Xeon-D Vs Xeon-E for embedded radar applications
September 13, 2017
Embedding Xeon E devices in to military applications requires rugged packaging, reliable/efficient cooling, fast and unrestricted pipes and banks of memory.
For airborne radars, longer, higher and further means more processing power in smaller, rugged efficient packages. The best implementations of ANSI/VITA 65 (OpenVPX), the de facto embedded military open system compute architecture meets the ruggedness and compact requirements, and adds scalability. Especially when implemented as VPX-REDI (VITA 48) even greater ruggedness and increased functional density is possible, as is the ease of two-level maintenance (2LM). Processing power is achieved by leveraging the best commercial Intel Xeon data-center compute capability. Xeon processors are available as either “mobile” devices (Xeon D) which are designed for laptop applications which require lower power, less cores, lower memory bandwidth and less connectivity. Xeon E processors power data-centers and cloud facilities worldwide. Such processors typically have larger core counts, faster memory and increased connectivity such as QPI enabling efficient use of multiple on-board processor and SMP.
Embedding Xeon E devices in to military applications requires rugged packaging, reliable/efficient cooling, fast and unrestricted pipes and banks of memory. Mercury’s proven (fourth generation) OpenVPX Xeon E powered blades have these enabling technologies and are known as the Ensemble® HDS (High Density Server) series of blades., How do these embedded blades with data-center performance usher in the next generation of radar systems? In the past Space Time Adaptive Processing (STAP) was a challenge for embedded systems. This white paper studies STAP processing approaches, using Xeon D and E processors for comparison.