DevConf.us 2018 is the 1st annual, free, Red Hat sponsored technology conference for community project and professional contributors to Free and Open Source technologies held at the Boston University in the historic city of Boston, USA.
When: Friday, August 17 to Sunday, August 19, 2018
This session will cover what is a memory address space, what it means to have different memory address space between CPU and GPU. A familiarity with existing compute API such as OpenCL or CUDA is expected (https://www.khronos.org/opencl/ or https://developer.nvidia.com/cuda-zone)
Threads on the GPU have their own memory address space, distinct from the CPU memory address space. This is one of the biggest issue for programmers wishing to leverage GPU. Programmers have to replicate their data structure in the GPU memory address space. New Linux kernel features and hardware features are now breaking this boundary and offer a unified memory address space for CPU and GPU. Programmers can share pointers between the two without having to do anything.
