What’s does it take to power protein-folding research as part of a fierce academic competition? Look to high-performance computing (HPC) solutions from HP and Intel.
This was the case at the University of Michigan Medical School. It needed to create an HPC environment with maximum availability for 24x7 processing of protein-folding research, including work in a competitive event likened to the Olympics of protein folding.
In this event—known as the Critical Assessment of Techniques for Protein Structure Prediction (CASP)—research teams race against each other in a 24-hour contest to predict the structure of a protein model. In this competition, the sporting equipment consists of massive compute resources.
The university found its HPC answer in a game-changing server configuration: HP ProLiant DL1000multi-node servers with four nodes in each 2U chassis for high-density compute power and increased energy efficiency. That product paved the way for a solution that helped the university satisfy the protein research team’s needs while meeting its space, power, and cooling targets.
The resulting solution was built around 67 DL1000 chassis, each with four dual, quad-core Intel Xeon processor 5500 series for a total of 2,144 cores within 268 servers. To manage the cluster, the university also deployed two HP ProLiant DL180 G6 servers as head nodes, also using Intel Xeon processor 5500 series. For the university, this was just the right combination.
“HP came out with their DL1000 series servers, which were a great solution for us—it’s the perfect middle ground between blades and a single-unit form factor for a really good price,” notes Traci Ruthkoski, HPC Team Lead, University of Michigan Medical School.
To get the full story, check out this HP customer case study.