The core Flight System (cFS) High Performance Computing Framework (HPCF) IRAD will develop the flight software components to efficiently manage computing resources on upcoming high-performance flight computer architectures including:
These architectures feature a more reliable/hardened real time subsystem coupled with a less reliable but very high-performance multi-core subsystem. The goal of the cFS High Performance Computing Framework is to run the flight proven cFS flight software on the real time subsystem along with the necessary components to start, monitor, and control computing jobs on the high-performance computing subsystem.
The development of high-performance space computing architectures such as the HPSC Chiplet and the Spacecube v3.0 are part of a growing trend to dramatically increase compute power available on space data systems. The HPSC Chiplet will provide eight 64-bit ARM cores, and the Spacecube v3.0 will provide four 64-bit ARM cores. Both of these systems are capable of running a wide range of software and mature device drivers on the Linux operating system. Millions of lines of code are in use for high speed networking applications, vision processing, artificial intelligence, autonomous navigation, and parallel processing applications just to name a few.
The principal idea behind this project is to allow this library of Linux software to run on the high-performance spacecraft processors, without requiring a prohibitively expensive conversion of the software, or sacrificing the reliability of the host spacecraft. This goal is accomplished by running the proven Core Flight System (cFS) flight software on the real time subsystems of the HPSC Chiplet, Spacecube v3.0, or MARES architectures, and providing a framework that allows the cFS to power up, launch, monitor, and report results from less trusted computing jobs on the high-performance computing subsystem.
This IRAD will deliver a completed cFS High Performance Computing Framework within two years, which will be ready for infusion for the upcoming high-performance flight computer architectures such as the HPSC, Spacecube v3.0, and MARES SmallSat/Cubesat platform. It is expected that the software will be delivered as feature complete, but that the first experiments and missions that adopt the framework will have to certify the cFS components to the level necessary for the mission.
More »The cFS High Performance Computing Framework has the potential to provide significant software development cost savings without compromising reliability, allowing missions to take full advantage of the advanced flight computer architectures and the vast library of scientific computing software available. This is especially important with Cubesat/SmallSat missions where large software development and certification efforts are cost prohibitive. The ability to use the vast library of Linux based software will allow missions to provide advanced flight functionality promised by these new flight architectures. Potential users of this framework include Spacecube v3.0 based experiments, HPSC/SPLICE flight technology demonstrations, and Cubesat/Smallsat missions.
This is a game-changing technology in the way that it provides a synergistic approach to deploying flight software on upcoming high-performance computing architectures. The cFS framework is a perfect fit for the real time, reliable, more resource constrained subsystems, while the high-performance subsystems provide workstation levels of computing resources.
More »Organizations Performing Work | Role | Type | Location |
---|---|---|---|
Goddard Space Flight Center (GSFC) | Lead Organization | NASA Center | Greenbelt, Maryland |