HDS6605 Multiprocessing Board
6U OpenVPX Xeon (Cascade Lake) multiprocessing server blade
6U OpenVPX Xeon (Cascade Lake) multiprocessing server blade
Give your application the most powerful embedded general processing power available. Powered by a 22-core Intel Xeon Scalable processor with Deep Learning Boost, which extends Intel’s Advanced Vector Extensions-512 (AVX-512), accelerates inference applications and other big data processing and AI applications. New embedded Vector Neural Network Instructions (VNNI) accelerators speed up dense computations, characteristic of convolutional neural networks (CNNs) and deep neural networks (DNNs), delivering up to a 14x improvement in inference performance compared to first-generation scalable processors.
HDS6605 data center server blades are the cornerstone of a truly composable high performance embedded edge computing (HPEEC) solution. When combined with other EnsembleSeries high-performance compute modules, including the latest data center GPU coprocessors, and interconnected with wideband, low-latency PCIe switch fabrics, your application leverages the power of a data center. These big data processing solutions can be tailored to your specific platform and deployed in the harshest and most SWaP constrained environments.
Deploy faster by building your next HPEEC processing solution with EnsembleSeries server blades. The HDS6605 is based on open architecture standards including Sensor Open System Architecture (SOSA) for interoperability, scalability and reuse. Drive affordability, performance and lower your program risk by using an architecture that evolves at the speed of technology, follows Intel’s technology roadmap and leverages commercial data center hardware, software and IP.
EnsembleSeries server blades use proven liquid and air cooling, rugged packaging and high-speed in-system switch fabric interconnect technologies which are open systems compliant and have a technology readiness level of nine (TRL-9). HDS6605 blades are well-suited for systems that require the utmost in reliable processing performance. The HDS6605 is available with 100 Gbps in-chassis fabrics and optional MOTS+ and BuiltSECURE technologies for extreme environmental protection and proven systems security engineering (SSE).
With an architecture designed for big data and AI processing performance, wider, faster channels and 192GB of shared DDR4-2933 system memory, HDS6605 server blades create an elegant compute solution for complex on-platform C4I, AI, radar, sensor fusion and autonomous platform management challenges.
HIGH-PERFORMANCE OPEN ARCHITECTURES
Our high-performance embedded edge compute (HPEEC) architecture packages the data center in a compact OpenVPX™ form factor so you can collocate AI processing next to sensors on edge platforms.
OPEN SYSTEM ARCHITECTURES
Mercury’s modular solutions, designed using open system architectures and open standards, can be quickly and affordably updated with the latest technologies to support changing AI software or algorithm requirements.
PROFOUNDLY ACCESSIBLE
We partner with industry leaders like Intel® and NVIDIA to incorporate the most advanced AI-enabled data center and multi-core edge processors in to the rugged, trusted and secure solutions that power your most ambitious on-platform processing applications.
6U OpenVPX GPU board coprocessing engine powered by dual NVIDIA Turing GPUs
6U OpenVPX wideband PCIe gen 3 network switch board
6U OpenVPX SSD plug-in storage with PCIe switching and removable canister