There is a very well defined param

NumCores = -2; // use all logical processors but two

However for strategies like Z12 which are pre-compiled, Zorro seems to be using a single core even in S.

I tried adding this param to Zorro.ini (hoping it gets read in at the start of program) but that didn't work.

Is there another way to allow Learning using multiple cores on built-in strategies?

As to my knowledge, Z12 uses multiple cores for a full training run. But a single WFO cycle uses only a single core.

Ciao, am I wrong or full training of z system is not possible?

No, I see in the script that you're right. Training by users only trains the last cycle, not the full period.

Thanks