D. S. Nikolopoulos and T. S. Papatheodorou, A Quantitative Architectural Evaluation of Synchronization Algorithms and Disciplines on ccNUMA Systems: The Case of the SGI Origin2000. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Integrating Non-blocking Synchronisation in Parallel.. - Tsigas, Zhang (2002) (Correct)
....the aim to lower the contention when the system is in a high congestion situation. These implementations give di erent execution times under di erent contention instances. But still the time spend by the processes on the synchronisation can form a substantial part of the program execution time [9, 15, 16, 18, 29]. The reason for this is that typical synchronisation is based on blocking that introduces performance bottlenecks because of busywaiting and convoying. Busy waiting tends to produce a large amount of memory and interconnection network contention. The convoying e ect that takes place when a ....
....hardware coherent machines with physically distributed memory have become very popular for moderate to large scale computing. It is important to evaluate the bene ts of non blocking synchronisation in a range of interesting applications running on top of modern realizations of these systems. In [18] the authors assess the performance and scalability of several software synchronisation algorithms, as well as the interrelationship between synchronisation, multiprogramming and parallel job scheduling. In their evaluation, minor modi cations are applied in the synchronisation code of small ....
D. S. Nikolopoulos and T. S. Papatheodorou, A Quantitative Architectural Evaluation of Synchronization Algorithms and Disciplines on ccNUMA Systems: The Case of the SGI Origin2000.
Integrating Non-blocking Synchronisation in Parallel.. - Tsigas, Zhang (2002) (Correct)
....the aim to lower the contention when the system is in a high congestion situation. These implementations give dioeerent execution times under dioeerent contention instances. But still the time spend by the processes on the synchronisation can form a substantial part of the program execution time [9, 15, 16, 18, 29]. The reason for this is that typical synchronisation is based on blocking that introduces performance bottlenecks because of busy waiting and convoying. Busy waiting tends to produce a large amount of memory and interconnection network contention. The convoying eoeect that takes place when a ....
....hardware coherent machines with physically distributed memory have become very popular for moderate to large scale computing. It is important to evaluate the benets of non blocking synchronisation in a range of interesting applications running on top of modern realizations of these systems. In [18] the authors assess the performance and scalability of several software synchronisation algorithms, as well as the interrelationship between synchronisation, multiprogramming and parallel job scheduling. In their evaluation, minor modications are applied in the synchronisation code of small number ....
D. S. Nikolopoulos and T. S. Papatheodorou, A Quantitative Architectural Evaluation of Synchronization Algorithms and Disciplines on ccNUMA Systems: The Case of the SGI Origin
System Software Support for Reducing Memory Latency on.. - Nikolopulos.. Self-citation (Papatheodorou) (Correct)
....hits against the memory latency wall. This observation motivated intensive research work on providing sophisticated runtime and operating system support for improving the memory performance of parallel applications on DSM systems. The reader is referred to related publications of the authors [9, 10, 11, 12, 13, 14, 15] for more elaborate discussions and analyses of the issues outlined in this paper. Sections 2 through 5 provide a brief overview of this work and Section 6 reports on lessons learned and some planned future work. 2. Memory Management for Lightweight Multithreading Multithreading a shared memory ....
.... explored in the past, the sufficiency of scalable synchronization algorithms, as well as the ability of existent algorithms to actually improve the performance of parallel applications on modern DSM systems is an issue of considerable debate [4] A recent study conducted by the authors [11] indicated that several highly sophisticated synchronization algorithms, such as concurrent queues and tree barriers, fail to scale on a 64processor tightly integrated DSM system. Interestingly, the reason behind this behavior lies on the poor scalability of the elementary synchronization ....
D. Nikolopoulos and T. Papatheodorou. A Quantitative Architectural Evaluation of Synchronization Algorithms and Disciplines on ccNUMA Systems: The Case of the SGI Origin2000.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC