Background
Previous flight software (FSW) projects, such as the Cyclone Global Navigation Satellite System (CYGNSS) and the Polarimeter to UNify the Corona and Heliosphere (PUNCH), did not fully utilize the processing power available on the SwRI Centaur Avionics platform. This platform is based on the Frontgrade-Gaisler GR712RC processor, which features two cores capable of operating at up to 100 MHz. However, only one core was operated at 25 MHz during these missions.
This research project aims to fully leverage the GR712RC's multicore architecture for high-speed instrument applications. The project seeks to determine the feasibility of using both processor cores for such tasks, while also characterizing the performance of the processor for future applications requiring intensive processing capabilities.
Approach
The project involves modifying the existing PUNCH flight software to implement an image compression task, which will serve as a benchmark for assessing performance. Three versions of the FSW will be developed for comparison: a single-processor version, a symmetric multiprocessing (SMP) version, and an asymmetric multiprocessing (AMP) version.
In symmetric multiprocessing (SMP), both processor cores work in tandem to share tasks equally, optimizing processing power for parallel workloads. Asymmetric multiprocessing (AMP), on the other hand, dedicates specific tasks to each core, allowing one core to handle real-time operations while the other manages non-time-sensitive tasks.
The modified software will receive images from instrument simulators via SpaceWire connections and will use the open-source openjpeg library to perform image compression. Performance data will be gathered and analyzed to evaluate the capability of the GR712RC to handle high-speed image processing through these multiprocessing techniques.
Accomplishments
The research project is in its early stages, but progress has been made. The openjpeg library has been successfully compiled for the target board and operating system. SpaceWire drivers have been integrated into the FSW, and modifications have been made to the software to enable image compression, aligning with the project’s approach.