Ingero is an open-source community-driven research project focused on solving one of the most persistent blind spots in high-performance computing: the gap between the Linux kernel and the application layer.
Born out of independent research into eBPF and GPU orchestration, the project provides low-overhead, kernel-level observability for performance-critical environments.
The Architecture
In environments where every CPU cycle counts, traditional “black box” monitoring falls short. Ingero project uses Extended Berkeley Packet Filter (eBPF) to hook directly into the kernel, monitoring CUDA calls to capture the exact CPU-GPU interactions. By tracing how the CPU feeds data to the GPU, developers can pinpoint resource starvation, latency, and throughput bottlenecks instantly—all without modifying a single line of application code or introducing heavy overhead.
Contact
[email protected]