Tom Deakin explains performance portability and discusses AWS Graviton

Brendan Bouffler
Date : February 2, 2023
Categories : AWS ParallelCluster ,Amazon NICE DCV ,Elastic Fabric Adapter ,Life Sciences
Tags : cpus ,dcv ,ec2 ,efa ,gpus ,hpc ,high performance computing ,lustre ,mpi ,parallel cluster ,schedulers ,storage ,autoscaling ,bioinformatics ,cloud computing ,elastic ,elastic fabric adapter ,graviton ,infiniband ,performance portability ,scientific computing ,technical computing ,tightly coupled ,virtualization ,vizualization ,techshorts

Tom Deakin (and his colleague Simon McIntosh Smith) work on the hard problem of performance portability - the task of getting great performance out of software running on a range of computer architectures. This means they’re often digging deep into workloads that find specific challenges with not just hardware, but compilers and libraries - virtually everything in the stack is fair game.

Tom and I caught up in Bristol before Christmas and he filled me in on a number of things they’ve found already, and also gave us an early report card on the AWS Graviton3, which is our Arm-architecture CPU - but he covers a lot of other ground, too. It’s a great discussion.

If you have ideas for technical topics you’d like to see us cover in a future show, let us know by finding us on Twitter (@TechHpc) and DM’ing us with your idea.