This repository provides a framework to model Large Language Models on dedicated, single-core accelerators and facilitates early identification of energy bottlenecks within the hardware architecture. It is built upon the ZigZag Design Space Exploration tool and inherits its hardware definition format.
Energy Cost Modelling for Optimizing Large Language Model Inference on Hardware Accelerators
Robin Geens, Man Shi, Arne Symons, Chao Fang, Marian Verhelst
Paper: https://ieeexplore.ieee.org/document/10737844
$ python3 -m venv env
$ source env/bin/activate
$ pip install -r requirements.txt
$ python main.py
The result should look like this:
