High Performance Computing

🖥️ slurm-usage: Save SLURM job efficiency metrics to Parquet before SLURM purges them after 30 days, and chart CPU, memory, and node usage per user and group. 📈

adaptive-scheduler: Run simulations adaptively on many cores (>10k) and easily manage starting and restarting using mpi4py.futures, ipyparallel, or dask-mpi 🎉

hpc05: a thin wrapper around ipyparallel to easily start workers remotely on a HPC cluster 🖥

adaptive: tools for adaptive parallel sampling of mathematical functions to speed up simulations by orders of magnitude ⚙️