Return to Video

Why is Dynamic Parallel Quicksort is More Efficient - Intro to Parallel Programming

  • 0:00 - 0:02
    Okay, final quiz.
  • 0:02 - 0:07
    What I've got here is a list of potential reasons why the dynamic parallel version of quicksort
  • 0:07 - 0:10
    ends up more efficient than the CPU-controlled one.
  • 0:10 - 0:15
    We've got more efficient partitioning of the data inside the sort.
  • 0:15 - 0:20
    We've got launching on the fly from the GPU rather than returning to the CPU to do the launch.
  • 0:20 - 0:24
    Simpler code, because the code is a lot shorter and easier to write,
  • 0:24 - 0:27
    and greater GPU utilization.
  • 0:27 - 0:29
    Remember, that GPU utilization is making sure that
  • 0:29 - 0:33
    as much of GPU is busy doing useful work at any given time.
  • 0:33 - 0:40
    So as long as you've got enough threads going on at one time to keep the GPU busy, you have high utilization.
  • 0:40 - 0:42
    So check all the boxes that apply.
Tytuł:
Why is Dynamic Parallel Quicksort is More Efficient - Intro to Parallel Programming
Opis:

more » « less
Video Language:
English
Team:
Udacity
Projekt:
CS344 - Intro to Parallel Programming
Duration:
0:43

English subtitles

Revisions Compare revisions