Get Started With CUDA Memory Management
In CUDA programming, memory management functions are essential for optimizing data transfer between the host (CPU) and the device (GPU).
Copy from/to Pageable Memory
In this case you move data manually from host/device side to the other side.
You first malloc
memory on host and copy it to device via cudaMemcpy
. When the computation on device is finished, you copy the result back via cudaMemcpy
again.