Research Computing Support Services Winter Update

The past year has seen a lot of changes for research computing on campus and a lot are in the works. The beginning of the year saw the commissioning of more than one petabyte of General Purpose Research Storage (GPRS) for researchers on campus to have an affordable way to store large amounts of data (currently at $10/TB/month). The MU Cyberinfrastructure Council was instrumental in the development of the policy for the storage system and will help develop new services and policies in the future. The Lewis HPC upgrade started with the commissioning of 32 new nodes in conjunction with a new experimental cluster and 32 additional nodes sponsored by the National Science Foundation. The combined system has more than 2,000 cores and is growing. Because the two systems share infrastructure, the Lewis3 system can take advantage of cycles not used by grant activities.

In addition to the GPRS service, we launched our HPC Investor service, which allows researchers to invest in the cluster in a grant-friendly way and get priority access to those resources. This partnership allows researchers to obtain computational capacity without having to manage their own systems. For more information see

The year 2015 also saw the transition of the stewardship of the research computing infrastructure from Research Support Computing (RSC) led by Gordon Springer to Research Computing Support Services (RCSS) led by myself, Timothy Middelkoop. Gordon Springer retired after many years of dedicated service to the campus community; however, he is still active in the community as an Emeritus faculty in the Computer Science department. The new team and I look forward to building upon the work that Gordon and his team did.

As we look forward into the new year we will be upgrading a lot of our infrastructure. Of note we will be:

  • Upgrading the HPC storage system
  • Decommissioning the old Lewis cluster (Lewis2) this summer
  • Upgrading and migrating services connected with the Lewis2 cluster

This will have an impact on a lot of existing services, so we will attempt to contact all those involved. Researchers that have invested in the past with RSC, or have active services, are encouraged to contact us directly to develop a migration plan.

We are also building new services to serve researchers on campus in a sustainable way. Of note we are:

  • Building an OpenStack environment to provide virtual machines for researchers to be used in their research
  • Building a Secure Research Compute cluster for researchers that need a secure compute environment for DCL3 and DCL4 data (
  • Building a student cluster to provide computational resources for teaching
  • Upgrading the campus Research Network (RNet) infrastructure that provides a Science DMZ for researchers on campus.

As a part of the upgrade to Lewis3, we have been providing training sessions for new and existing users to get the most out of the cluster. Last fall we also brought Intel to campus for a course on parallel computation. We plan on expanding the training opportunities for researchers on campus in the upcoming year. We will also be holding periodic user forums where we can talk with researchers about the cluster and to help improve our services. Watch this space for updates.

Feb. 17, 2016

gprs, hpc, Lewis2, lewis3, RNet, Science DMZ