Lab3 CUDA

lab3.cu

/*
Copyright 2017, Paweł Czarnul pawelczarnul@pawelczarnul.com

Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are
met:

1. Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.

2. Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.

3. Neither the name of the copyright holder nor the names of its
contributors may be used to endorse or promote products derived from
this software without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/

#include <stdio.h>
#include <stdlib.h>
#include <cuda_runtime.h>
#include <sys/time.h>
#include <curand.h>
#include <curand_kernel.h>

#define PRECISION 0.000001
#define SHOTS_PER_RANGE 10000

__global__ 
void initRNG(curandState *state){

  int idx = threadIdx.x+blockDim.x*blockIdx.x;
  curand_init(1234, idx, 0, &state[idx]);
}

__host__
void errorexit(const char *s) {
    printf("\n%s",s);
    exit(EXIT_FAILURE);
}

__device__ 
void float_rand(curandState d_state, double min, double max, double* result){
    double uniform_rand = curand_uniform_double(&d_state); /* [0, 1.0] */
    *result = (double) (min + uniform_rand * ( max - min )); /* [min, max] */
}

__device__ 
void f (double x, double* result){
    *result = sinf(x) * sinf(x); 
}
__device__ 
void monte_carlo(double hit, double shots, double a, double b, double* surface){
    double hit_shots_ratio =(double)hit/(double)shots;
    double square_surface = b-a;
    *surface = hit_shots_ratio * square_surface;
}

__global__ 
void simpleIntegration(curandState *const rngStates, double *result) {
    long my_index=blockIdx.x*blockDim.x+threadIdx.x;
    double a = (double) my_index;
    double b = a + 1.0;
    unsigned long start=my_index;
    int hit = 0;
    
    
    int page = 0;

    for(int i = 0; i < SHOTS_PER_RANGE; i++){
        double rand_x = 0.0;
        
        float_rand(rngStates[i+page] ,a, b, &rand_x);

        double sin_value = 0.0;
        f(rand_x, &sin_value);
        double random_y = 0.0;

        float_rand(rngStates[i+page] ,0.0, 1.0, &random_y);
        
        if(sin_value <= random_y){
            hit++;
        }     
    }

    double surface = 0.0;
    monte_carlo(hit,SHOTS_PER_RANGE,a,b, &surface);
    double surface_final = surface; 
    double difference = 0;
    int iterations = 0;
    do{
        page++;
        iterations++;
        int additional_shots = (int) ((float) SHOTS_PER_RANGE * 0.1f);
        int additional_hits = 0;
        for(int i = 0; i < additional_shots; i++){
            double rand_x = 0.0;
            
            float_rand(rngStates[i+page], a, b, &rand_x);
            
            double sin_value = 0.0;
            f(rand_x, &sin_value);
            double random_y = 0.0;

            float_rand(rngStates[i+page], 0.0, 1.0, &random_y);
            
            if(sin_value <= random_y){
                additional_hits++;
            }
        }
        double surface_with_adition = 0.0;
        
        monte_carlo(hit + additional_hits, SHOTS_PER_RANGE + additional_shots, a, b, &surface_with_adition);
        difference = surface_with_adition / additional_hits;
        surface_final = surface_with_adition;

        iterations++;
    }while(difference > 0.01f);
    
    
    result[my_index]=surface_final;
}



__host__
void printtime(struct timeval *start,struct timeval *stop) {
  long time=1000000*(stop->tv_sec-start->tv_sec)+stop->tv_usec-start->tv_usec;

  printf("\n%ld microseconds\n",time);

}

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


    struct timeval start,stop,start1,stop1;
    double result;
    int threadsinblock=100;
    int blocksingrid=100;	
    long threadcount=threadsinblock * blocksingrid;

    

    
    gettimeofday(&start,NULL);

    long size=threadcount*sizeof(double);
    double *hresults=(double *)malloc(size);
    if (!hresults) errorexit("Error allocating memory on the host");	

    double *dresults=NULL;
    if (cudaSuccess!=cudaMalloc((void **)&dresults,size))
        errorexit("Error allocating memory on the GPU");

    // Allocate memory for RNG states
    curandState *d_rngStates = 0;
    if (cudaSuccess!=cudaMalloc((void **)&d_rngStates, SHOTS_PER_RANGE * 100 * sizeof(curandState)))
        errorexit("Error allocating memory for curandState");

    gettimeofday(&start1,NULL);

    initRNG<<<blocksingrid, threadsinblock>>>(d_rngStates);

    // start computations on the GPU
    simpleIntegration<<<blocksingrid,threadsinblock>>>(d_rngStates, dresults);
    if (cudaSuccess!=cudaGetLastError())
      errorexit("Error during kernel launch");

    if (cudaSuccess!=cudaMemcpy(hresults,dresults,size,cudaMemcpyDeviceToHost))
       errorexit("Error copying results");

    cudaDeviceSynchronize();
    gettimeofday(&stop1,NULL);


    double total_surface = 0.0;
    for(long i=0;i<threadcount;i++)
        total_surface += hresults[i];

    gettimeofday(&stop,NULL);

    printf("\nThe final result is %f\n",total_surface);
    printtime(&start1,&stop1);
    printtime(&start,&stop);

    // release resources
    free(hresults);
    if (cudaSuccess!=cudaFree(dresults))
      errorexit("Error when deallocating space on the GPU");

}