---
title: "Tech Tip #15: Even tuned code has energy left in it"
description: "An AI agent found savings in code experts already optimised."
url: https://greencode.ai/tech-tips/15-even-tuned-code-has-energy-left/
---

# Tech Tip #15: Even tuned code has energy left in it

**An AI agent found savings in code experts already optimised.**

[Watch the video](https://media.greencode.ai/shorts/even-tuned-code-has-energy-left-a3b82a1cdbd3.mp4)

Video created by [explainapaper.com](https://explainapaper.com/?utm_source=greencode.ai&utm_medium=film&utm_campaign=made-by)

### The problem

It is tempting to assume mature, heavily optimised projects have nothing left to give. GreenCode research still found 4.7% GPU energy in llama.cpp and 28.4% in FFmpeg's 4K transcoding.

### The fix you can use today

- Measure mature code before assuming it is already efficient.
- Look at the hot paths under your real workload, not the benchmark the project tunes for.
- Upstream what you find so everyone running the code saves.

GreenCode researchers found that an AI agent could still save energy in code that experts had already tuned.

### Tools that help

- [GreenCode Constitution skill](https://greencode-constitution.org/?utm_source=greencode.ai&utm_medium=referral&utm_campaign=greencode) (GreenCode researchers): open-source energy principles and a catalogue of 37 energy anti-patterns that an AI coding agent can follow to find and fix waste like this
- [GreenCode Constitution profiler (profile.sh)](https://greencode-constitution.org/skill.md?utm_source=greencode.ai&utm_medium=referral&utm_campaign=greencode) (GreenCode researchers): a script from the Constitution tooling that measures the energy of a command from processor counters, GPU sensors and an optional smart plug, and writes the result as JSON
- [green-languages](https://github.com/PLEnergyDev/green-languages?utm_source=greencode.ai&utm_medium=referral&utm_campaign=greencode) (GreenCode researchers): a command-line tool that measures the energy of whole programs or marked sections of code, across languages

### How GreenCode helps

GreenCode helps tackle this by publishing the Constitution for any team's AI agent to follow. The agent profiles the code, fixes ranked energy anti-patterns, measures again and keeps only the changes that save energy. [More on this](https://greencode.ai/publications/greencode-constitution-agent-principles-for-optimizing-code-and-llm-inference-en/).

### Tags

#profiling#compiled-languages#video

### Sources

- [GreenCode Constitution (Kapenekakis et al., GreenAI workshop, ECML PKDD 2026): llama.cpp 4.7% GPU energy (3.7% total); FFmpeg 4K transcoding 28.4%; data synthesis pipeline 9.4%](https://greencode.ai/publications/greencode-constitution-agent-principles-for-optimizing-code-and-llm-inference-en/)

[All GreenCode Tech Tips](https://greencode.ai/tech-tips/)

#### Key Facts

- Tip **#15**
- For **Developers**
- Category **Code efficiency**
- Format **short video**
![GreenCode](https://greencode.ai/wp-content/uploads/2024/12/greencode-logo-website-small-colours-2.webp)

Want this checked on your own system? Talk to the GreenCode team.

[Get in touch](https://greencode.ai/contact/)
