COMS 5270: High-Performance Deep Learning
(Cross-listed with CPRE 5270).
Credits: 3. Contact Hours: Lecture 3.
Prereq: COMS 3520 or CPRE 3080
Fundamentals and advances in HPC in the context of GPUs, TPUs, multicore and HPC systems with specific focus on parallel programming models. Discussing of high-performance computing, GPGPU, scaling deep neural network training and machine learning algorithms, high-performance deep learning, intersection of HPC and AI for HPC, distributed parallel training, engineering parallel software and parallel design patterns.