Hi everyone, 
I've developed myself a walkforward optimization and also some bruteforce generator that uses multiple instances of LEAN Engine, to generate multiple backtests all together with different parameters for each instance run.

I'm talking about generally 1 instance per CPU Core, everything is handled with a main program that creates and manages multiple background workers.


Now the problem comes in.
I have noticed that if i run multiple instances , the speed of LEAN does not remain constant, infact it decreases a lot.
 

The thing i've noticed particularly is that all the bars consolidation process takes more time as there are more instances.
I'm afraid tha there might be locks on file readings or such, and it's so annoying that the engine does not even go full throttle and that wastes a lot of time and power.
Can anyone confirm this, or give a hint?

Thanks.

Author