Introduction/Overview

HPC not HCP!

This is about High-Performance Computing (HPC) not the Human Connectome Project (HCP)!

The High-Performance Computing Cluster at the U of A is a giant free Linux computing system. It consists of many interconnected computers, and there are other similar systems across the country and the world. Note that there are NO HIDDEN CHARGES, but it is a shared resource, so everyone has to be reasonable about sharing. You'll get cut off if you use too much time! Options for long-term storage, and HIPAA-compliant storage are being explored. This is a good place to get your neuroimaging processing done.

Because you share the HPC with everyone else at the university, the HPC needs a system to make sure that everyone has a fair share of the resources. You'll hear a little about that here (interactive mode and batch jobs), but more detail and practice is coming in subsequent lessons.

This lesson provides an overview of the HPC hardware, and the difference between login mode, interactive mode, and batch submissions. The HPC has a nice explorer called OOD (Open OnDemand), and allows you to open an interactive desktop and run graphical applications like FSLeyes, CONN, SPM, etc.

Practice

The accompanying practice lesson will give you hands-on experience logging on to the HPC with OOD, and configuring your environment for neuroimaging work.

Table of Contents

● What is the HPC
● HPC Hardware
● Pros and Cons
● Time Allocations
● Storage Allocations

HPC 1 (10)