Kreps
1.0.0
Breast cancer detection through machine learning
|
This project is done in the context of the Master Calcul Haute Performance et Simulation at the Université de Versailles St-Quentin.
The goal of this project is to implement a tool for breast cancer detection, and optimize it accordingly.
While the first semester focused on the sequential implementation (Versions <= 0.5), the second semester sees us switching to a GPU backend for maximum performance and scalability.
Furthermore, we also this project as an opportunity to explore software architecture on a real scale.
You will also need a blas implementation, and an OpenCL runtime (we recommend POCL for starter) Note that the Cuda toolkit already provides the necessary OpenCL files.
We advise you to run the tests to ensure everything went well:
Raw data will be outputted to <output path>/.
Benjamin Lozes \ Ugo Battiston \ Mathys Jam
Mohammed-Salah Ibnamar, for supervising us during this project \ Thomas Dufaud, for providing us access to a GPU cluster