Jorge Luis Galvez Vallejo
High Performance Computing (HPC) specialist , National Computational Infrastructure (NCI)
Jorge Luis main area of work is enabling scientific codebases to use exascale resources to the max of their abilities. This implies taking traditionally CPU based codebases (serial, shared, or distributed memory) and port the underlying algorithms to use Graphics Processing Units (GPUs) to produce speedups from 5 to 100x depending on the initial codebase architecture.
Jorge luis holds a PhD Physical Chemistry, Iowa State University, Thesis title: An exciting extravagant extraordinarily exceptional exascale experience-from water and bonding, to 27,000 GPUs and beyond! His postdoctoral experience include a Research Fellow at the school of computing.
Historically, he has been in the area of computational quantum chemistry, where he was part of the exascale computing project to enable the GAMESS codebase to run effectively on Intel, AMD, and NVIDIA GPUs. Since moving to the NCI, he has worked in the GPU acceleration of the ocean model MOM6 in collaboration with ACCESS-NRI and NOAA – while also porting the ANUGA flood modelling code to use GPUs in collaboration with CDAC in India. In my spare time I am developing a GPU accelerated hydrodynamic coastal model.
Jorge Luis’ areas of interest in terms of research is the exploration of portable approaches within programming languages for GPU acceleration and cross CPU-GPU compatibility, such as directive based approach or standard parallelism. He has also a passion for software engineering and architecture.