Project 2: Janine Liu

[j9liu]

• Repo link
• Implemented all CPU and GPU versions of the Scan and Compact algorithms as discussed in class.
• Collected and graphed the runtimes of these algorithms over a variety of array sizes.
• Discussed the contrasts in performance between the CPU and GPU algorithms.
  • Notable concern: The Naive Scan runs faster than my Work-Efficient Scan at all array sizes tested. I noted that it's because the algorithm takes twice as much time to run, evidenced by the runtimes shown in the NSight Analysis window, and could probably be optimized despite its increased kernel calls.
• Extra Credit: Implemented radix sort with test cases.