- Parallel Bubble Sort & its implementation

- Sequencing issues will be caused for coalesced hardware when acceding array elements 0 & 1 in thread 0, accessing 2 & 3 in thread 1 (Issue with shared memory & why we avoided it).

- Blocks will run in random orders (One of the reasons why we encountered an error with the single kernel with a for loop inside odd-even approach). We were able make small numbers work & it was much faster but for big numbers approximately 1 percent of the array size would be incorrectly sorted.

- We came across some interesting sorting algorithms (Bucket, radix, rank, hyper quick and merge sort) which are better options than the parallel bubble sort.