CPU-GPU Parallel programming (Python)

Accelerated Computing CUDA CUDA Programming and Performance
1

Hi,
Is there a way we could concurrently run functions on CPU and GPU (using Python)? I’m already using Numba to do thread level scheduling for compute intensive function on the GPU, but I now also need to add parallelism amongst CPU-GPU. Once we ensure that the GPU shared memory has all the data to start processing, I need to trigger the GPU start and then in parallel run some functions on the host using the CPU. Is there a standard library/way to achieve this? Appreciate any pointers in this regard.

2

numba execution characteristics will be similar to CUDA C++ in this regard.

Launch your numba kernel. Thereafter (until you do some sort of synchronization such as a copy of data from device to host) any CPU code you write after the kernel launch will run concurrently (on the CPU) with the kernel (running on the GPU).

3

Thanks Robert. I was thinking on similar lines but wasn’t very sure. I checked that until I put some synchronization for task completion between the cores, I’m effectively running GPU-CPU in parallel. Thanks again.