CosmoCoffee

Hello --

When resuming a stopped run using the checkpoint = T option, what covmat is used? I am using MPI Learning to resolve the covmat. Do the resumed chains use the latest state of the learned covmat from when the chains stopped, or do they use the original covmat specified in params.ini?

Thanks for any help!

Brian

With checkpointing on they use the latest covmat (i.e. the covmat at the last checkpoint).