The Impact of Resource Sharing on Performance and Performance Prediction
Time and Place
The kick-off meeting will take place on Friday, April 20th at 2pm in room 328 in building E1 3!
ScheduleWe will meet at 2pm on the following dates:
- May 18th: Please read the problem specifications of all papers by this date. Prepare a paragraph for each paper describing the problems it tries to solve. Come up with a classification of the problems. Based on this first reading round, we will eliminate some of the papers.
- June 15th: By this meeting, you should have read all of those papers completely that were not eliminated on the first meeting. Prepare a paragraph for each paper describing the proposed solution. Try to classify the methods used in the different papers. Can they be applied in a setting where strict worst-case guarantees are required?
- June 22nd
- July 6th
- July 12th-13th: Retreat at Schloss Dagstuhl.
- July 20th
- July 27th
Multi-processor and multi-core architectures use a combination of private and shared resources, to achieve high average-case performance at low implementation cost.
Resource sharing leads to resource contention, several threads/tasks compete for the resource. Resource contention degrades performance. It also makes performance prediction difficult.
In this doctoral privatissimum, we will study the existing literature about the effects of resource sharing on performance and will attempt to understand its impact on performance predictability. The focus is on the timing predictability of architectures as it is needed for static timing analysis of real-time systems.
As part of the privatissimum, we have written a survey on the impact of resource sharing on performance and performance prediction, including a discussion of open problems. This survey will appear in CONCUR 2013.
O. Tickoo, R. Iyer, R. Illikkal, D. Newell: Modeling Virtual Machine Performance: Challenges and Approaches
- R. Iyer, R. Illikkal, O. Tickoo, L. Zhao, P. Apparao, D. Newell: VM3: Measuring, modeling and managing VM shared resources
- R. Nathuji, A. Kansal, A. Ghaffarkhah: Q-Clouds: Managing Performance Interference Effects for QoS-Aware Clouds
- S. Zhuravlev, S. Blagodurov, A. Fedorova: Addressing Shared Resource Contention in Multicore Processors via Scheduling
A. Fedorova, S. Blagodurov, S. Zhuravlev: Managing Contention for Shared Resources on Multicore Processors
- K. J. Nesbit, J. Laudon, J. E. Smith: Virtual Private Caches
- P. Radojkovic, S. Girbal, A. Grasset, E. Quinones, S. Yehia, F. J. Cazorla: On the Evaluation of the Impact of Shared Resources in Multithreaded COTS Processors in Time-Critical Environments
- Y. Koh, R. Knauerhase, P. Brett, M. Bowman, Z. Wen, C. Pu: An Analysis of Performance Interference Effects in Virtual Environments
J.E. Jensen, J.L. Baer: A Model of Interference in a Shared Resource Multiprocessor
- J. Nowotsch, M. Paulitsch: Leveraging Multi-Core Computing Architectures in Avionics
- F. Boniol, H. Casse, E. Noulard, C. Pagetti: Deterministic Execution Model on COTS Hardware
- A. Schranzhofer, R. Pellizzoni, J.-J. Chen, L. Thiele, M. Caccamo: Timing Analysis of Resource Access Interference on Adaptive Resource Arbiters
- Y. Xie, G. H. Loh: PIPP: Promotion/Insertion Pseudo-Partitioning of Multi-Core Shared Caches
- X. Zhang, S. Dwarkadas, K. Shen: Towards Practical Page Coloring-based Multi-core Cache Management
- L. Zhao, R. Iyer, R. Illikkal, J. Moses, S. Makineni, D. Newell: CacheScouts: Fine-Grain Monitoring of Shared Caches in CMP Platforms
- V. Suhendra, T. Mitra: Exploring Locking & Partitioning for Predictable Shared Caches on Multi-Cores
- N. Guan, M. Stigge, W. Yi, G. Yu: Cache-Aware Scheduling and Analysis for Multicores
- Stefan G. Berg: Cache Prefetching
- D. Eklov, N. Nikoleris, D. Black-Schaffer, E. Hagersten: Cache Pirating: Measuring The Curse of the Shared Cache