University of Twente Student Theses
Code Green : Evaluating the Carbon and Energy Implications of LLM Integration in Software Development
Belchev, Boris (2025) Code Green : Evaluating the Carbon and Energy Implications of LLM Integration in Software Development.
|
PDF
7MB |
| Abstract: | Large Language Models (LLMs) have demonstrated incredible growth in their capabilities and the opportunities they provide, which has caught the public's attention. The domain of software development is naturally more inclined to bear the fruits of these advancements. However, deploying and using LLMs on a large scale comes with the burden of using more energy and releasing more carbon emissions. Although research has focused on the training costs of these models, there is a gap left uncovered and questions to be answered about the deployment phase. Therefore, this study aims to fill the gap and evaluate the use of LLMs in their inference stage. More specifically, in the context of software development. Through a series of experiments, this study quantifies the energy consumption of coding and general-purpose models across code-related tasks, such as code generation, bug fixing, documentation and testing. The findings provide insights into the trade-offs between accuracy and energy efficiency, helping guide future research and development toward more sustainable and effective LLM deployment. |
| Item Type: | Essay (Master) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 54 computer science |
| Programme: | Computer Science MSc (60300) |
| Link to this item: | https://purl.utwente.nl/essays/105219 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page