search

stanfordnmbl / osim-rl

Reinforcement learning environments with musculoskeletal models
http://osim-rl.stanford.edu/
MIT License
882 stars 249 forks source link

GPU computation #175

Closed Poddiachyi closed 5 years ago

Poddiachyi commented 5 years ago

Hello.

Is there a way to do environment computation on a GPU? As far as I understand there is no problem with training an agent but the environment is a bottle neck.

Sorry if there was such question before, I couldn't find it.

kidzik commented 5 years ago

No, unfortunately, due to complex model dynamics these operations are not easy to parallelize on a GPU at the moment. Typical approach that participants take is to run simulations on multiple CPUs and the network training on GPUs.