Iterative solutions to linear systems are popular because they lend themselves to easier solutions via parallel computing. Simplifications to the domain decomposition process are frequently made to obtain a code that is simple to write, but whose performance is subpar. This presentation will discuss one approach for improving the parallel performance of one such code.
|Presenter:||Babajide Osatuyi (Undergraduate Student)|
|Time:||1:45 pm (Session III)|
American Democracy Project Lecture: Janet Poppendieck
5 pm - 5:45 pm