Slurm’s job management API, coupled with ParallelCluster’s own API for managing cluster infrastructure opens up a lot of doors to create new ways for your users to interact with HPC resources - in this case without having to …
Read More
In a previous show (https://youtu.be/6gAwAK5IJ2w), we talked about ‘Infrastructure as Code’, or how the ParallelCluster 3 YAML config translated to an actual cluster running, ready to take jobs. Today, Austin Cherian joins us …
Read More
On Jan 10, the new Hpc6a instance became generally available in two regions. This instance is built from the ground up for HPC and comes with a number of interesting innovations. Neil and Heidi from our application performance team came …
Read More
In December, we introduced EnginFrame with AWS HPC Connector, which is a new feature in EnginFrame to help customers leverage managed HPC resources in AWS as well as their on-prem systems - that means a single interface so administrators …
Read More
reInvent ‘21 is over, but it was a busy week with new tech developments we’ve built for customers in pretty much every area of HPC: storage, networking and compute. We talk with our HPC General Manager, Ian Colle, about ZFS, …
Read More
CDI (Cloud Digital Interface) uses EFA to blast broadcast quality video, uncompressed, from instance to instance in the cloud as live video feeds its way through complex stacks of software and makes its way to your TV set. It’s one of …
Read More
Mike Lin is the creator and maintainer of miniWDL and has been working with our Genomics and ML teams to integrate it with Sagemaker Studio and AWS Batch. The result is a working environment that’s completely familiar and easy to use …
Read More
If you’re a frequent user of Batch queues, you have opinions about whether it’s scheduling is “fair” on your users. But ‘fair’ means different things to everyone - what’s needed are some levers and dials that allow you to figure …
Read More