معرفی
Dr. Mary Hall is a Professor in the School of Computing at the University of Utah, specializing in compiler optimization, parallel computing, and high-performance computing (HPC). Her work focuses on autotuning techniques, compiler-driven performance optimization, and minimizing data movement in computations to enhance efficiency. She has contributed significantly to frameworks like Bricks and Peak, advancing code generation for GPUs and block-structured grids. Her research also addresses educational initiatives, such as improving student retention in introductory computing courses and fostering diversity in the computing workforce through NSF-funded programs.
Her research interests span compiler technology, stencil computations, and energy-efficient HPC applications. Key projects include optimizing geometric multigrid methods, developing communication-avoiding algorithms, and integrating machine learning into autotuning. She has led efforts to streamline performance portability across heterogeneous architectures and has published extensively on scheduling languages and compiler-driven optimizations.
Mary Hall’s contributions include advancing data layout strategies for sparse tensors and DNNs, as well as fostering reproducibility in computational research through collaborative NSF REU programs. Her work emphasizes practical tools like ytopt and Rigel, which automate performance tuning for scientific applications. She remains active in both academic and industrial HPC communities, addressing challenges in extreme heterogeneity and scalable computing.
