Home > Speakers >

Dr. Giulio Corradi

Dr. Giulio Corradi is a Xilinx ISM (Industrial Scientific Medical) System Architect based in Munich, Germany. He brings 25 years of experience in management, software engineering and development of ASICs and FPGAs in the fields of control, communication, machine intelligence, DSP algorithms, and functional safety. Prior to joining Xilinx in 2006, Giulio managed several projects involving train communication networking and wireless remote diagnostic systems. Recently, Giulio has been focused on providing analytics for Xilinx edge platforms and real-time mixed criticality systems. Giulio is a recurrent keynote speaker at IEEE conferences.

Zynq ® Ultrascale+™ delivers Deterministic Processing for Mixed Criticality Applications in Industrial, Automotive, and Aviation Markets (2020)

Status: Available Now

Today's market requirements are forcing increased computational requirements across all embedded applications through the use of multi-core SoCs, while simultaneously requiring the preservation of legacy real-time code often developed decades ago for single core processors. Often, the performance limitations of real-time processors lead designers to consider and use application processors to achieve desired performance at expense of determinism and worst case execution time (WCET). This webinar describes how to use the ARM Cortex® A53 application processor cluster in Zynq® Ultrascale+™ to implement real-time asymmetric multiprocessing (RTAMP). This approach results in improving worst case execution time (WCET) and reducing latency by isolating and partitioning the cluster such that software developed for single cores can be reused. Demand for this solution is has skyrocketed in Industrial, Automotive and Avionics applications because software architects strongly prefer to use an application processing cluster like a set of single cores when executing real time code. Shared resources like the level 2 cache and memory controller guarantee performances on average, however worst case execution time is affected by interference amongst cores when accessing shared caches and memories. The combination of programmable logic technology and coloured lockdown concepts for shared cache management in conjunction with the open-source Jailhouse hypervisor make it possible to use Linux and bare-metal isolated applications running independently in the cluster. The overhead introduced by a hypervisor is also reduced, making the overall approach very lean.

Go to Session