-
Notifications
You must be signed in to change notification settings - Fork 86
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
Disable core affinity in ValkyrieAffinity when running in Gazebo environment #99
Comments
I can confirm that more than one person has this problem. |
I assume you mean using the |
Correct you should only need to redeploy the controller once the release is available. |
Thank you for working to get this rectified so quickly! Will this issue be resolved sometime this coming week? |
I am also interested in the release status of this -- thanks. Short of buying a new computer, I can't help my team right now. |
Standing by for this release, gone as far as I can with two cores............ |
Excellent! Thanks for this fix. Should we re-install with these changes or change the build locally? What is the best way to install these changes over an existing build? |
I'm currently writing up a tutorial for how to apply these changes to an existing build. Should be done by the end of the day. |
I just uploaded a new valkyrie_controller.tar.gz (step 10 in installation instructions), so please delete the |
…e/icp_optimization_cleanup to develop * commit '8bd250bd571e81b452e611ea10ff9ab6269b0a47': (33 commits) cleaned up more stuff removed the old multiplier calculation fixed the remaining icp optimization tests finished fleshing out all the multipliers for the different states expanded the functionality of the recursion multipliers to include one cmp fixed a bug with calculation of some of the current projection multipliers started fixing the reconstruction of the icp trajectory started using the time in the current state rather than the remaining time got it mostly walking with the new parameters fixed a major bug with the new ICP optimization in single support finished out the tests of the multiplier started improving a lot of the tests added a bunch of unit tests added a bunch more test scripts for the new state matrices started adding more tests got it computing the final ICP location correctly set up the new calculator finished first draft of new multipliers setup the recursion multipliers started overhauled the multipliers calculation, and the way we get reference points ...
When running Val in Gazebo, core isolation is not needed because we run in lockstep with the physics loop. It's reasonable to assume that developers won't be using machines that line up with what the real robot hardware looks like. At least one person has run in to a launch problem where they cannot run the simulation software on a 2 physical core computer.
We should use the parameterization added in 292a659 to parameterize ValkyrieAffinity to avoid this.
Contributors with relevant knowledge or expertise (optional)
@dljsjr
The text was updated successfully, but these errors were encountered: