Filters








7,804 Hits in 3.1 sec

Efficiently adapting to sharing patterns in software DSMs

L.R. Monnerat, R. Bianchini
Proceedings 1998 Fourth International Symposium on High-Performance Computer Architecture  
In this paper we introduce a page-based Lazy Release Consistency protocol called ADSM that constantly and efficiently adapts to the applications' sharing patterns.  ...  Our main conclusions are that our categorization and adaptation strategies are useful techniques for improving the performance of page-based software DSMs, while ADSM is a highly-efficient option for low-cost  ...  Acknowledgements The authors would like to thank the adaptive TreadMarks group, in particular Cristiana Amza, for letting us borrow the implementation studied in this paper.  ... 
doi:10.1109/hpca.1998.650568 dblp:conf/hpca/MonneratB98 fatcat:rh4w6tk7mbaxtc55xxznm3nj7a

Shared Memory in the Many-Core Age [chapter]

Stefan Nürnberger, Gabor Drescher, Randolf Rotta, Jörg Nolte, Wolfgang Schröder-Preikschat
2014 Lecture Notes in Computer Science  
With the evolution toward fast networks of many-core processors, the design assumptions at the basis of software-level distributed shared memory (DSM) systems change considerably.  ...  But efficient DSMs are needed because they can significantly simplify the implementation of complex distributed algorithms.  ...  The design of efficient software-level DSMs depends a lot on the underlying hardware's structure and the interface to the applications on top of the shared memory.  ... 
doi:10.1007/978-3-319-14313-2_30 fatcat:ah6h3rubebh4zj52w4wfnhaz2m

Dynamic adaptation of sharing granularity in dsm systems

Ayal Itzkovitz, Nitzan Niv, Assaf Schuster
2000 Journal of Systems and Software  
One way to resolve the tradeoff is to dynamically adapt the granularity to the application memory access pattern.  ...  We show that applications may benefit from adapting the memory sharing to the memory access pattern, using both coarse grain sharing and fine grain sharing interchangeably in different stages of the computation  ...  Acknowledgments We are grateful to the anonymous referees, whose comments helped us improve the paper.  ... 
doi:10.1016/s0164-1212(00)00044-3 fatcat:dkf3bst6snftralqg43ozes6om

A nationwide parallel computing environment

Ken Kennedy, Charles F. Bender, John W. D. Connolly, John L. Hennessy, Mary K. Vernon, Larry Smarr
1997 Communications of the ACM  
Through a scalable software environment coupling the small parallelism (a few processors) of the desktop to the large-scale parallelism (thousands of processors) in Alliance machines.  ...  Alliance (NCSA) hopes to accelerate the creation of a nationwide parallel computing environment for the US by developing hardware and software that will ease desktop access to the National Technology Grid  ...  Parallel adaptive methods. The Parallel Computing team is developing technologies to support advanced algorithmic methods, such as adaptive meshing efficiency on DSM architectures.  ... 
doi:10.1145/265684.265693 fatcat:xtfge7keynabjbyxfmdcbm3kci

On the design of global object space for efficient multi-threading Java computing on clusters

Weijian Fang, Cho-Li Wang, Francis C.M. Lau
2003 Parallel Computing  
We leverage runtime object connectivity information to detect distributed-shared objects (DSOs) that are reachable from threads at different nodes to facilitate efficient memory management in the distributed  ...  We report the performance results and give an in-depth analysis of the effects of the proposed adaptive solutions.  ...  Scientific applications exhibit diverse execution patterns. To execute these applications efficiently in software DSM systems, many cache coherence protocols have been proposed.  ... 
doi:10.1016/j.parco.2003.05.007 fatcat:cvo2lmmq5vdfleyhtzysrbwdlm

Improving the performance of software distributed shared memory with speculation

M. Kistler, L. Alvisi
2005 IEEE Transactions on Parallel and Distributed Systems  
Our protocol employs a predictor that uses patterns in past accesses to shared memory to predict future accesses.  ...  We study the performance benefits of speculation in a release consistent software distributed shared memory system.  ...  They also thank Ram Rajamony for many helpful discussions on the Treadmarks DSM system and software DSMs in general. Equipment for the performance evaluation was provided by IBM.  ... 
doi:10.1109/tpds.2005.110 fatcat:3ktjei2yefgbjoo7bkp5vsloxu

Comparing latency-tolerance techniques for software dsm systems

R. Pinto, R. Bianchini, C.L. Amorim
2003 IEEE Transactions on Parallel and Distributed Systems  
This paper studies the isolated and combined effects of several latency-tolerance techniques for software-based distributed shared-memory systems (software DSMs).  ...  More specifically, we focus on data prefetching, update-based coherence, and single-writer optimizations for page-based software DSMs.  ...  ACKNOWLEDGMENTS The research described in this paper was partly supported by FINEP/Brazil grant # 56/94/0399/00 and CNPq/Brazil.  ... 
doi:10.1109/tpds.2003.1247677 fatcat:dbndp7do3bb6veaqukqcori3ia

Lazy Home-Based Protocol: Combining Homeless and Home-Based Distributed Shared Memory Protocols [chapter]

Byung-Hyun Yu, Paul Werstein, Martin Purvis, Stephen Cranefield
2005 Lecture Notes in Computer Science  
We present an in-depth analysis of the effects of the protocol on DSM applications.  ...  This paper presents our novel protocol design and implementation of an all-software page-based DSM system. The protocol combines the advantages of homeless and home-based protocols.  ...  There have been similar ideas of using the two coherence protocols selectively in order to implement a more efficient coherence protocol in a software DSM system.  ... 
doi:10.1007/11557654_84 fatcat:wjmsgr5lrvcnpnszwls3fmrvke

Distributed shared memory: concepts and systems

J. Protic, M. Tomasevic, V. Milutinovic
1996 IEEE Parallel & Distributed Technology Systems & Applications  
She is currently working toward her PhD in the field of DSM. Her research interests are in computer architecture, distributed systems, and performance analysis.  ...  His research interests are computer architectures, multiprocessor systems, and distributed shared-memory systems. He can be reached a t etomasev@ubbg.etf.bg.ac.yu.  ...  We also want to thank Vojislav Protiit for his help in providing up-to-date literature, and Liviu Iftode, who kindly provided some of his most recent papers.  ... 
doi:10.1109/88.494605 fatcat:56jusk7vobepvhcvroadujiiae

MP-LOCKs: replacing H/W synchronization primitives with message passing

Chen-Chi Kuo, J. Carter, R. Kuramkote
1999 Proceedings Fifth International Symposium on High-Performance Computer Architecture  
MP-LOCKs reduce the design complexity and runtime occupancy of DSM controllers and can exploit software's inherent flexibility to adapt to differing applications lock access patterns.  ...  We conclude that locks based on message passing should be considered as a replacement for hardware locks in future scalable multiprocessors that support efficient message passing mechanisms.  ...  We took great pains to conduct a fair comparison by including a detailed 4.4BSD-based kernel in our simulation environment to accurately simulate the software overhead of the proposed message passing mechanisms  ... 
doi:10.1109/hpca.1999.744381 dblp:conf/hpca/KuoCK99 fatcat:egdu26242jatrlgkxg3t7urwpa

Efficient Embedded Software Migration towards Clusterized Distributed-Memory Architectures

Rafael Garibotti, Anastasiia Butko, Luciano Ost, Abdoulaye Gamatie, Gilles Sassatelli, Chris Adeniyi-Jones
2016 IEEE transactions on computers  
This is illustrated in an open-source realistic design framework, including tools from ASIC to microkernel.  ...  A large portion of existing multithreaded embedded sofware has been programmed according to symmetric shared memory platforms where a monolithic memory block is shared by all cores.  ...  ACKNOWLEDGMENTS The research leading to these results has received funding from the European Community's Seventh Framework Programme (FP7/2007-2013) under the Mont-Blanc Project: www.montblanc-project.eu  ... 
doi:10.1109/tc.2015.2485202 fatcat:v2lbaqig5zd6zdr3h5afrhit3u

Merging the Publish-Subscribe Pattern with the Shared Memory Paradigm [chapter]

Loïc Cudennec
2018 Lecture Notes in Computer Science  
Both the publish-subscribe and the shared memory models offer such abstraction. However they are intended to be used in different application contexts.  ...  In this paper we propose to merge these two models into a new one.  ...  Software-Distributed Shared Memory (S-DSM) is used to federate distributed physical memory and provide a high level of abstraction to the application.  ... 
doi:10.1007/978-3-030-10549-5_37 fatcat:tabsh73brrckxks67scwgto3fy

Combined compile-time and runtime-driven, pro-active data movement in software DSM systems

Seung-Jai Min, Rudolf Eigenmann
2004 Proceedings of the 7th workshop on Workshop on languages, compilers, and run-time support for scalable systems - LCR '04  
We present a new compiler algorithm to detect such repetitive data references and an API to an underlying software distributed shared memory (software DSM) system to orchestrate the learning and pro-active  ...  reuse of communication patterns.  ...  Introduction Software Distributed-Shared-Memory (software DSM) Systems have been shown to perform well on a limited class of applications [3] [12] .  ... 
doi:10.1145/1066650.1066661 fatcat:d36vuh3b5ve7nhjef6vehltxwq

Heaps and stacks in distributed shared memory

M. Pizka, C. Rehn
2002 Proceedings 16th International Parallel and Distributed Processing Symposium  
Software-based distributed shared memory (DSM) systems do usually not provide any means to use shared memory regions as stacks or via an efficient heap memory allocator.  ...  Stack and heap organization is adapted to the changed requirements in DSM environments and both, stacks and heaps, are transparently placed in DSM space by the operating system.  ...  Introduction Software-based distributed shared memory (DSM) systems do usually not provide any means to use shared memory regions 1 as stack or heap memory.  ... 
doi:10.1109/ipdps.2002.1016494 dblp:conf/ipps/PizkaR02 fatcat:cxjxmlku6vdgtlr7lvkmo2qya4

Programming Distributed Memory Sytems Using OpenMP

Ayon Basumallik, Seung-Jai Min, Rudolf Eigenmann
2007 2007 IEEE International Parallel and Distributed Processing Symposium  
We present key concepts and describe techniques to analyze and efficiently handle both regular and irregular accesses to shared data.  ...  patterns.  ...  Software DSMs typically adapt a page-based coherence mechanism that intercepts accesses to remote data, at runtime, and requests the data from its current owner.  ... 
doi:10.1109/ipdps.2007.370397 dblp:conf/ipps/BasumallikME07 fatcat:cdpbjy7ghndcxa6kh5zlryc6q4
« Previous Showing results 1 — 15 out of 7,804 results