Adapting Programs for Deep Code Model

Empirical Study on GPU Programming Bugs

Numerical Issues in GPU Programming 

Multi GPU-Programming Pros and Cons: A Case Study
May 2020 - Jul 2020 

A study about advantages and disadvantages of using Multi-GPU programming vs Single GPU programming vs Sequential programming. The modern-day supercomputers have the ability to shrink large numbers and solve tasks in seconds. GPU programming using CUDA is one of the areas that has the ability to solve computations involving large numbers within seconds. The research is a study and analysis on how the Multiple GPUs can perform in solving problems for large numbers rather than computing using a single GPU or traditional computer programs. We have been able to evaluate the pros and cons of using multiple GPUs and how the change in performance evaluation affects with the increase of number of data sets.  GitHub