This repository provides customized Linux kernel builds optimized for Ubuntu and MINT Linux systems. While these distributions are excellent for desktop and development use, their default kernel support is often limited — especially when working with cutting-edge hardware such as NVIDIA GPUs, robotics controllers, and AI inference accelerators.
Due to limited resources, we only maintain kernel builds that are actively used in our internal and partner projects. If you require a specific kernel version or hardware patch support, please contact us directly via email.
📧 Contact: Jon@spmatrix.net
(replace # with @ when sending email)
2025-10-24 — ROS Edition
- Base System: Ubuntu 20.04 LTS (Focal Fossa)
- Kernel Version: Linux 6.6.11
- Target Use: ROS / Robotics middleware and real-time modules
- Compatibility: Verified with NVIDIA drivers v570+ and v580+
- Status: Stable in production (tested on multiple NVIDIA RTX systems)
This kernel is tuned for low-latency robotics workloads and optimized for ROS real-time message scheduling and GPU acceleration.
SPMatrix Linux Kernel Project focuses on maintaining high-performance kernels for:
- Robotics platforms (ROS/ROS 2, Unitree, Jetson, RK3588, etc.)
- Edge computing (AI inference, vision, and speech workloads)
- Industrial applications where stability, determinism, and device compatibility are crucial
Each kernel is built and validated with:
- Optimized GCC / LLVM toolchains
- Optional PREEMPT_RT patches for real-time control
- Verified compatibility with modern NVIDIA and AMD GPU stacks
- Tested on both x86-64 and ARM64 environments
SPMatrix is a technology company founded by former Alibaba engineers, dedicated to advancing edge computing and robotics infrastructure.
Our core mission is to deliver reliable, production-grade software stacks for intelligent devices — combining embedded hardware, AI inference, and secure network communication into unified industrial solutions.
We collaborate globally with partners in:
- Intelligent robotics
- Industrial automation
- AI-powered embedded systems
- Edge cloud orchestration
We welcome collaboration from developers and organizations interested in:
- Custom kernel configurations
- ROS/ROS 2 real-time integration
- Edge AI driver compatibility testing
- Cross-platform build automation
To request a custom kernel or contribute patches, please reach out via email. Include details such as:
- Target hardware (CPU / GPU / SoC)
- Kernel version requested
- Expected use case (ROS, AI, networking, etc.)
All kernel builds and scripts in this repository are distributed under the GPLv2 license, consistent with the upstream Linux kernel licensing model.