Should I use parfor on an already multithreaded function (decision tree,...)?

Hi, I want to train many decision trees in parallel by using parfor, but I've heard elsewhere that they are multithreaded, so should I use parfor with them? I tried on some datasets, but they seem to be slower. However I don't know if things would be the same for very large datasets. Please help me, my understanding of parallel processing is very limited, I don't even understand the difference between process and thread, thank you very much

Answer 1

[Jan]

Trying is the best way to find out, if it runs faster or not. There is no general best way and it depends on the data and the algorithm.