A Case for Task Sampling Based Learning for Cluster Job Scheduling

A Case for Task Sampling Based Learning for Cluster Job Scheduling

USENIX via YouTube Direct link

History-based Learning: Assumptions and Reality

5 of 17

5 of 17

History-based Learning: Assumptions and Reality

Class Central Classrooms beta

YouTube videos curated by Class Central.

Classroom Contents

A Case for Task Sampling Based Learning for Cluster Job Scheduling

Automatically move to the next video in the Classroom when playback concludes

  1. 1 Authors Introduction
  2. 2 Challenges in Cluster Scheduling
  3. 3 Learning Runtime Properties for Cluster Scheduling
  4. 4 Widely Used Approach for Learning: History-based Learning
  5. 5 History-based Learning: Assumptions and Reality
  6. 6 Poor Performance of the State-of-the-Art History-based Predictor
  7. 7 SLearn: A Novel Approach for Learning Runtime Properties
  8. 8 Learning in Time History vs Learning in Space SLearn
  9. 9 Comparing Prediction Accuracy: Large Scale Trace-based Analysis
  10. 10 Comparing Coefficients of Variations CoVs across Space and Time
  11. 11 Varying the History Length in CoV comparison
  12. 12 Comparing Prediction Overhead: Simulation and Testbed Experiments Using GS
  13. 13 SLearn's Implementation and Design
  14. 14 Baselines and Experimental Setup
  15. 15 Simulation and Testbed Experimental Results
  16. 16 SLearn for DAG and Future Work
  17. 17 SLearn Summary

Never Stop Learning.

Get personalized course recommendations, track subjects and courses with reminders, and more.

Someone learning on their laptop while sitting on the floor.