Mohammad Anas

PROJECTS:- - Worked with the HPC Team in the GPU enablement, optimizations as well as debugging of existing large-scale Climate Forecasting Scientific (CFS) codes using OpenACC. - Worked on porting single-node Fluid Simulation (FS) program to multi-node multi-GPU on Nvidia PSG cluster. INITIATIVES:- - Completed the code standardization and added the program documentation using Doxygen. - Executed Performance and Scaling analysis and Optimizations - Upgraded CFS code from Fortran 77 to Fortran 90 and added modular paradigm. - Added error checking macros to the programs and debugged various compile time and runtime errors. IMPACT:- - Increased code readability and reusability though code standardization and Modularity. - Observed 4x performance in the Fluid Simulation program. - Observed 10x to 5x performance gain in different modules of CFS code. - Helped team to debug errors quickly and deliver project on time.

PROJECT:- Worked on Nvidia CUDA API for the development of a GPU based efficient parallelization algorithm to numerically solve waveform inversion from geophone measurement. INITIATIVES:- -Introduced a GPU detector which detects and assigns parallel tasks on GPU. -Used heterogeneous computing to make GPU and CPU perform different independent operations in parallel. IMPACT:- - Successfully solved waveform inversion in GPU leading to 9x performance gain. - Further use of asynchronous programming increased the performance gain to 11x and lead to more compute-efficient program. - Co-authored a journal paper, summarizing all the results, with Prof Frank titled “Enhancement of In-plane Seismic Full Waveform Inversion with CPU and GPU Parallelization”.

Courses: • Analysis of Algorithms • Computer Graphics • Distributed Operating System • Computer Network Security • Trustworthy Machine Learning • Advance Datastructures • Natural Language Processing • Deeplearning in Medical Imaging

Relevant Coursework - • Object-Oriented Programming • Operating Systems • Database Management Systems • Computer Networks • Information Security