A deep dive as we love it on ecoCompute 
Max talk features two stages. First he explores a deep dive into how to detect energy hotspots in code.
- Code benchmarking is not always straight forward as modern software has ofte 400+ configuration options and behaviour is starkly different depending on configuration
- By creating a system with 3.5 kHz sampling on component basis they analyse a dataset of 10+ open source tools along with different configurations
- To analyze software on the code level they train an ML model that maps execution paths to energy readings
In the second part they interview developers and their bug fixing strategies. Effectively investigating how, how quick and with what strategy they approach a problem. This helps to create better tools to aid developers making more energy efficient code.
If you liked the Sneak Peak you can find the full slides on the ecoCompute website: https://www.eco-compute.io/files/slides_2025/01_Thursday/01_SoHa/13_Weber_Green_Configuration.pdf