HPC Terminology

HPC Terminology

Understanding the key concepts and jargon

  • It’s ok to not remember all of these!
  • We just want these terms to be familiar to you.

HPC

  • HPC stands for High Performance Computing
  • HPC systems or platforms can also be called clusters or supercomputers

HPC systems allow multiple processors to work on the same problem at the same time…

allowing complex problems to be solved faster than they could be on a single computer.

HPC

HPC systems allow multiple processors to work on the same problem at the same time…

  • Years to run on a single processor ⇨ might run 10 or 100 times faster using 10 or 100 processors in parallel
  • It’s not magic - there is no fairy dust that can just make code run on hundreds of processors
  • But it could transform your workflow if you put the effort in to shape your code to work on a HPC system

Cluster

  • A system of computers that work together to allow large-scale computing
  • Imagine a network of laptops hooked up together on a high speed network
    • Each laptop is a “node” in the cluster and acts as a server

Network of connected computers and storage forming a cluster

Node

  • A node is an individual physical machine/server, that when linked to other nodes makes up the cluster
    • Linked together by a high-speed network which allows very rapid communications between the nodes
  • In current systems, a node would typically include one or more processors, as well as memory and other hardware

Think of it like this:

  • Your laptop = 1 node
  • HPC cluster = many nodes connected together

Processor

The processor or central processing unit (CPU) inside the node, which contains one or more cores.

  • Your laptop or desktop will have a CPU, for example “Intel Core i7”
  • Machines can also have a graphical processing unit or GPU (often used in gaming and machine learning)

Core

This refers to the basic computation unit of the CPU. This is the unit that carries out the actual computations.

  • You will have to define how many cores you want to run your code across
  • You can use cores across multiple nodes

Aire Specifications

  • Aire is the University of Leeds’ latest HPC system
  • It uses a Linux operating system
  • The system has 52 CPU nodes (and 28 GPU nodes)

Each CPU node has:

  • 168 cores (ARC4 had 40 cores)
  • 768GB memory

You can read more about the Aire system on our website