The Rockets carrier board features a sophisticated 6-layer design, carefully crafted to fit both the Lattepanda Mu N100 and the V100, which is based on the SXM2 interface, into a compact form factor. 

Although the V100 has been discontinued, it remains widely available, making this carrier board a valuable tool for leveraging the full capabilities of the V100. 

By utilizing the Lattepanda Mu N100, this solution allows users to take full advantage of CUDA capabilities within an x86 environment. Whether running Windows or Linux, users can harness the power of this system for demanding AI and computational tasks. 

The design is optimized to ensure seamless integration and efficient performance. We will provide complete PCB design files, enabling easy reproduction and customization for your specific needs. This project aims to deliver a powerful, portable, and flexible AI solution for developers and engineers alike.

This project offers the opportunity to build an AI system with 32GB of VRAM for under $1000, and we hope to see more interesting AI applications on this system.

This project has received technical support from the Lattepanda team, as well as strong support from Minerval, Luke, hqchip, and others. We would like to express our sincere gratitude.