New Tundra Open Compute Servers Optimized for HPC and Hyperscale Workloads
Penguin Computing has confirmed availability of the company’s Open Compute Project (OCP) compliant “Tundra” server family based on Cavium’s 64-bit ARMv8 ThunderX workload optimized processors. Penguin Computing provides customized build-to-order server solutions for customers with specialized hardware requirements in enterprise, financial, federal government, bioinformatics and Internet segments. The company has successfully delivered hundreds of data center deployments, from departmental HPC clusters to TOP500 class supercomputers, and scale-out server farms powering some of the largest web properties. The development of ThunderX-based servers enables Penguin Computing to provide best-in-class ARMv8 performance with reduced TCO, using SOCs optimized for specific workloads, on customized platforms.
These platforms will focus on highly-scalable hyperscale and HPC-type workloads including big data, large-scale graph analytics, molecular dynamics, and Ceph / cloud storage. Performance will be driven by ThunderX ARMv8 SOCs optimized for these workloads, with up to 48 high performance custom cores, very high memory bandwidth, and large memory capacity. Adding to these high-performance features are integrated accelerators for storage, networking or security, offloading from the main ARMv8 CPU cores and enabling best-in-class performance. In addition, integration of Ethernet fabric interfaces and storage controllers reduces both power and cost, driving outstanding reductions in TCO.
Penguin has been the leader in developing open, Linux-based cloud and HPC solutions for well over a decade. This latest Tundra server family allows customers to benefit from the efficiency and flexibility enabled by OCP infrastructure combined with Cavium ThunderX architecture’s workload-optimized performance. Customers may initiate evaluation in September 2015, with general availability in the fourth quarter of the year.
Visit our Tundra page for more information.