Job Scheduling Strategies for Parallel Processing

1. Front Matter

   Pages I-IX

   PDF

2. Experience and Practice of Batch Scheduling on Leadership Supercomputers at Argonne

   • William Allcock, Paul Rich, Yuping Fan, Zhiling Lan

   Pages 1-24

3. Analysis of Mixed Workloads from Shared Cloud Infrastructure

   • Dalibor Klusáček, Boris Parák

   Pages 25-42

4. Tuning EASY-Backfilling Queues

   • Jérôme Lelong, Valentin Reis, Denis Trystram

   Pages 43-61

5. Don’t Hurry Be Happy: A Deadline-Based Backfilling Approach

   • Tchimou N’takpé, Frédéric Suter

   Pages 62-82

6. Supporting Real-Time Jobs on the IBM Blue Gene/Q: Simulation-Based Study

   • Daihou Wang, Eun-Sung Jung, Rajkumar Kettimuthu, Ian Foster, David J. Foran, Manish Parashar

   Pages 83-102

7. Towards Efficient Resource Allocation for Distributed Workflows Under Demand Uncertainties

   • Ryan D. Friese, Mahantesh Halappanavar, Arun V. Sathanur, Malachi Schram, Darren J. Kerbyson, Luis de la Torre

   Pages 103-121

8. Programmable In Situ System for Iterative Workflows

   • Erich Lohrmann, Zarija Lukić, Dmitriy Morozov, Juliane Müller

   Pages 122-131

9. A Data Structure for Planning Based Workload Management of Heterogeneous HPC Systems

   • Axel Keller

   Pages 132-151

10. ScSF: A Scheduling Simulation Framework

    • Gonzalo P. Rodrigo, Erik Elmroth, Per-Olov Östberg, Lavanya Ramakrishnan

    Pages 152-173

11. DJSB: Dynamic Job Scheduling Benchmark

    • Victor Lopez, Ana Jokanovic, Marco D’Amico, Marta Garcia, Raul Sirvent, Julita Corbalan

    Pages 174-188

12. Back Matter

    Pages 189-189

    PDF

This book constitutes the thoroughly refereed post-conference proceedings of the 21st International Workshop on Job Scheduling Strategies for Parallel Processing, JSSPP 2017, held in Orlando, FL, USA, in June 2017.
The 10 revised full papers presented in this book were carefully reviewed and selected from 20 submissions. The papers cover topics in the fields of design and evaluation of new scheduling approaches; performance evaluation of scheduling approaches; workloads; consideration of additional constraints in scheduling systems; scaling and composition of very large scheduling systems; cloud provider issues; interaction between schedulers on different levels; interaction between applications/workloads; experience reports from production systems or large scale compute campaigns. 

• Parallel computing
• Parallel processing
• Distributed systems
• Scheduling
• Job scheduling
• System modeling
• Workload modeling
• System evaluation
• Performance optimization
• Simulation
• Cloud computing
• Grid computing
• scheduling algorithms
• resource allocation
• processors
• scheduling problem

• CESNET, Prague, Czech Republic

  Dalibor Klusáček

• Google, Mountain View, USA

  Walfredo Cirne

• Google, Seattle, USA

  Narayan Desai

Policies and ethics