Huge slowdown in cmdstanr

Hey everyone, I just had some quick questions. I just updated my R to 4.3.3 from 4.3.2 and then updated cmdstanr to the 4.3.3 version as well and I have noticed a huge slowdown in my runtimes for running my BSEM model, like previously it was taking me about 10 minutes and now I am pushing 30 minutes under the same conditions. Maybe that is a coincidence, but I have been saving backup older versions for weeks of this code and that code is running slow too, so I don’t think it is just recent updates to my code that have really bogged me down.

I am using lavaan to simulate my data and have had no problems for some time passing my data over to stan. I recently did start using cmdstanr over rstan and that had proved to be a time saver.

Has anyone else noticed a large bump in run times? Does anyone know if running from a previous version is faster?

My code compiles and runs as it did just a few days ago, but about 3-5 times slower.

This may be best reported as a performance regression here:

It would help if you could provide a reproducible example.

Did you previously have anything in make/local? If so, did you make sure to port it over when updating cmdstanr?