CS 4523 - Programming Massively Parallel Processors

3 Class Hours 0 Laboratory Hours 3 Credit Hours
Prerequisite: CS 3304, CS 3502
A study of practical parallel algorithms with an emphasis on implementation and performance issues on massively parallel processors. Design and implement high performance computing applications using CUDA running on Graphics Processing Unit (GPU). Topics include heterogeneous parallel programming, hardware threading models, synchronization, parallel blocking algorithms, register allocations, memory performance, and inter-thread communication.