Hafiz Adnan Niaz

My PhD research focused on energy measurement and modelling for parallel hybrid applications on heterogeneous CPU/GPU platforms. I developed one of the first methodologies for accurately measuring the energy cost of inter-device data transfers and introduced systematic techniques for constructing runtime energy predictive models. I also introduced the first concurrent and orthogonal software power meters for accurate runtime profiling of computation and communication energy in parallel hybrid applications. This work advances energy-aware high-performance computing and contributes to the development of sustainable computing infrastructures.