Skip to main content

About This Knowledge Base

This site is Zhiyu Ding's long-running public knowledge base. It is also the more workbench-like counterpart to my personal homepage. Built with Docusaurus 3, Markdown / MDX, and React, it organizes material from learning, competitions, engineering practice, server environments, and everyday writing.

At the moment, the site contains about 300 articles and roughly 418,000 words. It is not a textbook with a fixed chapter plan. It is closer to a continuously revised technical archive: some pages build fundamentals, some document troubleshooting in real environments, and some are retrospectives from projects, competitions, reading, and longer-term observation.

What I Write Here​

The content is organized around several main threads:

  • CUDA, SuperComputer, and high-performance computing: GPU programming models, parallel programming, Slurm, supercomputing environments, competition retrospectives, and operator optimization notes.
  • Linux, containers, and engineering environments: remote development, SSH, Docker, proxies, system setup, server deployment, and common commands.
  • Python, data structures, and deep learning: language fundamentals, data processing, course material, experiment environments, and quick-reference notes.
  • WebNotes and LabNotes: site building, DNS, HTTPS, CDN, Docusaurus customization, automation scripts, comments, audio, maps, and implementation notes for this site itself.
  • Look Around, Travel, and Blog: long-form research beyond pure engineering, travel albums, life fragments, and time-ordered reflections.

The categories are not meant to be perfectly uniform. What matters more is that each article tries to answer a concrete question: what happened, why this approach was used, what the limits were, and where to start next time.

Why I Keep Writing​

For me, writing is not just about showing results. It is a way to make the learning process clear enough to reuse.

Many ideas feel understood the first time I meet them, but they only become durable after I restate them, add context, correct details, and record the failed paths. That is the value of this site: it turns one-off searches, experiments, and tuning sessions into context that can be reused later.

How This Site Is Maintained​

I try to make each article meet three standards:

  • Keep the necessary context, not just the conclusion.
  • Explain pitfalls, constraints, and where the solution applies.
  • Keep revising content when it can be improved, instead of leaving old notes frozen at their first version.

So this is not a site that is written once and left alone. It is closer to a growing archive: new pages are added, old pages are corrected, and the site itself keeps changing as real needs appear.

How To Browse​

If this is your first visit, start from All Articles, then use the top navigation, sidebars, or search to find a topic.

For the technical core, CUDA, SuperComputer, Linux, and Python are good starting points. For site implementation and content workflows, read LabNotes. For longer observations and research, read Look Around. If you just want to learn more about me, the homepage and CV are more direct.

Contact​

If you find an error, a broken link, or a topic that should be expanded, you can contact me through the links below, or leave a comment at the bottom of this page.

Comments