B. Nikolic Software and Computing Blog

Simulation using SLURM code itself. 10,000 homogeneous nodes Mixture of 50, 500 and 5000-Node jobs Submission rate at average ideal throughput of cluster All jobs take 300s with 30s dispersion Standard SLURM scheduling algorithm

Reviewing the source code of the SLURM sluster scheduler I happened on a file “skilling.c”.

Some Friday afternoon thoughts on this perennial question which yet seems to have little good info found through Google : Git is designed to store, retrieve and index trees, i.e., file-system directories, and their sub-directories and files.

The Spack package manager has excellent in-built capabilities for building on system disconnected form the internet.

Had cause to use the STL parallel algorithms this week with excellent results.

Mostly from: https://hpc.nih.gov/docs/job_dependencies.html .

A very quick experiment to try a traditional data cleaning example with Excel and xlOil: calculate the Levenshtein distance between every pair of words.

Of course PostgreSQL is a enough for a room reservation system.

At the time of writing the most recent version of Python is 3.14.2.

Marimo is a reactive notebook system in Python.