Mercury Systems Announces High-Performance Compute Blade for Artificial Intelligence and Other Advanced On-Platform Processing Applications
Scalable Xeon and UltraScale FPGA processing power, and extensive memory packaged in rugged, secure, small form-factor blade for deployment anywhere
Mercury Systems, Inc. announced the EnsembleSeries™ LDS3517 processing blade for advanced on-platform processing, machine learning and artificial intelligence (AI) applications. Each blade combines an Intel Xeon® D server-class processor, a Xilinx® UltraScale® FPGA and a mezzanine site in a 3U OpenVPX™ form-factor. Wafer-stacking and system-in-package (SiP) miniaturization technologies enable the latest general and FPGA processing capabilities, their extensive supporting memory and a versatile mezzanine expansion site to fit into this compact form-factor. The LDS3517 blade is ideally suited to on-platform cognitive electronic warfare (EW), next-generation radar, machine learning and AI applications that require small, powerful and scalable processing engines. Optionally configured with embedded BuiltSECURE™ systems security engineering and packaged with modified-off-the-shelf-plus (MOTS+) technology, the LDS3517 compute blade can support military missions anywhere.
“Our customers are asking us for small, powerful processing blades they can run their most compute-intense applications on and these resources need to be rugged and secure for deployment at the tactical edge,” said Joe Plunkett, Mercury’s Senior Director and General Manager for Sensor Processing Solutions. “The EnsembleSeries LDS3517 meets these needs, making all manner of new missions possible from advanced on-platform electro-optical/infrared (EO/IR) and EW cognitive processing to platform autonomy and AI.”
EnsembleSeries LDS3517 blades feature:
- Embedded BuiltSECURE technology: Proven, built-in, system-wide security across software, firmware and hardware in the domains of system security engineering (SSE), trust and cyber-hardening.
- Dense, versatile processing capability: The LDS3517 is a 3U OpenVPX Xeon D family-powered processing blade with powerful UltraScale FPGA support and a mezzanine site for customization and broad-streaming I/O pre-processing, off-boarding and deep packet inspection.
- Versatile mezzanine site: An XMC mezzanine site supports I/O customization, making the LDS3517 exceptionally versatile for a wide variety of high-bandwidth signal processing applications. Each blade is interoperable with Mercury’s other EnsembleSeries OpenVPX processing building blocks for rapid processing subsystem pre-integration.
- Rugged packaging: The LDS3517 optionally includes MOTS+ technology that equips blades with the highly rugged embedded packaging. MOTS+ is an additional layer of environmental protection and ruggedness that requires the soldering of all board-level devices, regardless of their native packaging/terminations, to their respective substrates for reliability.
- On-board AI processing: AI applications require higher levels of processing power and larger memories. The LDS3517’s Xeon D family processor is supported with a powerful UltraScale FPGA and the system memory required to run machine learning and AI applications.
Mercury’s EnsembleSeries™ processing solutions are designed, made, programmed and supported in the USA in DMEA-accredited facilities using devices from our trusted, managed supply chain. Customer orders are currently being accepted for LDS3517 blades in conduction and Air Flow Through (VITA 48.8) one-inch packages with shipments beginning in Q3, 2018.