Make sure timing from pool emcee scales well for large runs. So far only made an estimate comparing 1 core in a notebook with 128 in a node, but didn't run the full notebook example, and also comparing a notebook in a shared jupyter node vs timing from a full node might not be an accurate comparison. Should rerun properly.
Make sure timing from pool emcee scales well for large runs. So far only made an estimate comparing 1 core in a notebook with 128 in a node, but didn't run the full notebook example, and also comparing a notebook in a shared jupyter node vs timing from a full node might not be an accurate comparison. Should rerun properly.