Anzeige
Erschienen in:
2019 | OriginalPaper | Buchkapitel
13. Class D Results and Simulations
verfasst von : Art Sedighi, Milton Smith
Erschienen in: Fair Scheduling in High Performance Computing Environments
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Abstract
Class D workloads represent cases in which one user submits a small, noise-like workload, and one or more disproportionately larger users submit much larger workload. In Simulation 5, one user submits a large workload with 5000 tasks, and another user submits 1 task for every two PCs up to a total of 10 tasks. This submission pattern continues, with 6 users in Simulation 9, 11 users in Simulation 13, and 21 users in Simulation 17. In Simulation 9, users 1–5 submit 1000 tasks each for a total of 5000 tasks. In Simulation 13, users 1–10 submit 500 tasks each, while user 11 submits 5000 tasks on their own. In Simulation 17, users 1–20 submit 250 tasks each.