The NANOhpc-obc from SkyLabs d.o.o. is a High-Performance, Fault-Tolerant Onboard Computer (OBC) designed for the most demanding Low Earth Orbit (LEO) missions requiring advanced onboard processing, artificial intelligence, situational awareness, and cybersecurity capabilities. Built on a 64-bit multicore RISC-V architecture, this space-grade computing platform delivers exceptional processing performance while maintaining a best-in-class performance-to-power ratio.
The OBC features a main processing cluster with four 64-bit RISC-V processors operating at up to 600 MHz per core, complemented by a dedicated RISC-V supervisor core for system monitoring and fault management. A matrix coprocessor accelerates tensor and arithmetic operations, while support for AI accelerators such as Intel® Movidius and Microchip VectorBlox enables advanced AI-at-the-edge applications, onboard analytics, machine learning, and autonomous mission operations.
Designed for high-performance embedded computing, the NANOhpc-obc delivers up to 6,500 CoreMarks at just 1.3 W, achieving 3.125 CoreMarks/MHz and 1.714 DMIPS/MHz. Its memory subsystem includes 64 kB of L1 cache and 2 MB of SECDED-protected L2 memory, along with 2 GB of LPDDR4 main memory. For data storage, the system integrates 128 GB of SECDED-protected NAND Flash mass storage and 256 kB of MRAM with SECDED protection for telemetry and mission-critical data retention.
The platform provides a rich set of communication interfaces, including redundant CAN buses, dual-redundant LVDS channels, redundant 1 Gigabit Ethernet, PCIe Gen2 x4/x2 Root Port, SMBus, and GPIO interfaces. This extensive connectivity enables seamless integration with payloads, sensors, communication systems, and spacecraft subsystems.
Supporting both Linux (Yocto and Buildroot distributions) and real-time operating systems (RTOS), the NANOhpc-obc allows developers to deploy deterministic real-time applications alongside high-level Linux workloads within a coherent multicore architecture. This combination of radiation-tolerant design, AI acceleration capability, high computational density, and low power consumption makes it ideally suited for AI@Edge processing, Earth observation, situational awareness, cybersecurity, autonomous spacecraft operations, and next-generation LEO satellite missions.
