The NVIDIA® Jetson™ platform is an ARM64-based AI supercomputer on a module powered by NVIDIA’s Pascal™ and Volta™ architectures. The Jetson offers high performance in a small, power-efficient form factor that’s ideal for intelligent edge devices like robots, drones, smart cameras, and portable medical devices. Concurrent Real-Time offers a set of products especially designed to support the NVIDIA Jetson TX2, TX2i, AGX Xavier and Nano in real-time applications.
For mission-critical applications where microsecond-level response is a must, Concurrent’s RedHawk Linux provides guaranteed real-time response and determinism needed for embedded solutions. On the Jetson, RedHawk assures a maximum process dispatch latency of less than 50 microseconds to ensure critical events are recognized and responded to. RedHawk is the ideal Linux solution for a broad range of server and embedded applications such as hardware-in-the loop simulation, training systems, data acquisition, industrial control and medical imaging systems.
RedHawk Linux users can take full advantage of NightStar, Concurrent’s powerful, integrated tool set for developing time-critical CPU and GPU applications. NightStar tools run with minimal intrusion, thus preserving application execution behavior and determinism. Users can quickly and easily debug, monitor, schedule, analyze and tune applications in real-time. NightStar GUI-based tools reduce test time, increase productivity and lower development costs. Time-critical applications require debugging tools that can handle the complexities of multiple processors and cores, multitask interaction and multi-threading. NightStar’s advanced features enable system builders to solve difficult problems quickly.
RedHawk Linux is specially optimized for real-time CUDA performance. RedHawk’s NVIDIA graphics and GPU drivers are enhanced to allow applications to achieve maximum determinism during image rendering and GPU computation. RedHawk also includes a CUDA Persistent Threads (CuPer) API that improves determinism by launching a CUDA kernel only once, at the start of the application, and allowing it to run until the application ends.
SIMulation Workbench delivers a complete framework for developing and executing real-time hardware-in-the-loop and man-in-the-loop simulations. Its powerful GUI allows users to conveniently configure, start, stop, record and play back simulation runs. SimWB allows complex simulations to be executed on a single multi-processor platform running Concurrent’s RedHawk Linux RTOS. SimWB supports modeling environments such as Simulink, VI-grade, SIMPACK, Adams, AMESim, Dymola, veDYNA, MapleSim, GT-Suite and CarSim.