CUDA學習筆記（三）CUDA並行編程2

CUDA並行編程

教材《GPU高性能編程CUDA實戰》第四章 CUDA C並行編程 基於CPU和GPU的Julia集對比

基於CPU的Julia集

#include "book.h"
#include "cpu_bitmap.h"

#define DIM 1000

struct cuComplex {
    float   r;
    float   i;
    cuComplex( float a, float b ) : r(a), i(b)  {}
    float magnitude2( void ) { return r * r + i * i; }
    cuComplex operator*(const cuComplex& a) {
        return cuComplex(r*a.r - i*a.i, i*a.r + r*a.i);
    }
    cuComplex operator+(const cuComplex& a) {
        return cuComplex(r+a.r, i+a.i);
    }
};

int julia( int x, int y ) { 
    const float scale = 1.5;
    float jx = scale * (float)(DIM/2 - x)/(DIM/2);
    float jy = scale * (float)(DIM/2 - y)/(DIM/2);

    cuComplex c(-0.8, 0.156);
    cuComplex a(jx, jy);

    int i = 0;
    for (i=0; i<200; i++) {
        a = a * a + c;
        if (a.magnitude2() > 1000)
            return 0;
    }

    return 1;
}

void kernel( unsigned char *ptr ){
    for (int y=0; y<DIM; y++) {
        for (int x=0; x<DIM; x++) {
            int offset = x + y * DIM;

            int juliaValue = julia( x, y );
            ptr[offset*4 + 0] = 255 * juliaValue;
            ptr[offset*4 + 1] = 0;
            ptr[offset*4 + 2] = 0;
            ptr[offset*4 + 3] = 255;
        }
    }
 }

int main( void ) {
    CPUBitmap bitmap( DIM, DIM );
    unsigned char *ptr = bitmap.get_ptr();

    kernel( ptr );

    bitmap.display_and_exit();
}

基於GPU的Julia集

#include "cuda_runtime.h"
#include "device_launch_parameters.h"//包含blockIdx.x
#include "book.h"
#include"cpu_bitmap.h"

#define DIM 1000
struct cuComplex {
    float r;
    float i;
    __device__ cuComplex(float a,float b):r(a),i(b){}//前面要加__device__否則會報錯

    __device__ float magnitude2(void) {
        return r*r + i*i;
    }

    __device__ cuComplex operator*(const cuComplex& a) {
        return cuComplex(r*a.r - i*a.i, i*a.r + r*a.i);
    }

    __device__ cuComplex operator+(const cuComplex& a) {
        return cuComplex(r+a.r,i+a.i);
    }
};
__device__ int julia(int x, int y) {
    const float scale = 1.5;
    float jx = scale*(float)(DIM / 2 - x) / (DIM / 2);
    float jy = scale*(float)(DIM / 2 - y) / (DIM / 2);

    cuComplex c(-0.8, 0.156);
    cuComplex a(jx, jy);
    int i = 0;
    for (i = 0; i < 200; i++) {
        a = a*a + c;
        if (a.magnitude2() > 1000)
            return 0;
    }
    return 1;
}

__global__ void kernel(unsigned char *ptr) {
    //將threadIdx/BlockIdx映射到像素位置
    int x = blockIdx.x;
    int y = blockIdx.y;
    int offset = x + y*gridDim.x;
    //現在計算這個位置上的值
    int juliaValue = julia(x, y);
    ptr[offset * 4 + 0] = 255 * juliaValue;
    ptr[offset * 4 + 1] = 0;
    ptr[offset * 4 + 2] = 0;
    ptr[offset * 4 + 3] = 255;
}

int main() {
    CPUBitmap bitmap(DIM, DIM);
    unsigned char *dev_bitmap;

    HANDLE_ERROR(cudaMalloc((void**)&dev_bitmap, bitmap.image_size()));
    dim3 grid(DIM, DIM);
    kernel << <grid, 1 >> >(dev_bitmap);

    HANDLE_ERROR(cudaMemcpy(bitmap.get_ptr(), dev_bitmap, bitmap.image_size(), cudaMemcpyDeviceToHost));
    HANDLE_ERROR(cudaFree(dev_bitmap));

    bitmap.display_and_exit();
}

注意：構造函數__device__ cuComplex(float a,float b):r(a),i(b){} 前面要加_device__，否則會報錯

 從device上調用host的函數是不允許的。