2,494 Hits in 2.0 sec

Brief Announcement: Update Consistency in Partitionable Systems [article]

Matthieu Perrin, Achour Mostéfaoui, Claude Jard
2015 arXiv   pre-print
We propose a new consistency criterion, update consistency (UC), in which the convergent state must be obtained by a total ordering of the updates, that contains the sequential order of each  ...  Eventual consistency [2] guaranties that all replicas eventually converge to a common state when the participants stop updating.  ...  We can prove that update consistency is universal, in the sense that every object has an update consistent implementation in a partitionable system, where any number of crashes are allowed.  ... 
arXiv:1501.02175v1 fatcat:35i266ip6fflnehts7kwsvpp7u

The Jgroup Distributed Object Model [chapter]

Alberto Montresor
1999 Distributed Applications and Interoperable Systems II  
Aim of Jgroup is to support the development of dependable applications in partitionable distributed systems.  ...  Jgroup consists of a partitionable group communication service that simplifies the cooperation among groups of replicated server objects, and a client-side mechanism to transparently invoke methods on  ...  Examples can be found in areas such as computer-supported cooperative work (CSCW), mobile systems, weak-consistency data sharing.  ... 
doi:10.1007/978-0-387-35565-8_31 fatcat:pqxb5eaf5vdxvj6r5ieegc6kk4

Incrementalization of graph partitioning algorithms

Wenfei Fan, Muyang Liu, Chao Tian, Ruiqi Xu, Jingren Zhou
2020 Proceedings of the VLDB Endowment  
We show that we can deduce A∆ from both vertex-cut and edge-cut partitioners A, retaining their bounds.  ...  Given a (vertex-cut or edge-cut) partition C(G) of a graph G and updates ∆G to G, it is to compute changes ∆O to C(G), yielding a partition of the updated graph such that (a) the new partition is load-balanced  ...  It is a fundamental problem in graph theory, and is crucial to parallel graph systems for supporting computations on large-scale graphs [18, 22, 23, 33, 45] .  ... 
doi:10.14778/3389133.3389142 fatcat:itwgevngkfcalng2zn6vf7wlci

Pragma: an infrastructure for runtime management of grid applications

M. Parashar, S. Hariri
2002 Proceedings 16th International Parallel and Distributed Processing Symposium  
, predictive models for system behavior and application performance, and an agent based control network.  ...  Pragma combines 4 key components: system characterization and abstraction component, application characterization component, active network control, and policybase.  ...  The evaluation consisted of comparing the runtimes and load-balance generated for the system sensitive partitioner with those for the default partitioning scheme.  ... 
doi:10.1109/ipdps.2002.1016579 dblp:conf/ipps/ParasharH02 fatcat:4ws7iextyzc4rfjyzvfiuq7hgi

Scalable Shared-Memory Hypergraph Partitioning [article]

Lars Gottesbüren, Tobias Heuer, Peter Sanders, Sebastian Schlag
2020 arXiv   pre-print
In terms of solution quality, we outperform the distributed hypergraph partitioner Zoltan on 95 just four cores,Mt-KaHyPar is also slightly faster than the fastest sequential multilevel partitioner PaToH  ...  To this end, we present Mt-KaHyPar, the first shared-memory multilevel hypergraph partitioner with parallel implementations of many techniques used by the sequential, high-quality partitioning systems:  ...  We iterate in parallel over all vertices to collect boundary vertices in a shared task queue consisting of thread-local vectors.  ... 
arXiv:2010.10272v2 fatcat:wkh6ar27ifaoljqrwxf3iyirqe

DeDiSys Lite: an environment for evaluating replication protocols in partitionable distributed object systems

S. Beyer, A. Sanchez, F.D. Munoz-Escoi, P. Galdamez
2006 First International Conference on Availability, Reliability and Security (ARES'06)  
Distributed object systems for partitionable systems present a challenge, in that there is a trade-off between availability and consistency.  ...  In the DeDiSys project we aim at allowing this trade-off between consistency and availability to be configurable.  ...  DeDiSys Lite serves as a simulation environment for evaluating replication protocols in partitionable distributed object systems.  ... 
doi:10.1109/ares.2006.44 dblp:conf/IEEEares/BeyerSMG06 fatcat:qj7avn7lrveevjqw5llbaghwgi

The Transis approach to high availability cluster communication

Danny Dolev, Dalia Malki
1996 Communications of the ACM  
A unique large-scale multicast service designed for partitionable operation is examined here.  ...  Various projects of many other students have helped in bringing Transis to its current state.  ...  Acknowledgments We acknowledge the contribution of members of the Transis project: Yair Amir and Shlomo Kramer of the original Transis team who designed and shaped Transis, and those who currently extend the system  ... 
doi:10.1145/227210.227227 fatcat:55sfx6aa7ngtvdc2kaomiilrwm

Automatic parallelization of a class of irregular loops for distributed memory systems

Mahesh Ravishankar, John Eisenlohr, Louis-Noël Pouchet, J. Ramanujam, Atanas Rountev, P. Sadayappan
2014 ACM Transactions on Parallel Computing  
Automatic parallelization of a class of irregular loops for distributed memory systems.  ...  Traditional optimizations that rely on purely static analysis fail to generate parallel code in such cases.  ...  The conjugate gradient (CG) method to solve a linear system of equations consists of five partitionable loops within a convergence loop.  ... 
doi:10.1145/2660251 fatcat:24ghpaagpzbmpgzt2bjb2wi364

Code generation for parallel execution of a class of irregular loops on distributed memory systems

Mahesh Ravishankar, John Eisenlohr, Louis-Noel Pouchet, J. Ramanujam, Atanas Rountev, P. Sadayappan
2012 2012 International Conference for High Performance Computing, Networking, Storage and Analysis  
The inspector captures the data-dependent behavior of the computation in parallel and without requiring complete replication of any of the data structures used in the original computation.  ...  In this paper, we propose an approach for effective parallel execution of a class of irregular loop computations in a distributedmemory environment, using a combination of static and runtime analysis.  ...  CG kernel The conjugate gradient (CG) method to solve linear system of equations consists of several partitionable loops within a convergence loop.  ... 
doi:10.1109/sc.2012.30 dblp:conf/sc/RavishankarEPRRS12 fatcat:nl2q52snera63falotlboplqvu

Strengthening Consistency in the Cassandra Distributed Key-Value Store [chapter]

Panagiotis Garefalakis, Panagiotis Papadopoulos, Ioannis Manousakis, Kostas Magoutis
2013 Lecture Notes in Computer Science  
In this short note we argue that the resulting system fits a wider range of applications, and is more robust and easier to reason about.  ...  The Apache Cassandra system is one such popular store combining a key distribution mechanism based on consistent hashing with eventually-consistent data replication and membership mechanisms.  ...  We improve data consistency over original Cassandra by prohibiting multimaster updates.  ... 
doi:10.1007/978-3-642-38541-4_17 fatcat:pn4svinvubaaxn7fgujksbptvq

An Effective Resource Partitioning Heuristic for Embedded Applications on an MPSoC

Hassan Salamy, Olalekan Sopeju
2014 International Journal of Computer Applications  
An MPSoC is a system consisting of multiple heterogeneous processing cores, memory hierarchies, and communication infrastructure to effectively overcome the power and clock constraints from single core  ...  This article presents effective techniques to partitioning the processing cores and memory budget in an MPSoC among multiple embedded applications possibly entering the system at different times.  ...  The proposed heuristic then updates the list L, the M in and Curr values of the remaining applications based on the remaining SPM budget in the system. It also updates the PRF values.  ... 
doi:10.5120/17040-7350 fatcat:xi2bxvyg2zavzjhzbpxnovv47y

Design and Implementation of the Fiduccia-Mattheyses Heuristic for VLSI Netlist Partitioning [chapter]

Andrew E. Caldwell, Andrew B. Kahng, Igor L. Markov
1999 Lecture Notes in Computer Science  
If used in pass-based partitioners, it incrementally computes the change in gains due to a move, and updates the Gain Container.  ...  A pass consists of legal moves, chosen and applied by the move manager. Within a pass, a partitioner can request that the Move Manager undo some of the moves, i.e. perform inverse moves.  ...  The average cutsize for a at partitioner can increase by rather stunning percentages if the worst combination of choices is used instead of the best combination.  ... 
doi:10.1007/3-540-48518-x_11 fatcat:ztjyifd2zfe6dmpwbn37xkmeqm

Runtime support and compilation methods for user-specified irregular data distributions

R. Ponnusamy, J. Saltz, A. Choudhary, Yuan-Shin Hwang, G. Fox
1995 IEEE Transactions on Parallel and Distributed Systems  
The second mechanism is a conservative method for compiling irregular loops in which dependence arises only due to reduction operations.  ...  This mechanism in many cases enables a compiler to recognize that it is possible to reuse previously computed information from inspectors (e.g., communication schedules, loop iteration partitions, and  ...  The authors would like to thank Chuck Koelbel and Sanjay Ranka for many enlightening discussions about universally applicable partitioners and how to embed such partitioners in compilers.  ... 
doi:10.1109/71.406958 fatcat:soh3ilwnpfe2fejg7s6gw6ju4u

DRONE: a Distributed Subgraph-Centric Framework for Processing Large Scale Power-law Graphs [article]

Xiaole Wen, Shuai Zhang, Haihang You
2019 arXiv   pre-print
Thus, the SC model is less competitive in practice. In this paper, we present an innovative distributed graph computing framework, DRONE (Distributed gRaph cOmputiNg Engine).  ...  Nowadays, in the big data era, social networks, graph databases, knowledge graphs, electronic commerce etc. demand efficient and scalable capability to process an ever increasing volume of graph-structured  ...  Experimental Setup Our in-house test bed is a 4-node system, each node consists of 8 Intel Xeon E7-8830 2.13GHz CPU total of 64 cores and 1TB memory. The system is connected via 1GigE Ethernet.  ... 
arXiv:1812.04380v2 fatcat:sbomvxlnm5cfjecuyt6lcyvpi4

HYPE: Massive Hypergraph Partitioning with Neighborhood Expansion

Christian Mayer, Ruben Mayer, Sukanya Bhowmik, Lukas Epple, Kurt Rothermel
2018 2018 IEEE International Conference on Big Data (Big Data)  
data record in a specific replicated shard.  ...  However, existing hypergraph partitioners that scale to very large graphs do not effectively exploit the hypergraph structure when performing the partitioning decisions.  ...  As no partitioner in group (I) consistently outperforms the other partitioners in terms of partitioning quality, scalability and partitioning time, we decided for the well-established and widely used hypergraph  ... 
doi:10.1109/bigdata.2018.8621968 dblp:conf/bigdataconf/MayerMBER18 fatcat:55qu2nkg5rbsth576narblotxu
« Previous Showing results 1 — 15 out of 2,494 results