Real Time System Simulation

The development of large scale electronic systems, such as vehicles or aircraft, involves system architecture, hardware and software generation, diagnostics and test development, and system integration. Most often, interdependent subsystems are acquired from a wide range of suppliers who use a variety of design methodologies. The result of this disparity is very long system integration cycles in which complex system interaction problems must be diagnosed and debugged. The vast majority of platform-level debugging can only occur when the subsystems are brought together in a system integration facility.

The traditional integration laboratory is constructed of production subsystem hardware, prototype hardware, custom test equipment, computer workstations, cabling, power supplies and cooling systems in an attempt to mimic the behavior of a complete large scale system. These labs are expensive to construct, operate, and maintain. It is impractical to replicate integration labs to allow all of the teams that need to use them to work in parallel. Just like the early days of computers, users must schedule lab time and wait in the queue for their opportunity to do their job, and then step aside to analyze their results while the next user steps in. This part-time access to test and integration facilities makes the system integration process far more inefficient and time consuming than it would be if each team had its own lab.

Compounding the inefficiencies of the traditional integration lab is the fact that the platform subsystems, known as Line Replaceable Units (LRUs), typically provide very limited visibility into their internal operation. At the large scale system level any such visibility is almost non-existent. Lack of visibility makes the process of locating bugs and resolving them exponentially more complicated. It becomes difficult or impossible to predict the amount of time it will take to resolve system integration problems and schedules are typically missed, sometimes dramatically, as a result. In fact, many problems remain undiagnosed and unresolved until well after systems are in production

	A Traditional Integration Laboratory

Platform-Level Models

Engineering teams have had access to models of large scale systems for mechanical and physical analysis for years and have realized significant benefits from these models. But such platform-level models have not been available for electronic systems. Though some tools exist for electronic system modeling at very high levels of abstraction, these models do not represent the actual electronic systems with enough fidelity to run real software or diagnostics, nor do they facilitate problem resolution during system integration. For electronic systems, there is a clear need for a solution that provides much deeper visibility and understanding of a large scale system’s behavior with enough fidelity to run actual software at real time. Such a machine would greatly benefit system architects, software developers, and diagnostics teams, and would lead to a dramatically improved time-to-market and product quality. CPU Tech fills this need with its SystemLab PS™, a revolutionary system modeling environment that enables for the first time the simultaneous simulation of every electronic system on a major platform at real-time speeds.

The SystemLab PS combines the power of CPU Tech’s SuperQ X3 supercomputer running the industry leading SystemLab modeling application. The SuperQ X3 accelerates the SystemLab software to real time speeds. SystemLab models can be created for systems consisting of existing LRUs or new LRU designs or concepts, including any mix of both new and existing LRUs. The models run existing software without modification and new software can be developed and debugged even before the physical implementation of the LRU exists. Platform-level models are achieved through virtual modeling of subsystem interconnections – busses, discrete signals, cables – and a virtualization of the human-machine interface.

Users access the SystemLab PS using workstations at their desks, making it both practical and economical to provide a platform-level model for each team performing development and integration activities.