Loading presentation...

Present Remotely

Send the link below via email or IM


Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.


Improving Process Scheduling Though Machine Learning Techniques

No description

Mark Burton

on 4 December 2013

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Improving Process Scheduling Though Machine Learning Techniques

When developing process scheduling techniques, we attempt to create a system in which every process receives balanced and consistent attention from the CPU. We do this so that processes may reach completion in a timely manner and without causing an undesirable number of context switches. While these scheduling techniques offer general solutions, they do not take into account the processes taking place on particular machines. Were the operating system to employ machine learning techniques in is scheduling, it might be able to optimize itself for use by particular programs and users.
Improving Process Scheduling
Though Machine Learning

Learning From the Past
Live Evaluations
Learning From Your Peers
Application Similarity
Historical Profiler Example
Researcher: Richard Gibbons

Profiler Object
- stores and retrieves historical dat
Execution Time Objects
- stores information about specific jobs and is returned whenever the Scheduler queries about specific jobs

Even imperfect records lead to improved performance.

However, must maintain permanent memory usage.
Researchers: Warren Smith,
Valerie Taylor, and Ian Foster

Developed a method for deriving the run times of parallel applications from the run times of similar applications that have executed in the past.

Resulted in better estimations than Gibbons method.
Still Works ^^^
Researchers: Hao Shen, Ying Tan, Jun Lu, Qing Wu, and Qinru Qiu

Developed a learning system that operates at runtime,
requiring no past knowledge.
Full transcript