EmbeddingGemma 2: an open, lightweight multimodal embedding model
deepmind.google·6d ago
AI Engineering
Lobsters·Sep 10·5 min read
Some (mostly historical) issues with the Unix load average
TL;DR
Unix load average has evolved from a simple concept in single-CPU environments to a more complex metric in multi-CPU and multi-threaded systems, leading to historical inconsistencies. Understanding these changes is crucial for accurately interpreting system load today.
✦ Why It Matters
Review your system's load average interpretation to ensure accurate performance monitoring.
Key Takeaways
- Recognize that load average reflects active processes, not CPU usage percentage.
- Check how your Unix variant calculates load average for accurate performance insights.
- Be aware of thread visibility in load calculations for accurate monitoring.
- Consider I/O wait times in load average for a comprehensive system load assessment.
How It Works
The Unix load average measures the number of processes in a runnable state, which can include those waiting for CPU time and I/O operations. In modern systems, this metric is calculated globally, meaning it reflects the total number of active processes regardless of the number of CPUs available.
This shift from a simple count to a more complex interpretation allows for better performance monitoring in multi-threaded environments.