45
edits
Changes
N/A
,→Assignment 1
'''Conclusion:'''
Based on the results from profiling 3 different sizes of arrays, we can assume that majority of the time taken to sort the arrays is taken by the O(n^2) algorithms (Bubble, Insertion, and Selection). However, the other sorting algorithms (Heap, Merge, Quick) that are O(n log(n)) are extremely fast even when the size of the arrays are large. As observed from the profile, the elapsed time increased as the size of the array increased. I went from approximately 8 seconds to execute the entire program to 13 minutes to execute.
'''Final Selection: Sorting Algorithms'''
Based on the profiled applications above, we think that the sorting algorithms would benefit a lot from offloading to the GPU. Sorting Algorithms are commonly used in programming and can have a strong impact on the programs speed and efficiency. Since they are so commonly used, we think it would be quite interesting to see if we can speed up the O(n^2) sorting algorithms to potentially match the sorting speed of the O(n log n) algorithms since there won’t much change for the parallelized version of them, as they are already fast.Bubble Sort is an elementary sort, it is the most basic sorting method out there however it is the worst sorting algorithm. By Speeding it up using the GPU, we plan to see how much we can improve this sorting algorithm. Also, we would like to look at Insertion Sort, mainly because we know that it is not possible to do on the GPU. We want to take this opportunity to be innovative and attempt or to the very least find a way to solve this issue. '''Summary of selection:''' Bubble & Insertion Sort
=== Assignment 2 ===
