Hello folks,
lately I’ve been working on a very simple CUDA/OpenGL program, and I’m going mad to figure out what’s wrong with it.
If I run my program in device emulation, I get the correct result:
[attachment=6317:attachment]
But when I run in hardware mode (i.e. non emulated), I get this seemingly senseless output:
[attachment=6318:attachment]
As I said, the code is very, very simple:
#include <GL/glew.h>
#include <GL/glut.h>
#include <cuda.h>
#include <cuda_runtime.h>
#include <cuda_gl_interop.h>
#define RESOLUTION_X 800
#define RESOLUTION_Y 600
typedef unsigned int uint;
uint buffer;
uint tex;
uchar4 *data;
void idle();
void draw();
void initGLUT(int argc, char** argv)
{
glutInit( &argc, argv );
glutInitDisplayMode( GLUT_RGBA | GLUT_DOUBLE );
glutInitWindowSize( RESOLUTION_X, RESOLUTION_Y );
glutCreateWindow( "AlphaVox" );
glutIdleFunc( idle );
glutDisplayFunc( draw );
}
void initGLEW()
{
glewInit();
if ( !glewIsSupported( "GL_EXT_pixel_buffer_object" ) )
exit(1);
}
void initGL()
{
int size;
glGenBuffers( 1, &buffer );
glGenTextures( 1, &tex );
glBindBuffer( GL_PIXEL_UNPACK_BUFFER, buffer );
glBufferData( GL_PIXEL_UNPACK_BUFFER, RESOLUTION_X * RESOLUTION_Y * 4, NULL, GL_DYNAMIC_DRAW );
glGetBufferParameteriv( GL_PIXEL_UNPACK_BUFFER, GL_BUFFER_SIZE, &size );
glBindBuffer( GL_PIXEL_UNPACK_BUFFER, 0 );
glBindTexture( GL_TEXTURE_2D, tex );
glTexParameteri( GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST );
glTexParameteri( GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST );
glEnable( GL_TEXTURE_2D );
cudaGLRegisterBufferObject( buffer );
}
void idle() {
glutPostRedisplay();
}
__global__ void sampleDrawing(uchar4* output)
{
int y = blockIdx.y * blockDim.y + threadIdx.y;
int x = blockIdx.x * blockDim.x + threadIdx.x;
__syncthreads();
output[ y * RESOLUTION_X + x ] = make_uchar4( threadIdx.x * 16, threadIdx.y * 32, 0, 1 );
}
void draw()
{
dim3 blockSize( 16, 8 );
dim3 gridSize( RESOLUTION_X / 16, RESOLUTION_Y / 8 );
cudaGLMapBufferObject( (void**) &data, buffer );
sampleDrawing<<< gridSize, blockSize >>>( data );
cudaGLUnmapBufferObject( buffer );
glBindBuffer( GL_PIXEL_UNPACK_BUFFER, buffer );
glTexImage2D( GL_TEXTURE_2D, 0, GL_RGBA, RESOLUTION_X, RESOLUTION_Y, 0, GL_RGBA, GL_UNSIGNED_BYTE, 0 );
glBindBuffer( GL_PIXEL_UNPACK_BUFFER, 0 );
glBegin( GL_QUADS );
glTexCoord2i( 0, 0 ); glVertex2i( -1, -1 );
glTexCoord2i( 1, 0 ); glVertex2i( 1, -1 );
glTexCoord2i( 1, 1 ); glVertex2i( 1, 1 );
glTexCoord2i( 0, 1 ); glVertex2i( -1, 1 );
glEnd();
glutSwapBuffers();
}
void shutdown()
{
cudaGLUnregisterBufferObject( buffer );
glDeleteBuffers( 1, &buffer );
}
int main(int argc, char** argv)
{
initGLUT( argc, argv );
initGLEW();
initGL();
glutMainLoop();
shutdown();
}
My computer is a Sony Vaio laptop with a 8400M GT device, with 128 MB of dedicated memory. I know it’s quite a low amount of memory (no wonder the majority of SDK samples doesn’t run properly), but it shouldn’t affect this very basic application.
