Introduction

Overview:

  • Teaching: 15 min
  • Exercises: 0 min

Questions

  • What is High Performance Computing?
  • Why would I want to use HPC?
  • What is our local service?
  • How do I acknowledge the HPC service in my publications?

Objectives

  • Understand that HPC is a collection of computing resourcesto deliver higher performance than a single machine.
  • Understand tha HPC can help you run your calculations more effieciently and more reliably.
  • Know that as well as the the HPC service there is a team to help you with your research, and how to contact them.
  • Know how to acknowledge the service in your papers and associate with publications in Pure.

What is HPC and why is it different to using your desktop?

“High Performance Computing most generally refers to the practice of aggregating computing power in a way that delivers much higher performance than one could get out of a typical desktop computer or workstation in order to solve large problems in science, engineering, or business.” - insideHPC

  • Aggregated computing power
  • Very large problem sizes
  • Faster time to solutions for applications that scale
  • Multiple problems simultaneously

Technical Specification

Hardware Count/Amount
Nodes 202
Main compute 3,072 Intel Ivy Bridge cores (16 per node)
Upgrade compute 408 Intel Skylake cores (24 per node)
Total Memory 18TB
Node Memory mostly 64GB/128GB
Skylake 192GB
Hi-mem 512GB
'Home' Storage ~50TB (backed up)
'Scratch' Storage 660TB (Parallel not backed up)
Accelerators GPUS, K20X and P100
Network Intel TrueScale 2:1 blocking
(worst case) QDR Infiniband fabric

Additional sevices:

  • Interactive development and testing nodes
  • Visualisation nodes

Balena sits in a room with other computingresources in cabinets:

Cabinets in a data centre

It has mood lighting to make it look pretty for marketing shots:

Balena lights

Connecting all the computing resources together takes a lot of special cables:

Balena cables 1

And can be quite pleasing to the eye:

Balena cables 2

Advancing Research Compute (ARC) team

ARC run and support the HPC service at Bath:
Steven Chapman ARC Manager
Roshan Mathew HPC System Developer
James Grant Research Software Engineer
Stefano Angioni Cloud Solutions Developer

We provide a range of support and consultancy services including to help users of all skill levels:

  • installing and supporting software applications
  • assistance with creating job scripts, workflow and troubleshooting issues
  • optimising software performance and streamline data handling and analysis process
  • guidance and support in meeting current and future research challenges
  • advice on including HPC and software development costs on research proposals
  • migrate to the Balena/Isambard HPC systems
  • software development consultancy
  • training course

Acknowledge use of Balena/Assistance

Reason

  • Important measure of success for the service
  • Help secure future funding for HPC

How:

  • Research publications
    • “This research made use of the Balena High Performance Computing (HPC) Service at the University of Bath”
  • University of Bath's Research portal, Pure, tag the Balena resouce:

Balena location as Equipment in Pure

Key Points:

  • HPC is a shared collection of computing resources, typically allocated via a scheduler
  • It is not magic to make your code faster
  • The University runs an HPC service Balena
  • The HPC team and RSE can help you to get started on Balena or port your codes
  • Acknowledge Balena/Isambard and tag them in Pure so we can demonstrate the impact of the service in delivering your research.

Previous

Schedule

Next