cuda实战chapter4_julia

结果

参考：书本。自己写了一遍。方法参考书本 

疑问：1：

运行书本例子显示the launch timed out and was terminated，参考网址

http://stackoverflow.com/questions/6185117/cudamemcpy-errorthe-launch-timed-out-and-was-terminated好像是单个网格数目不能负担1000*1000的并行数目？？  修改成200*200后，可以。  应该可以修改成多个GPU运行？？但现在不知道。不过直接增加每个线程块负担的数目应该可以。

2：参考http://blog.csdn.net/xizhibei/article/details/6664184，修改julia数据，但是显示的并不好。是不是颜色搭配问题？？

3：刚开始invalid argument，应该是数组没有设置好。a[(i*width+j)*6+m]才是正确的，但我a[i*width+y*6+m]，只检测了i=0的情况，是正确的。。悲剧

4：那开始用vbo绘制也不对，应该也是数组问题？？

5. 对C的指针不熟悉。

   怎么开多个线程网格？并运用？

   数组，最好多举例，实现。

6 opengl不熟悉。  把 glutCreateWindow()放在glutInitPosition前面了,一直显示初始化内容？？

直接cpu实现：

     

#include<iostream>
#include <time.h>
#include <gl/GLee.h>
#include <gl/glut.h>
#include <gl/GL.h>
#include<stdlib.h>

#include <math.h>
#define  width 1000
#define  height 1000

using namespace std;

void init(void )
{
	glClearColor(0,0,0,0);

	glShadeModel(GL_SMOOTH);

}


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)(width/2 - x)/(width/2);
	float jy = scale * (float)(height/2 - y)/(height/2);

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

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

	return 1;
}



void reshap(int w,int h)
{
	glViewport(0, 0, (GLsizei) w, (GLsizei) h);

	glMatrixMode(GL_PROJECTION);
	glLoadIdentity();
	//gluOrtho2D(-(GLdouble) w/2, (GLdouble) w/2, (GLdouble) -h/2, (GLdouble) h/2);
	gluOrtho2D(0, (GLdouble) w, 0, (GLdouble) h);

	glMatrixMode(GL_MODELVIEW);
	glLoadIdentity();
}


void keyboard(unsigned char key,int x,int y)
{
	switch(key)
	{
	case 27:
		exit(0);
		break;
	}
}


static float vertex_list[width*height][6];
static int index_list[width*height];
void kernel()
{
	
	for (int i=0;i!=width;++i)
	{
		for (int j=0;j!=height;++j)
		{
			vertex_list[i*width+j][0]=julia(i,j);

			vertex_list[i*width+j][1]=0;

			vertex_list[i*width+j][2]=0;//前三为颜色

			vertex_list[i*width+j][3]=i;

			vertex_list[i*width+j][4]=j;

			vertex_list[i*width+j][5]=0;//后三为坐标

		}
	}

	
	for (int i=0;i!=width*height;++i)
	{
		index_list[i]=i;
	}

}






void display(void )
{

	glClear(GL_COLOR_BUFFER_BIT);


	kernel();

	GLuint vbuffer=0;
	GLuint ibuffer=0;

	kernel();
	glGenBuffers(1,&vbuffer);
	glGenBuffers(1,&ibuffer);

	glBindBuffer(GL_ARRAY_BUFFER,vbuffer);
	glBufferData(GL_ARRAY_BUFFER,sizeof(vertex_list),vertex_list,GL_STATIC_DRAW);


	glBindBuffer(GL_ELEMENT_ARRAY_BUFFER,ibuffer);
	glBufferData(GL_ELEMENT_ARRAY_BUFFER,sizeof(index_list),index_list,GL_STATIC_DRAW);

	glEnableClientState(GL_COLOR_ARRAY);
	glEnableClientState(GL_VERTEX_ARRAY);

	glInterleavedArrays(GL_C3F_V3F,0,0);


	glDrawElements(GL_POINTS,width*height,GL_UNSIGNED_INT,NULL);
  
    glDrawElements(GL_POINTS,width*height,GL_UNSIGNED_INT,vertex_list);


	glutSwapBuffers();

}

int main(int argc,char** argv)
{
	clock_t start, end;

	start = clock();


	glutInit(&argc,argv);

	glutInitDisplayMode(GLUT_RGB|GLUT_DOUBLE);
	glutInitWindowSize(width,height);

	glutCreateWindow(argv[0]);

	init();
	glutReshapeFunc(reshap);
	glutDisplayFunc(display);
	glutKeyboardFunc(keyboard);

	end = clock();
	
	cout<<"Run time: "<<(double)(end - start) / CLOCKS_PER_SEC<<"S"<<endl;
	
    glutMainLoop();
	
	return 0;
}

gpu实现，但怎么修改，使1000*1000的可以/

#include<iostream>
#include <time.h>

#include<cuda.h>
//#include<gl/glee.h>
#include <gl/glut.h>

#include<stdlib.h>

#include <math.h>
#define  width 200
#define  height 200

//static float vertex_list[width*height][6];
 float vertex_list[width*height*6];
// float index_list[width*height];


using namespace std;

void init(void )
{
	glClearColor(0,0,0,0);

	glShadeModel(GL_SMOOTH);

}


struct cuComplex {
	float   r;
	float   i;
 __device__	cuComplex( float a, float b ) : r(a), i(b)  {}
 __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)(width/2 - x)/(width/2);
	float jy = scale * (float)(height/2 - y)/(height/2);

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

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

	return 1;
}



void reshap(int w,int h)
{
	glViewport(0, 0, (GLsizei) w, (GLsizei) h);

	glMatrixMode(GL_PROJECTION);
	glLoadIdentity();
	//gluOrtho2D(-(GLdouble) w/2, (GLdouble) w/2, (GLdouble) -h/2, (GLdouble) h/2);
	gluOrtho2D(0, (GLdouble) w, 0, (GLdouble) h);

	glMatrixMode(GL_MODELVIEW);
	glLoadIdentity();
}




void keyboard(unsigned char key,int x,int y)
{
	switch(key)
	{
	case 27:
		exit(0);
		break;
	}
}


//__global__ void kernel1(float a[width*height][6])
__global__ void kernel1(float *a)
{

	int i=blockIdx.x;
	int j=blockIdx.y;


			a[(i*width+j)*6]=julia(i,j);

			a[(i*width+j)*6+1]=0;

			a[(i*width+j)*6+2]=0;//前三为颜色

			a[(i*width+j)*6+3]=i;

			a[(i*width+j)*6+4]=j;

			a[(i*width+j)*6+5]=0;//后三为坐标
	
}


static void HandleError( cudaError_t err,
                         const char *file,
                         int line ) {
    if (err != cudaSuccess) {
        printf( "%s in %s at line %d\n", cudaGetErrorString( err ),
                file, line );
        exit( EXIT_FAILURE );
    }
}
#define HANDLE_ERROR( err ) (HandleError( err, __FILE__, __LINE__ ))



void display();
int main(int argc,char** argv)
{
	clock_t start, end;
	start = clock();



    float *cuda_vertex;


	//static float cuda_vertex[width*height][6];

 
	//float cuda_index_list[width*height];

	HANDLE_ERROR(cudaMalloc((void**)&cuda_vertex,6*width*height*sizeof(float)));
 //HANDLE_ERROR(cudaMalloc((void**)&cuda_vertex,sizeof(cuda_vertex)));

	
	dim3 grid(width,height);

	kernel1<<<grid,1>>>(cuda_vertex);

	HANDLE_ERROR(cudaMemcpy(vertex_list,cuda_vertex,sizeof(float)*width*height*6,cudaMemcpyDeviceToHost));


	glutInit(&argc,argv);

	glutInitDisplayMode(GLUT_RGB|GLUT_DOUBLE);
	glutInitWindowSize(width,height);

	glutCreateWindow(argv[0]);

	init();	


	glutReshapeFunc(reshap);
	glutDisplayFunc(display);
	glutKeyboardFunc(keyboard);

	 HANDLE_ERROR(cudaFree(cuda_vertex));


	end = clock();
	cout<<"Run time: "<<(double)(end - start) / CLOCKS_PER_SEC<<"S"<<endl;
    glutMainLoop();
	
	return 0;
}


void display()
{

	glClear(GL_COLOR_BUFFER_BIT);

		glBegin(GL_POINTS);
	

	for(int i=0;i!=width;++i)
		{
			for(int j=0;j!=height;++j)
			{

				//glColor3f(1,0,0);

             	//glColor3f(vertex_list[i*width+6*j+0],0,0);
               glColor3f(vertex_list[(i*width+j)*6+0],vertex_list[(i*width+j)*6+1],vertex_list[(i*width+j)*6+2]);

               glVertex2i(vertex_list[(i*width+j)*6+3],vertex_list[(i*width+j)*6+4]);
			   //glVertex2i(i,j);
			}
		}
	glEnd();


	glutSwapBuffers();

}