Thanks guys. It makes sense now why the anchored training would use only one core as the second core would have to wait for the first one to finish the training, the third one for the first two ... and so on.

A note added to the documentation would not hurt plus some heads up on the fact that multi core training does not kick in if the source file is not in the Strategy folder. I also tried to "organize" the scripts and I created another folder. The confusion was even bigger but everything seems to be clear now.

Thanks.