MultiObjectiveExploration

Table of Contents

• Background
• Install
• Usage
• Maintainers
• License

Background

1. evaluate various prediction models for DSE of CPU.
2. a hypervolume-improvement-based multi-objective optimization method and a uniformity-aware selection algorithm to select design points. Further, AdaBoost is first introduced to GBRT model to improve the prediction accuracy.

Install

This project uses torch, skopt, math, and several basic python packages. Go check them out if you don't have them locally installed.

Usage

This is only a documentation package.

# run main() in main.python
$ python main.python

Major running configurations are defined in config.py Performance metric dataset is listed in data_all_simpoint/ . For now, it only includes a demo sample record in 500.1-refrate-1.txt

Maintainers

@multiobjectiveDSE.

License

MIT