I want to be able to share a GPU memory address between two applications and have them both be able to read from and write to that address. In my quick hackings, I can’t seem to read what’s being written to the GPU memory. Is this something that’s wrong with my code or is it because I’m wanting to do something that can’t be done?
The GPU has memory protection schemes. One context cannot read another context’s global memory.
That’s what I was afraid of. I guess I’ll have to find a way around that. Thanks for your help.
I’m reading through the 2.1 Programming guide and it talks about contexts and how threads can attach and detach from them. Do you know if this would do what I mentioned above? Have one process write and detach from the context, then have another attach, read, and detach?