Multi-Core J - Parallelization

Is there a way to get J to use multiple cores? I thought that part of the advantages of APL / J is that language constructs are well suited for parallel solutions.

Looking at using my processor (I'm on OSX), it is obvious that only one processor is used.

I have a heavy function f acting on a list and I don’t understand why it couldn’t split the list into 4 parts and reassemble the results?

+6
source share
1 answer

ArrayFire may be interested. Its OpenCL with AMD / nvidia support and reverse CPU support. Its processing arrays. It should easily bind to J, as happens in matlab.

0
source

Source: https://habr.com/ru/post/984398/


All Articles