Copy
View this email in your browser

In This Issue

Research Computing and Cyberinfrastucture
Newsletter | October 2021

 


Updates


Introducing Pioneer

Research Computing has completed the early stages for deployment of our latest cluster “Pioneer”, named after an old Western Reserve University athletic mascot. Based on Redhat Enterprise Linux 8, one of the biggest changes users should expect is the move from the hierarchical module system found on Rider and Markov, to a flat module system that can incorporate multiple software building tools to help reduce the time required to add and update software on the cluster. The Pioneer cluster currently has sixteen standard compute nodes, and two GPU nodes available to test out functionality and software. Software and libraries are being added on an ongoing basis as part of this phase of deployment. The new cluster will run concurrently with the Rider and Markov clusters.

At this time Pioneer is open to initial testers and should be considered non-production. Interested research groups/faculty members can contact us at hpc-support@case.edu to request inclusion in the testing phase. 
 

HPC Maintenance - October 20, 2021

There will be a 24-hour scheduled maintenance for the CWRU HPC and the attached storage systems on Wednesday, October 20, starting at 9 am. We will be updating our cluster head nodes and compute nodes, the HPC storage system operating system, and a few other maintenance tasks.
 

New Compute Nodes and NVidia DGX A100

As part of the university's investment in high-performance computing, twenty standard compute nodes have been added to the CWRU HPC resource. Half of the servers will join the Rider cluster, and the other half will join the newly-built Pioneer cluster. 

In addition to general computing nodes, one NVidia DGX A100 server is being added to Rider. The NVidia DGX server has built-in-connectivity among the eight A100 GPU cards with 15TB of local storage. NVidia A100 GPU cards are the latest 40GB-memory, multi-purpose GPU cards from NVidia that provide up to 20X higher performance over the prior generation. Once the DGX server is available in early October, we will invite the community to test the server until the end of November 2021. After that, we are planning to prioritize member jobs that will make efficient use of the node, based on usage data collected during the trial period and on proposal submissions.

Nautilus

Ally logo

Case Western Reserve University has joined the Pacific Research Platform (PRP) and other national universities to build a “hyper-converged cluster” called Nautilus, which uses Google’s open-source Kubernetes to orchestrate software containers across the distributed system.  This collaborative resource spans across the US, but mostly on the west coast, and has different types of compute nodes available including GPU and FPGA accelerators. Nautilus provides support for network measurement and monitoring as well as distributed computing and shared storage, currently supported by award NSF #1826967 TNRP. Research conducted on Nautilus includes structure from motion, image processing, image segmentation, virtual reality, and geospatial imaging. If you are interested in using this resource, please contact us at hpc-support@case.edu. Information about how to use Nautilus is available at Nautilus User Documentation.

Google Cloud Platform Availability

Case Western Reserve University has signed an Internet2 agreement with Google Cloud Platform (GCP) that will provide some usage discounts through our vendor, Onix. Researchers can also apply for research credits, up to $5K, by providing the requested information at Google Cloud for Education - Researchers. Please contact us at hpc-support@case.edu if you have any questions or interest in this resource.

Training & Events


HPC Workshops

The start-of-semester training was completed on September 10. For anyone who missed the session, or wants to catch up on some of the topics discussed: Introduction, Basic Linux, and Job Submission presentations are available at the Canvas website: Canvas course 5600.
 

XSEDE Big Data Training, Oct 5-6, 2021

XSEDE and Pittsburgh Supercomputing Center will hold a workshop on Big Data and Machine Learning event on October 5-6, 2021. The registration pages are available below:

https://www.psc.edu/resources/training/hpc-workshop-series/xsede-hpc-workshop-big-data-october-5-6-2021/

https://portal.xsede.org/web/xup/course-calendar


Registration for XSEDE training requires an account at the XSEDE Create XSEDE Portal account

Need Support?


You can contact us at hpc-support@case.edu

Office Hours: Wednesday, 2:00pm - 4:00pm. Please email us in advance so that we can set up the virtual meeting with you.
HPC Wikihttps://sites.google.com/a/case.edu/hpcc/


Want to change how you receive these emails?
You can
update your preferences or unsubscribe from this list.

Copyright © 2021 Case Western Reserve University, All rights reserved.


Email Marketing Powered by Mailchimp