Multiple Envs performance

[leopt4]

Hi!
I am running learn.py with n_envs=8, open htop and see at one time, there is only one cpu core 100% while another are approximately 0. And there is only 1 thread running. Meanwhile I tried ./parallelism.sh, at least 4 cores running and there are 4 or 16 thread running.
It means I am running all envs in one core, right?
There are 16 cores in my cpu. Can I run each env in each core of cpu?
Thank you

[JacopoPan]

The n_envs variable in learn.py refers to the environments use by Stable Baselines3 make_vec_env utility
https://stable-baselines3.readthedocs.io/en/master/common/env_util.html#stable_baselines3.common.env_util.make_vec_env
You could try to use vec_env_cls SubprocVecEnv instead of DummyVecEnv