Uppsala 31 March 2025 – Rapidity Space is excited to announce a collaboration with Frontgrade Gaisler to evaluate and demonstrate the SIMD (Single Instruction, Multiple Data) capabilities of the GR765 multi-processor System-on-Chip (SoC) platform. This initiative aims to showcase how SIMD acceleration can unlock new performance gains for AI and other vectorizable/data-parallel workloads in space applications where processing power is a critical resource.
Frontgrade Gaisler, a world leader in embedded computer systems for space and harsh environments, is currently developing the radiation-hardened GR765, a highly versatile SoC designed to support both NOEL-V RISC-V and LEON5 SPARCv8e processor cores. Through this collaboration, Rapidity Space will conduct benchmarking and develop demonstrators that highlight how the GR765’s SIMD extensions can enhance performance in inference, signal processing, and scientific computing for space systems. These benchmarks and demonstrators will provide key insights, helping to optimize onboard processing for missions that demand real-time decision making and advanced data processing.
Rapidity Space brings extensive expertise in high-performance computing, compiler toolchains, and embedded system optimization to this collaboration. The company will lead efforts to evaluate SIMD acceleration by developing benchmarks and creating demonstrators that illustrate the efficiency of SIMD-based processing in relevant AI/ML and other computational tasks. The development work will be carried out in alignment with Frontgrade Gaisler’s GR765 roadmap, ensuring that the efforts provide clear, data-driven insights into the processor’s capabilities. The work will also involve rigorous testing and performance analysis to demonstrate the practical benefits of SIMD extensions for space application end users.
