as I understand the current benchmark status, it is: 1. GMM as plotted above, but with a faster accuracy calculation we can run on other data set sizes which should show our advantages more. 2. NB runs with accuracy calculation for hakrit and haskell backends, with hakrit much faster, but no JAGS code or other comparison currently 3. LDA runs in hakrit and haskell backends, but we don’t have any accuracy calculation

Is that accurate? Are those the benchmarks we’re going with, and if so, what do we need to do to unblock each?