Nov. 18 — Penguin Computing, experts in high performance and enterprise computing solutions, today announced Tundra Open HPC. This clustered architecture can support up to 108 high performance compute servers in a single rack with improved serviceability and investment protection.
Penguin’s Tundra Open HPC architecture is based on the Open Compute Project’s Open Rack form factor that re-engineers the conventional data center rack for maximum power efficiency, serviceability and investment protection. Penguin Computing’s Tundra Open Rack provides 36 Open Units (OUs) of space for compute equipment and dedicates additional space for shared power infrastructure and networking components. Using the new Penguin designed compute module, three nodes can be mounted in each Open Unit, enabling deployments with up to 108 nodes within the dimensions of a traditional 42U rack. Using Intel E5-2600 V2 processors, each rack can provide over 40TFLOPS and nearly 50TB of RAM within the dimensions of a traditional 19″ 42U rack.
“The Open Compute Project started by addressing the needs of Facebook datacenters and large scale internet/cloud environments,” said Charles Wuischpard, CEO of Penguin Computing. “As an OCP Solution Provider and leader in High Performance Computing solutions, we saw an opportunity to bring the benefits of OCP to the design and deployment of next generation supercomputers. Today’s announcement of Tundra Open HPC lays the foundation for future work in the areas of exascale computing, large scale distributed storage and open, software defined networking.”
“We are excited to see Penguin Computing introduce the Open Compute Project’s design philosophy and standards to the High Performance Computing market,” said Frank Frankovsky, Chairman and President of the Open Compute Foundation, “There are many similarities between today’s largest internet data centers and supercomputing and the Open Compute Project’s benefits apply to both. Penguin Tundra Open HPC is an example of innovation that we encourage from our OCP Solution Providers.”
Penguin offers Open Compute Ready TM compute, storage and interconnect solutions both for Open Rack and conventional 19” EIA racks. Tundra Open HPC designs will be contributed to the Open Compute Foundation, expanding Penguin Computing’s traditional high performance computing solutions and establishing a new architecture for current and future datacenter solutions.
Penguin Computing will be exhibiting at Supercomputing ’13 from November 18-21, in Denver, Colorado at booth #1907.
For more information, please visit www.penguincomputing.com.
About Penguin Computing
Penguin Computing is the largest private supplier of complete high performance computing (HPC) solutions in North America and has built and operates the leading specialized public HPC cloud service Penguin Computing on Demand (POD). Penguin Computing also applies its core expertise in the field of distributed large-scale enterprise computing delivering scale-out compute, storage, virtualization, and cloud solutions for organizations looking to take advantage of modern open data center architectures.
Source: Penguin Computing