01-18-2007, 09:21 AM
i have recently been developing application in linux, specifically a imageviewer application, and since all i had to do was blit a 2d image to the screen i programmed it entirely in xlib using the dbe and mit-shm extension. it works very well. Now the reason for this app was in a rendering cluster, machines(including the same one as the displaying machine) would volume render a piece of the total image and send it on to the machine that was the display. The rendering code utilizes opengl and glx. Now if I run the display on a different machine without a rendering code running, the imageviewer app works very well. The problems comes about when I have an instance of the rendering code up, my imageviewer app is no longer double buffered, but appears to be copied into the frame buffer. I have tried many ideas to correct this, as I thought that maybe the rendering code was using to much card memory and booting imageviewer out of the card. Now I did not write the rendering code, but the developer insists that he is reducing the memory footprint of this app, but the problem persists. I would very much like to hear any suggestions or insight into my problem. Thanks in advance.