Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Enable AVX-enabled builds #38

Open
traversaro opened this issue Feb 28, 2024 · 2 comments · May be fixed by #45
Open

Enable AVX-enabled builds #38

traversaro opened this issue Feb 28, 2024 · 2 comments · May be fixed by #45
Labels
question Further information is requested

Comments

@traversaro
Copy link
Contributor

Comment:

MuJoCo PyPI builds use AVX, and so it may be confusing for users not to get have these optimization in the conda package. Thanks to conda-forge/conda-forge.github.io#2091, we should be able to have AVX-enabled packages for MuJoCo.

@traversaro traversaro added the question Further information is requested label Feb 28, 2024
@traversaro
Copy link
Contributor Author

Once #41, we can easily use the mujoco-testspeed to roughly compare the different package variants performances.

@traversaro traversaro linked a pull request Mar 10, 2024 that will close this issue
5 tasks
@traversaro
Copy link
Contributor Author

Just to give an idea why this is worth, my quick&dirty benchmark is the following, run testspeed (mujoco-testspeed in conda packages) with the model https://github.com/google-deepmind/mujoco/blob/3.1.3/model/humanoid/humanoid.xml , i.e. mujoco-testspeed ./humanoid.xml 100000 . The current results are (this is simulation time, so less is better):

  • ~1.81 for the upstream build
  • ~2.45 for the conda-forge build

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
question Further information is requested
Projects
None yet
Development

Successfully merging a pull request may close this issue.

1 participant