GPU virtual address space per process

Accelerated Computing CUDA CUDA Programming and Performance
1

Hello,

I know it would seems obvious but i want to make sure about it.
Does each processes have their own GPU virtual address space? like CPU-only processes have their own virtual address space per process?

Additionally is it possible that if we share the cuda context among processes, does processes share the view of virtual address space?

Thanks.

2

Yes, see here.

I don’t think there is any way to do that. It’s an imaginary idea.

1 Like