AEROSPACE
CPUs, GPUs System Delivers Simulations in Real-Time at UCLan
- New HPC system provides 100 fold speed increase on previous system and enables viewing and editing of research simulations in real-time
- HPC system contains GPUs enabling scientists to test and develop more compute intensive simulation codes
A central, shared HPC system, scientists at the University can access the system via their local workstations, compile simulation code in user accounts and, once processed by the HPC, view the resulting simulation locally.
The HPC system replaces an existing, aging system for data processing and provides a 100 fold speed increase.
“Not everyone at UCLan has a super powerful workstation to create their own simulations,” says Dr. Steven Chapman, HPC manager, UCLan. “Using the new HPC, scientists can run simulation code on a more powerful, shared machine and view the resulting simulations via their local workstations. We can also now view and edit the simulations in real-time, on the fly; we no longer have to wait – some times days - to fully process the code, view and edit the simulations.”
The HPC system also includes a number of Nvidia GPUs enabling scientists to test and develop more compute intensive simulation codes.
The HPC system, funded entirely by the University without external financial support was designed, installed and configured by UK HPC and storage integrator OCF. OCF also made available a £15,000 bursary towards a PhD studentship.
“UCLan is a great example of how universities can successfully purchase and house a powerful compute resource centrally but make it available to a wide range of scientists with different research areas and working in different departments,” says Julian Fielden, managing director, OCF plc. “Central purchase, shared use is a growing theme in HPC.”
In the future, UCLan is aiming to extend use of the HPC system to additional departments including the bioinformatics and health teams.
The HPC system uses Sun Fire X2270 server hardware with Intel Nehalem CPU processors and Sun Fire X4170 server hardware.