Summary and Reflection on Fedora

Motivation

Currently, my PC (with Nvidia GPU) serves only one purpose: training models using CUDA as the second stage of my progressive development pipeline. In this pipeline, Windows 11 essentially acts as a launcher for WSL 2 to run my Linux-based code. This made me wonder - why not bypass Windows 11 entirely and install a Linux OS directly on the hardware?

Major Decisions

Having previously installed macOS on PC laptops and Windows on Intel MacBooks, installing another OS wasn’t daunting.

The first decision was choosing which Linux distribution to use. Since I planned to use it as the main system on my PC, I wanted something both functional and aesthetic. After browsing options, Elementary OS and Fedora caught my attention. I ultimately chose Fedora for its simplicity, clean interface, and ease of use. I tested both distributions in VMs on my macOS.

Next was creating the bootable USB drive. I used BalenaEtcher with the Fedora Workstation Live ISO to create the installation media. This step was quite straightforward.

The challenge came when partitioning the hard drive. Windows’ default disk management tool failed to provide enough space. After debugging for a full day without success, I tried AOMEI, but its free version didn’t support partition adjustment. Finally, MiniTool Partition Wizard solved the problem.

During the Fedora installation, I initially forgot to disable Secure Boot in BIOS. Once I addressed that, the installation process was straightforward.

I needed to remap keys since I’m more familiar with the macOS layout, which I accomplished using GNOME Tweaks. For Chinese input, I had to configure additional settings to use [] for navigating between Chinese character pages.

Finally, I installed my development environment, including VS Code, Docker, CUDA drivers, Claude Code, and other essential tools.

Exploring the new world!




Enjoy Reading This Article?

Here are some more articles you might like to read next:

  • Reddit Data
  • Reflections on Progressive Development, Docker and HPC