Technical Essays¶
Longer-form writing on HPC, scientific computing, and programming — things that deserve more than a how-to guide but don't belong in the main docs.
HPC
-
Why Linux on HPC?
The practical and historical reasons every serious cluster runs Linux, and what that means for researchers moving from Windows or macOS workflows.
-
Demystifying Compiler Toolchains
Why your HPC software stack depends on GCC versions and what toolchains actually do under the hood. Covers the GCCcore–GCC–foss hierarchy and why mixing toolchains silently breaks things.
-
GH200 Grace Hopper: LLM Inference Performance (Qwen2.5-72B, FSDP)
72-billion parameter large language model (Qwen2.5-72B) across 2 x GH200 GPUs and runs repeated forward passes at a sequence length of 8,192 tokens
Programming
-
Unit Testing for Scientific Code
What is Unit Testing and Why It Matters for Research Software¶
-
Migrating R Packages Between Versions
Using
.Rprofileto switch R library paths dynamically depending on module environment, so you don't end up with mixed library states across sessions. -
Version Control with Git - A Technical Primer
Understanding the core concepts of Git version control
-
SSH on Ghostty
Getting Ghostty's terminfo and italics working correctly over SSH to remote HPC login nodes, without corrupting your remote
TERMenvironment. -
Date Integrity (checksum)
verify transferred data with checksums — compact cryptographic summaries computed in a way that even a single changed bit produces a completely different value
