Skip to content

Conversation

@foreverska
Copy link

Some minor changes to the C to speed up environment execution

  • Avoid division in the hot loop (instead pre-compute reciprocal and multiply)
  • Early returns on common failure modes
  • Simplify clamps
  • Cache values where possible instead of multiple env structure reads
  • use '.0f' consistently to avoid possible cast

Locally, SPS went from 3.4M -> 4.0M. Ran a puffer train on the environment, agent successfully finds policy.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant