Abstract


The Sanders project researches compiler-based code optimisation for heterogeneous many-core systems. We aim to develop a machine learning based approach to optimise OpenCL programs for performance and energy consumption on heterogeneous systems such as CPU-GPU mixed systems and ARM big.LITTLE architectures.

This project is funded by EPSRC (EP/M01567X/1)

logo

 

Members:


Zheng Wang

Vicent Sanz Marco

Ben Taylor

Project Website


http://scc-sanders.lancs.ac.uk/

Media Coverage


Sciencedaily

Phys.org