Projects
HPC for Stata
High-performance computing · Wesleyan · Fall 2026
I benchmarked large Stata PPML regressions on Wesleyan’s computing cluster to see how CPU counts, node allocation, and parallel execution affected runtime, and I updated a faculty guide to running Stata on the cluster.
What I did
In a two-month course covering Linux, Bash, and Slurm, I compared serial and parallel configurations using job logs and measured runtimes.
What I learned
Allocating more computing resources did not reliably shorten a run. Measuring before scaling up turned out to matter more than asking for more cores.