Mercury Systems unveils 3U OpenVPX processing solution
By Loren Blinde
March 12, 2015
Mercury Systems, Inc., a Chelmsford, MA-based high-tech commercial provider of more affordable secure and sensor processing subsystems powering today’s critical defense and intelligence applications, announced on March 9 the rugged OpenVPX™ Ensemble® LDS3506 processing module that seamlessly integrates the Intel® Xeon® processor D system-on-a-chip (SoC) product family (formerly codenamed “Broadwell DE”) with Xilinx’s powerful Ultrascale™ FPGA in a SWaP-constrained 3U package. This dense union of best available commercial-item general processing and FPGA resources produces a highly versatile, affordable and interoperable building block for embedded, high-performance computing applications with additional low-latency, refresh and mission capabilities.
The Ensemble LDS3506 leverages Mercury’s fourth generation of highly SWaP-efficient packaging technology to securely deliver the Intel Xeon processor D product family for reliable deployment in scalable subsystems right to the tactical edge. The Ensemble LDS3506 secures and cools the best commercial technology to produce a dense, powerful combination of general server-class and low-latency front-end FPGA processing in an open systems architecture module that is designed and made in the USA.
“With over 256 peak GFLOPS of general processing power from the Intel processor alone, the Ensemble LDS3506 represents a disruptive force in the 3U embedded market space,” said Ian Dunn, Vice President of Mercury Systems’ Embedded Products group. “But, it is the unique combination of this latest Intel technology combined with the low-latency and versatile FPGA resources that delivers a potent compute solution for multidimensional SWaP-constrained applications, particularly those that are sensitive to performance and latency, such as electronic warfare (EW), electro-optical/infrared (EO/IR), image intelligence (IMINT) and other mission or sensor processing applications.”
The LDS3506 provides x4 Gen3 PCIe connectivity across the data plane via Xilinx’s FPGA device, with multiple DMA-enabled non-transparent bridge (NTB) interfaces, giving users the versatility needed to construct powerful processing subsystems quickly. The module’s latest Xilinx FPGA hosts Mercury’s Protocol Offload Engine Technology (POET™) to give each module the ability to refresh its mission capability, provide information assurance abilities, or even refresh or upgrade its switch fabric itself without affecting any hardware.
“The Intel Xeon processor D family featuring Intel Advanced Vector Extension 2.0 (Intel AVX2) support, SoC architecture and the integrated Platform Controller Hub enables server-class performance within smaller form factors, including 3U OpenVPX,” said Steve Price, General Manager, Communications Infrastructure Division, Intel. “The elegance of the SoC architecture of the Intel Xeon processor D-1540 supports a new range of dense, efficient system designs in both the micro-server market and the embedded defense application space.”
The Ensemble LDS3506 supports open data movement middleware, including Open MPI and OpenMPI/OFED™, VITA 46.11 system management, and standard optimized math libraries. Modules will be initially populated with the 45W TDP Xeon® D-1540 processor and will be available with additional Intel® Xeon® D series processors as they become available. Customer demonstrations will begin spring 2015 and production units will become available later this year for air-cooled, conduction-cooled, and Air Flow-By™ systems. Additional products in the 6U OpenVPX form factor based on the Intel Xeon processor D family are expected to be announced by Mercury later this year.
Source: Mercury Systems Inc.
