Process texture in CUDA and display result

Accelerated Computing CUDA CUDA Programming and Performance
1

I was wondering the way to process a texture in CUDA and then directly display it to the screen, without having to transfer the image data through host memory again.

In the sobel filter example provided with CUDA, they do the following to process a texture and display:

cutilSafeCall(cudaGraphicsMapResources(1, &cuda_pbo_resource, 0));
size_t num_bytes;
cutilSafeCall(cudaGraphicsResourceGetMappedPointer((void **)&data, &num_bytes,
cuda_pbo_resource));

// process texture and write result to data variable

glBindTexture(GL_TEXTURE_2D, texid);
glBindBuffer(GL_PIXEL_UNPACK_BUFFER, pbo_buffer);
glTexSubImage2D(GL_TEXTURE_2D, 0, 0, 0, imWidth, imHeight,
GL_LUMINANCE, GL_FLOAT, OFFSET(0));
glBindBuffer(GL_PIXEL_UNPACK_BUFFER, 0);

// display texture to screen by enabling GL_TEXTURE_2D, using GL_QUADS

Correct me if I’m wrong, but I believe that the allocated memory pointed to by the ‘data’ variable is in device memory, since we are just directly writing to it in the kernels, and since it is initialized with cudaGraphicsResourceGetMappedPointer.

So I’m thinking this is not going through host memory, but I’m not sure. Is this the best (ie. quickest) way to process a texture and display it directly without going through host memory? If not, how should I do this?

Also, could someone please explain the reason for using the pbo_buffer. Is this necessary?

Thanks a lot!

2

Could someone please comment on this so I have peace of mind about this?

Thanks!

3

Could someone please comment on this so I have peace of mind about this?

Thanks!

4

So are people not answering this thread because it’s a hard problem people don’t know how to solve or because it’s a really obvious question???

5

So are people not answering this thread because it’s a hard problem people don’t know how to solve or because it’s a really obvious question???

6

No, this code isn’t doing any transfers through host memory, although it is doing a copy in GPU memory.

The PBO is there because you can’t load an OpenGL texture directly from a CUDA device pointer.

7

No, this code isn’t doing any transfers through host memory, although it is doing a copy in GPU memory.

The PBO is there because you can’t load an OpenGL texture directly from a CUDA device pointer.

8

Thanks. So is the PBO sort of the openGL equivalent of a cuda array? You must use this to bind a texture to?

9

Thanks. So is the PBO sort of the openGL equivalent of a cuda array? You must use this to bind a texture to?

10

No, the PBO is just an abstraction of an array of bytes in memory, which the texture can be initialized from quickly.
[url=“http://www.opengl.org/registry/specs/ARB/pixel_buffer_object.txt”]http://www.opengl.org/registry/specs/ARB/p...ffer_object.txt[/url]

11

No, the PBO is just an abstraction of an array of bytes in memory, which the texture can be initialized from quickly.
[url=“http://www.opengl.org/registry/specs/ARB/pixel_buffer_object.txt”]http://www.opengl.org/registry/specs/ARB/p...ffer_object.txt[/url]