Filters








121,063 Hits in 2.5 sec

ParaGraphE: A Library for Parallel Knowledge Graph Embedding [article]

Xiao-Fan Niu, Wu-Jun Li
2017 arXiv   pre-print
We name our framework as ParaGraphE, which provides a library for parallel knowledge graph embedding.  ...  Recently, many methods [1, 5, 3, 2, 6] have been proposed to deal with this problem, but existing single-thread implementations of them are time-consuming for large-scale knowledge graphs.  ...  We name our framework as ParaGraphE, which provides a library for parallel knowledge graph embedding.  ... 
arXiv:1703.05614v3 fatcat:myujopmdijhtldfhii6b2ldlsq

Algorithm Parallelization Using Software Design Patterns, an Embedded Case Study Approach

Robbie Vincke, Sille Van Landschoot, Piet Cordemans, Joan Peuteman, Eric Steegmans, Jeroen Boydens
2013 2013 Eighth International Conference on P2P, Parallel, Grid, Cloud and Internet Computing  
First, an application is implemented with a minimum on algorithm modifications using the Map-Reduce design pattern. Next, the same algorithm is rewritten using the graph theory.  ...  Parallel Design Patterns can help in the migration process from legacy sequential to high-performing parallel code. Therefore we propose a layered model of parallel design patterns.  ...  A sequential variant of the Graph Theory implementation will be used as reference implementation for computing the speedup of the parallel Graph Theory algorithm.  ... 
doi:10.1109/3pgcic.2013.80 dblp:conf/3pgcic/VinckeLCPSB13 fatcat:fwxyknpltzfwfjsbwjqkgyshpe

Improving the expressiveness of deep learning frameworks with recursion

Eunji Jeong, Joo Seong Jeong, Soojeong Kim, Gyeong-In Yu, Byung-Gon Chun
2018 Proceedings of the Thirteenth EuroSys Conference on - EuroSys '18  
In this paper, we add recursion to the programming model of existing frameworks by complementing their design with recursive execution of dataflow graphs as well as additional APIs for recursive definitions  ...  between nodes for efficient execution based on parallel computation.  ...  Analysis of Recursive Graphs: Parallelization The performance difference between the iterative and recursive implementation of the same recursive neural network mainly comes from the parallelization of  ... 
doi:10.1145/3190508.3190530 dblp:conf/eurosys/JeongJKYC18 fatcat:yjt6dn2sw5fdppk664dtgliqg4

From Scilab to multicore embedded systems: Algorithms and methodologies

George Goulas, Panayiotis Alefragis, Nikolaos S. Voros, Christos Valouxis, Christos Gogos, Nikolaos Kavvadias, Grigoris Dimitroulakos, Kostas Masselos, Diana Goehringer, Steven Derrien, Daniel Menard, Olivier Sentieys (+8 others)
2012 2012 International Conference on Embedded Computer Systems (SAMOS)  
This paper presents the methodology and algorithms for the creation of parallel software written in Scilab source code for multicore embedded processors in the context of the "Architecture oriented paraLlelization  ...  for high performance embedded Multicore systems using scilAb" (ALMA) EU FP7 project.  ...  In fact, MIP has been used before for automatic parallelization of embedded software [28] .  ... 
doi:10.1109/samos.2012.6404184 dblp:conf/samos/GoulasAVVGKDMGDMSHSOBRSKM12 fatcat:ldkdwbhc5batjfkzi6vphfvxbq

Exploring Single Source Shortest Path Parallelization on Shared Memory Accelerators

Daniele Palossi, Andrea Marongiu
2016 Proceedings of the 19th International Workshop on Software and Compilers for Embedded Systems - SCOPES '16  
performance/watt, but calls for efficient parallel SSSP implementation.  ...  In this work we provide a detailed exploration of the ∆-stepping algorithm performance on a representative heterogeneous embedded system, TI Keystone II, considering the impact of several parallelization  ...  The emerging trend towards the adoption of low-power parallel accelerators in embedded systems 1 opens new opportunities to deliver superior performance/watt, but calls for efficient parallel implementation  ... 
doi:10.1145/2906363.2915925 dblp:conf/scopes/PalossiM16 fatcat:ptnriyia7fc7xdvfrri2lk4bnm

A Multi-Level Parallel Implementation of a Program for Finding Frequent Patterns in a Large Sparse Graph

Steve Reinhardt, George Karypis
2007 2007 IEEE International Parallel and Distributed Processing Symposium  
The parallel implementation required the addition of 21 directives and about 50 accompanying lines of code, in an original code of about 15,000 lines.  ...  embeddings.  ...  The OpenMP work was greatly accelerated by discussions with John Baron of SGI and Jay Hoeflinger and Grant Haab of Intel, especially concerning the use of the taskq/task constructs.  ... 
doi:10.1109/ipdps.2007.370404 dblp:conf/ipps/ReinhardtK07 fatcat:xqgt4v44xbacdf5q6fn4dpuxki

Pangolin: An Efficient and Flexible Graph Pattern Mining System on CPU and GPU [article]

Xuhao Chen, Roshan Dathathri, Gurbinder Gill, Keshav Pingali
2020 arXiv   pre-print
There is growing interest in graph pattern mining (GPM) problems such as motif counting.  ...  GPM systems have been developed to provide unified interfaces for programming algorithms for these problems and for running them on parallel systems.  ...  Compared to graph analytics, GPM algorithms are more difficult to implement on parallel platforms; for example, unlike graph analytics algorithms, they usually generate enormous amounts of intermediate  ... 
arXiv:1911.06969v2 fatcat:jqj6f2gxxrcylfybqt4mkq6xsm

FlowGNN: A Dataflow Architecture for Universal Graph Neural Network Inference via Multi-Queue Streaming [article]

Rishov Sarkar, Stefan Abi-Karam, Yuqi He, Lakshmi Sathidevi, Cong Hao
2022 arXiv   pre-print
The architecture features a configurable dataflow optimized for simultaneous computation of node embedding, edge embedding, and message passing, which is generally applicable to all models.  ...  Our implementation code and on-board measurement are publicly available on GitHub.  ...  ACKNOWLEDGEMENTS The authors would like to thank Zihang Qiao for his contributions to the development of PNA and DGN, as well as Parima Mehta for her contribution to the virtual node model.  ... 
arXiv:2204.13103v1 fatcat:5mbf62bdp5achp7cx5xb2psz2a

Force2Vec: Parallel force-directed graph embedding [article]

Md. Khaledur Rahman, Majedul Haque Sujon, Ariful Azad
2020 arXiv   pre-print
A graph embedding algorithm embeds a graph into a low-dimensional space such that the embedding preserves the inherent properties of the graph.  ...  We make Force2Vec highly parallel by mapping its core computations to linear algebra and utilizing multiple levels of parallelism available in modern processors.  ...  Other graph embedding and visualization algorithms can be implemented in this framework. • We present a highly-parallel algorithm that uses multicore processors and memory efficiently.  ... 
arXiv:2009.10035v1 fatcat:bm4dieccxnctfm4qcisdon46py

Hardware and Software Implementations of Prim's Algorithm for Efficient Minimum Spanning Tree Computation [chapter]

Artur Mariano, Dongwook Lee, Andreas Gerstlauer, Derek Chiou
2013 IFIP Advances in Information and Communication Technology  
Prim's algorithm is based on graph traversals, which are inherently hard to parallelize.  ...  We study two algorithmic variants and compare their performance against implementations on desktop-class and embedded CPUs.  ...  Parallelism Analysis Depending on the implementation of Prim's algorithm, it can exhibit some parallelism.  ... 
doi:10.1007/978-3-642-38853-8_14 fatcat:wjd3ttgqrfhllchsroeoul34dy

Pangolin

Xuhao Chen, Roshan Dathathri, Gurbinder Gill, Keshav Pingali
2020 Proceedings of the VLDB Endowment  
There is growing interest in graph pattern mining (GPM) problems such as motif counting.  ...  GPM systems have been developed to provide unified interfaces for programming algorithms for these problems and for running them on parallel systems.  ...  Compared to graph analytics, GPM algorithms are more difficult to implement on parallel platforms; for example, unlike graph analytics algorithms, they usually generate enormous amounts of intermediate  ... 
doi:10.14778/3389133.3389137 fatcat:53fbcjljv5hpxa5yeqv4ioviou

PecanPy: a fast, efficient, and parallelized Python implementation of node2vec

Renming Liu, Arjun Krishnan
2021 Bioinformatics  
We have developed PecanPy, a new Python implementation of node2vec that uses cache-optimized compact graph data structures and precomputing/parallelization to result in fast, high-quality node embeddings  ...  Learning low-dimensional representations (embeddings) of nodes in large graphs is key to applying machine learning on massive biological networks.  ...  Mancuso, Anna Yannakopoulos, and the rest of the Krishnan Lab for valuable discussions and feedback on the manuscript. We are grateful to Charles T.  ... 
doi:10.1093/bioinformatics/btab202 pmid:33760066 pmcid:PMC8504639 fatcat:ed6psnd4b5hq5fekiaqfs6aapm

PecanPy: a fast, efficient, and parallelized Python implementation of node2vec [article]

Renming Liu, Arjun Krishnan
2020 bioRxiv   pre-print
We have developed PecanPy, a new Python implementation of node2vec that uses cache-optimized compact graph data structures and precomputing/parallelization to result in fast, high-quality node embeddings  ...  Learning low-dimensional representations (embeddings) of nodes in large graphs is key to applying machine learning on massive biological networks.  ...  Mancuso, Anna Yannakopoulos, and the rest of the Krishnan Lab for valuable discussions and feedback on the software and manuscript.  ... 
doi:10.1101/2020.07.23.218487 fatcat:ukgmp3keobdwngfew3pxwuwluu

Dense edge-disjoint embedding of complete binary trees in interconnection networks

S. Ravindran, A.M. Gibbons, M.S. Paterson
2000 Theoretical Computer Science  
For the mesh and the shu e-exchange graphs each edge is regarded as two parallel (or anti-parallel) edges.  ...  The embeddings have the following properties: paths of the tree are mapped onto edge-disjoint paths of the host graph and at most two tree nodes (just one of which is a leaf) are mapped onto each host  ...  For the mesh and shu e-exchange graphs these embeddings required replacing each edge by a pair of parallel (or anti-parallel) edges. A number of problems remain open.  ... 
doi:10.1016/s0304-3975(00)00066-9 fatcat:zgilbrqsubfbvgyzebbyvajhou

Graph Data Mining with Arabesque

Eslam Hussein, Mohammed Zaki, Abdurrahman Ghanem, Vinicius Vitor dos Santos Dias, Carlos H.C. Teixeira, Ghadeer AbuOda, Marco Serafini, Georgos Siganos, Gianmarco De Francisci Morales, Ashraf Aboulnaga
2017 Proceedings of the 2017 ACM International Conference on Management of Data - SIGMOD '17  
Arabesque provides a simple programming model to express graph data mining computations, and a highly scalable and efficient implementation of this model, scaling to billions of subgraphs on hundreds of  ...  Most current parallel graph analytics systems do not provide good support for graph data mining. One notable exception is Arabesque, a system that was built specifically to support graph data mining.  ...  Arabesque explores the input graph in a series of bulk synchronous parallel (BSP) steps, and maintains a set of candidate embeddings at each step.  ... 
doi:10.1145/3035918.3058742 dblp:conf/sigmod/HusseinGDTASSMA17 fatcat:yfesugez3rcqhjsbusxixb423m
« Previous Showing results 1 — 15 out of 121,063 results