- Rugged data processing/graphics module features AVX2-optimized CPUs with cache-coherent shared memory to OpenCL-programmable GPUs
- Flexible built-in Serial RapidIO and 10 Gigabit Ethernet fabric compatibility
CHELMSFORD, Mass., June 4, 2013 (GLOBE NEWSWIRE) -- Mercury Systems, Inc. (Nasdaq:MRCY), a best-of- breed provider of commercially developed, open sensor and Big Data processing systems for critical commercial, defense and intelligence applications, has expanded its family of powerful system building blocks with the new rugged Ensemble ® Series 6U OpenVPX ™ HDS6502 High Density Server (HDS) module based on 4th generation Intel ® Core ™i7 processors (formerly codenamed "Haswell").
Skillful integration of dual quad-core Intel Core i7-4700EQ processors with native Intel Advanced Vector Extensions 2.0 and PCIe Gen 3, an OpenCL ™ programmable GPU and up to 16GB of 1600MHz DDR3 SDRAM within a rugged OpenVPX module continues the company's tradition of performance-driven, highly-configurable, open-standard systems that leverage the organization's scalable system building blocks. The HDS6502 is designed for OpenCL deployment and, like all Mercury products, is configured with others to form subsystems to meet unique application-specific requirements.
"The HDS6502 is a natural and complementary evolution to our other HDS/LDS (High Density Server/Low Density Server) product solutions ideally suited to high performance, real-time embedded applications," said Gregg Ogden, Director of Solutions and Product Marketing, Mercury Systems. "But that is only half the story: this module is also a critical building block for our existing customers, enabling them to future-proof their legacy systems through targeted upgrades. Upgrades using the HDS6502 preserve their software, optimizeSWaP (Size, Weight and Power) performance, and, true to Mercury tradition, more again can be achieved with less. And this is all done while enabling full OpenCL implementation. The whole concept of open systems is broadened by through the use of OpenCL programming support across multiple platforms."