Pick a task on the left
Each task opens with the steps to follow, the prerequisites that must already be in place, and a diagram of how the pieces relate to each other, so you can see what a task actually builds rather than only what you type.
Work from task 1 to task 100 in order and you will have drawn a GPU node and its fabric, prepared Linux hosts, installed drivers and DCGM on a real GPU, run Slurm and Kubernetes with the GPU Operator, built Prometheus, Grafana and Loki, simulated a RoCE fabric and read InfiniBand reports, recorded an NCCL baseline, automated it all with Terraform, Ansible and Argo CD, written the runbooks the job descriptions ask for, and rehearsed eight timed scenarios. Tick tasks off as you go; progress is kept in this browser.
Environments: P needs only paper. A is your laptop. B is one rented GPU instance for a few hours. C is two instances and is optional.
Keyboard: J and K, or the arrow keys, move between tasks. D marks the open task done.