Can we accelerate functions from other packages using juliaGPU package?

Specific Domains GPU
1

Hello,

I want to accelerate functions imported from other julia packages. In the past, I always use

using Distributed
addprocs(Num)

To accelerate computation, and what I need to do is just write the functions and execute it. In fact, I don’t need to do anything special about it.

I am new to juliaGPU and what I want to know is: If I am importing an external function from outside packages, for example funCPU(a,b), and I want to do is use juliaGPU to accelerate it. How can I achieve that? Since this function is write in outside packages and defined in CPU code.

2

In Julia, the GPU or CPU location is encoded in the type of an array. For instance, a CuArray from CUDA.jl lives on an NVIDIA GPU. So if you have a function generic enough to accept arbitrary array types, then applying it to a CuArray will just work.

3 Likes
3

Many thanks for that, I will have a try!