How to build a supercomputer in AWS with spot instances

Note: This meetup is 1:30pm AEST (https://www.google.com.au/webhp?sourceid=chrome-instant&rlz=1C1CHLW_enAU513AU513&ion=1&espv=2&ie=UTF-8#q=1:30pm+AEST)

Want to know how to replicate the capability of large supercomputer facilities in AWS?

The ANU's NCI (National Computational Infrastructure (http://nci.org.au/)) and Link Digital (http://linkdigital.com.au/), supported by AWS and others, have worked to together since April to create something special.

Join with us as Shane Davis from Link Digital (http://linkdigital.com.au/) gives a live demo of orchestration software which has been built to replicate the NCI's Raijin (http://nci.org.au/systems-services/national-facility/peak-system/raijin/)capability on AWS.

From NCI itself we'll be privileged to also have Dr. Muhammad Atif, Manager of HPC and Cloud Systems co present this demonstration session.

A global pre-release

The scripts and software supporting this demo will be released as open source in the near future, along with a case study on how they have been bench-marked 'in the field' via research teams.

The technology, automated within AWS, provides access to simulation techniques which people would have to otherwise gain access to via research agreements with facility managers. Rather than replace existing supercomputer workloads the software and approach opens the door for peer review, executable white papers and skill development for those learning about big data, data analytics and scientific computing.

Built to be a low cost yet effective option for researchers the scripting makes use of spot instances and placement groups that competes equally on performance in a variety of workload scenarios.

This meetup can be attended at Link Digital, or can be attended via video conference.

Join from PC, Mac, iOS or Android:
https://zoom.us/j/481769504

Or join by phone:
Dial in numbers available at
https://zoom.us/zoomconference?m=c19mDMLBONtpa3u0dYw4hbjJIO0fKm6q