Skyler Medeiros

@alsora @jmachowinski I wanted to give a quick update on running these benchmarks ahead of the client library working group this friday. I got our benchmark container running with ros2...

@jmachowinski I was doing these benchmarks with ROS rolling. The topology was a newer one we made on top of some other tooling that builds on top of `ros2-performance`, including...

@jmachowinski unfortunately after doing a clean rebuild of my workspace with the top of your rolling branch (and top of rclcpp and rcl rolling from today) I'm still getting a...

For the actions freeze, running with valgrind gets us a `Access not within mapped region` warning when I'm able to repro the freeze: https://gist.github.com/skyegalaxy/3b40c8d51ca1d874270825041f2cda38 For the 50 topic 2000 hz...

Forgot to mention: for the actions freeze, the topmost thread (whether run through valgrind or just the program itself) enters a zombie state that prevents the thread itself from dying,...

I got caught up with other work stuff yesterday but I'll see if I can find time to set up the lxc container either today or tomorrow. In the meantime,...

My analysis / writeup of the data I collected is [here](https://gist.github.com/skyegalaxy/71ba2696e6d3a081bdcb0b9fb51a4c71). tl;dr: - there doesn't appear to be too much of a difference across all the characteristics we tested. EventsExecutor...

Next steps as a follow-up to the discussion in the client library meeting today: - [ ] Rerun these benchmarks on more constrained hardware (raspberry pi 4) - [ ]...

@jmachowinski - I finally had some time today to set up the LXC container and do a clean rolling build, with cm_executors (rolling) and ros2-performance. I am still able to...

you'll also wanna make sure these benchmarks are being run with sudo/root so that the CPU governor can be set to performance mode. In my case, I simply set up...