Run Experiment Lite problem

[tizpibes]

Hi all,
is it normal that the number of trajectories printed by the reinforcement learning algorithm varies changing the num_parallel input for running the code in parallel?
Also the performance of the code in terms of returns of the algorithm seems to get worse.

Thank you in advance