Software Tools for Rapid Algorithm Development on HPC, Data-Processing Hardware

Current and future NASA missions are confronted with the challenge that the amount of data collected by its sensors significantly outpaces the down-link bandwidth. As a consequence, initial data reduction has to occur on-board, using algorithms sophisticated enough to sufficiently reduce the amount of data while simultaneously ensuring the preservation of valuable information. This requires enhanced processing power on-board the spacecraft/device/telescope. While future generations of rugged, radiation-hardened hardware can be expected to deliver the necessary performance by employing highly heterogeneous architectures consisting of a combination of FPGAs, DSPs and massively multicore architectures, software development for such devices poses a significant challenge. We prepose to develop tools that facilitate and accelerate algorithm development on these hardware devices. The tools will enable scientists to prototype algorithms in the High-Level language that they are most familiar with while simultaneously building code for high-performance computational hardware.

Modern computational accelerator technologies like FPGAs and GPUs can be difficult to program efficiently and effectively due to the complex nature of the underlying hardware. Many companies provide libraries with highly specialized algorithms that can be accessed from compiled languages like C/C++/Fortran. However, this presents another barrier for many scientists who generally program in high-level languages like MATLAB and IDL. The tools developed in this project, can make these computational accelerators easily accessible to a much wider range of scientists through open source high level languages.

The software tools in this project can benefit a variety of current NASA missions including the Landsat Continuity Mission (LDCM), or planned missions like HyspIRI, the hyperspectral infrared imager. These missions offer a glimpse of the flood of data that is going to be expected from far-future missions. This is not only a challenge encountered with Earth Science missions, but with all future missions, including the Solar Dynamics Observatory or the James Webb Space Telescope. Mechanisms currently used or investigated to reduce the amount of data prior to transmission to Earth ranges from fairly simple trigger functions as encountered on board of RHESSI or planned for XMM, to more complex cloud detector algorithms on board the Earth Observing System (EOS) mission.

This is early/mid-stage (TRL 4) — the most realistic path in is NASA SBIR/STTR, which funds small businesses and research institutions to develop technology aligned with NASA's needs (equity-free, phased funding). Check whether a current SBIR/STTR solicitation topic overlaps with this project's technology area, or contact the project directly (above) to ask.

None of these are guaranteed paths for this specific project — TechPort itself doesn't have an "apply" button. Reaching out to the contact(s) above with a specific question is usually the fastest way to find out what's actually open.