a, n, b))
                printf("Test PASSED\n");
        else
                printf("Test FAILED\n");

        //Deallocate the CPU memory 
        free(a);    

        // deprecated
        // cudaThreadExit();

        return 0;
}