Rugged Server-Class Blade Enables 4X Subsystem Performance Increase Over Previous Generation With 16 SMP Cores Per 6U OpenVPX Slot
CHELMSFORD, Mass., Mar 6, 2012 (GlobeNewswire via COMTEX) — Mercury Computer Systems, Inc., a trusted provider of commercially developed application-ready ISR and EW subsystems for defense prime contractors, announced the server-class Ensemble(R) Series 6U OpenVPX(TM) HDS6601 rugged compute blade based on the Intel(R) Xeon(R) processor E5-2600 family (formerly codenamed Romley-EP). Utilizing two of the new 8-core Intel(R) Xeon(R) processors E5-2648L, the HDS6601 is a 16-core Intel(R) architecture-based module, configured as an easy-to-program symmetric multiprocessing (SMP) cluster. This is the embedded computing industry’s first 16-way, 32-thread SMP server blade compliant with the OpenVPX standard, enabling new levels of rugged subsystem performance with deployed systems.
With the server-class HDS6601, applications can be easily developed to target the high-performance SMP operating environment. Any of the SMP cluster’s CPUs can access any region of the very large, 64 Gbyte on-board memory directly, eliminating the need to program explicit data movement. Subsystems can be easily scaled to multiple boards with Mercury’s OpenMPI/OFED software, industry-leading switch fabric modules and Protocol Offload Engine Technology (POET(TM)). Together, these enable an easy-to-use, high-performance open architecture platform, available with Serial RapidIO(R) and 10 Gigabit Ethernet-based OpenVPX dataplane configurations.
“Key to enabling our sever-class rugged processor module are Mercury’s mechanical packaging innovations for the Intel(R) Xeon(R) processor E5-2648L and the development of new rugged large memory packaging methods,” said Steve Patterson, Vice President of Defense Product Management at Mercury Computer Systems. “And when the server-class processing capability is configured with Mercury’s GPGPU blades and mobile Intel(R) processor-based products, our customers can configure solutions that quadruple the SWaP performance of currently deployed applications,” Patterson continued.
“The Intel(R) Xeon(R) processor E5-2600 family is ideal for high-end processing due to the SMP model supported across multiple devices and enabled by Intel(R) QuickPath Interconnect chip-to-chip bus technology,” said Steve Price, Director of Marketing, Communications Infrastructure Division, Intel. “Combined with the Intel(R) Advanced Vector Extensions processing capability on each SMP core, this creates a high-performance SMP cluster that performs well in ruggedized servers for high-end embedded real-time applications.”
The server-class HDS6601 delivers new levels of performance for both large and small 6U OpenVPX systems. Small subsystems are typically configured with one or two HDS6601 modules for up to 32 cores, while larger systems of up to 10 HDS6601 modules can be configured with up to 160 cores. For extreme computing needs, these larger systems can be augmented with multiple Mercury GPGPU modules and high-performance switch modules to fill a standard 19″ rack mountable OpenVPX chassis. The server-class HDS6601 delivers the defense industry’s best size, weight and power (SWaP) optimized solutions for both large and small footprints in radar, ISR, EW and EO/IR applications.
Utilizing the new Intel(R) Xeon(R) processor E5-2648L with Intel(R) Advanced Vector Extensions (Intel(R) AVX), the HDS6601 delivers nearly half a teraflop in a single standard OpenVPX slot, more than four times what was possible with prior generation products. To create a balanced memory design matching the high performance of the server-class Intel(R) Xeon(R) processor E5-2648L, Mercury developed high-capacity, rugged memory packaging that enables the HDS6601 to support four times more memory per processor than traditional Intel-based OpenVPX boards — up to 64 GBytes per board. For applications that demand the fastest inter-core communications, such as radar beam-forming, the two 8-core Intel(R) Xeon(R) processors E5-2648L are connected with two link pairs of the Intel(R) QuickPath Interconnect (Intel(R) QPI) Bus, quadrupling the inter-processor bandwidth over previous designs, while providing a balanced system of I/O, memory and processing performance.
Building on the success of the previously announced HDS6600 module, the HDS6601 is Mercury’s second generation of signal processing modules based on the server-class Intel(R) Xeon(R) processor family. It is supported by Mercury’s MultiCore Plus(R) software development environment, which delivers high-performance data movement interfaces and processing libraries with industry-standard open interfaces. Users can migrate code seamlessly from Mercury’s workstation-based Virtual MultiComputer simulator (VMC) to the HDS6601, enabling quick deployment of lab-based algorithms.
The HDS6601 is available now and can be configured in air-cooled or conduction-cooled rugged versions.
For more information on the HDS6601, visit www.mc.com/HDS6601 , or contact Mercury at (866) 627-6951 or firstname.lastname@example.org.