High-performance computing (HPC)
Running work across many machines at once so it finishes far sooner. The technique behind weather models, genomics, and large-scale simulation.
Historically this meant an on-premises supercomputer: a fixed cluster, a job scheduler, a queue, and a procurement cycle measured in months. The compute model has not changed. Work is still split across nodes and recombined. What changed is access to it: what used to require owning a machine room is now something you request for the duration of one run. Everything these courses teach is HPC; the vocabulary is just less formal than the field's own.
FundamentalsDistributed compute
Related terms
- HPC as a service (HPCaaS): High-performance computing consumed on demand rather than owned: request compute for one run, stop paying when it ends. No hardware, no scheduler to run.
- Cluster: A set of machines, one head node and one or more compute nodes, launched together to run distributed work.
- Compute node: A machine in a cluster that executes tasks. Compute nodes do the work; the head node coordinates it.
- Distributed computing: Splitting one job across several machines that work on it simultaneously. The core technique of HPC, and what the distribute decorator gives you in Python.
- Parallel computing: Many calculations at once rather than one after another. Two scales on Eugo: across machines in a cluster, and inside a CPU core through vectorization.