Distributed Algorithms with Theoretic Scalability Analysis of Radial and Looped Load flows for Power Distribution Systems

Total Page:16

File Type:pdf, Size:1020Kb

Distributed Algorithms with Theoretic Scalability Analysis of Radial and Looped Load flows for Power Distribution Systems Electric Power Systems Research 65 (2003) 169Á/177 www.elsevier.com/locate/epsr Distributed algorithms with theoretic scalability analysis of radial and looped load flows for power distribution systems Fangxing Li *, Robert P. Broadwater ECE Department Virginia Tech, Blacksburg, VA 24060, USA Received 15 April 2002; received in revised form 14 December 2002; accepted 16 December 2002 Abstract This paper presents distributed algorithms for both radial and looped load flows for unbalanced, multi-phase power distribution systems. The distributed algorithms are developed from tree-based sequential algorithms. Formulas of scalability for the distributed algorithms are presented. It is shown that computation time dominates communication time in the distributed computing model. This provides benefits to real-time load flow calculations, network reconfigurations, and optimization studies that rely on load flow calculations. Also, test results match the predictions of derived formulas. This shows the formulas can be used to predict the computation time when additional processors are involved. # 2003 Elsevier Science B.V. All rights reserved. Keywords: Distributed computing; Scalability analysis; Radial load flow; Looped load flow; Power distribution systems 1. Introduction Also, the method presented in Ref. [10] was tested in radial distribution systems with no more than 528 buses. Parallel and distributed computing has been applied More recent works [11Á/14] presented distributed to many scientific and engineering computations such as implementations for power flows or power flow based weather forecasting and nuclear simulations [1,2]. It also algorithms like optimizations and contingency analysis. has been applied to power system analysis calculations These works also targeted power transmission systems. [3Á/14]. Parallel computing is usually referred to as Instead of using shared memory, a scheme of message computing carried out on a single machine with multiple passing was used. CPUs and shared memory. Distributed computing is This paper discusses distributed algorithms to calcu- usually thought of as a less ‘coupled’ computing carried late load flows for radial and looped power distribution out among multiple, separate machines without shared systems based on tree traverses. The algorithms are memory. implemented in up to 8 workstations distributed in an Parallel schemes for power flow based on sparse Ethernet LAN. Ways in which this work differs from previous works are now considered: matrices have been investigated in previous works [3Á/ 9] for transmission systems. These works employed (1) Previous works presented scalability based on test results. This work not only presents test results of factorization to utilize sparse matrices in parallel scalability, but also presents theoretic scalability formulas computers. Power transmission systems were the target from the point of view of algorithm analysis (AA). AA of these works. The previous work [10] presented a [1,2] is an approach to explore the performance of parallel power flow method for radial distribution algorithms from the point of view of mathematics. systems. Like [3 /9], a Jacobin matrix was employed. Á Scalability, including speedup and efficiency, is the most important measurement of performance for par- allel and distributed algorithms. It is a function of n and p, where n is the number of elements in a power * Corresponding author. Tel.: /1-919-807-5707; fax: /1-919-807- 5060. distribution system, and p is the number of processors E-mail addresses: [email protected], [email protected] (F. Li). involved in parallel or distributed schemes. This paper 0378-7796/03/$ - see front matter # 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0378-7796(03)00021-X 170 F. Li, R.P. Broadwater / Electric Power Systems Research 65 (2003) 169Á/177 explores theoretic scalability analysis and estimates this work are based on the (sequential) algorithms in the scalability for the proposed distributed algorithms. previous work [15Á/19], the advantages of the tree-based Estimations will be checked with test results. approach flow naturally to distributed algorithms. (2) Previous works were based on dedicated parallel/ (4) Previous works focused on power transmission distributed architecture, while this work is based on systems or small distribution systems, while this work Ethernet LANs that are widely available. Some previous focuses on power distribution systems having a much parallel solutions of load flows were implemented in larger system scale. The number of elements (including dedicated and/or expensive high-performance parallel customer loads) in a large-scale power distribution computers [4Á/10], such as the Sequent Symmetry S81, system model may well exceed the number of elements iPSC/2, Transputer. Other previous work utilized spe- in a transmission system model. The previous works [3Á/ cific software like parallel virtual machine [11Á/13] or 9,12Á/14] were tested on transmission systems with at specific virtual architecture [14] like ring and hyper- most two thousand elements. The previous work [10] cube. Here we use existing, generic computer networks handled small radial distribution systems with less than like Ethernet LANs, without any extra hardware or 600 elements. This work is tested with radial and looped software, to fulfill distributed solutions for load flows. distribution systems consisting of approximately 40 000 This makes the proposed algorithms very practical since elements. Ethernet LANs are widely available. The communica- Section 2 presents a general approach for distributed tion delay in a LAN is generally longer than in a specific algorithms of tree-based systems. Section 3 presents a architecture, but there is no additional cost. Fig. 1 shows distributed algorithm for radial load flow. Section 4 the Ethernet LAN architecture. Since there is no shared presents a distributed algorithm for looped load flow. memory or storage, message passing through the Section 5 employs techniques in AA to derive formulas standard TCP/IP protocol is used to exchange informa- of the scalability (speedup and efficiency) of the tion among different machines. Here multi-instruction distributed algorithms discussed in Sections 3 and 4 and multi-data is used in the distributed algorithm. and Section 6 presents test results of a distribution Also, a multithreading approach is employed to fully system with approximately 40 000 elements. utilize the concurrency of hardware. (3) Previous works used matrix-based load flows, while this work uses a tree-based approach for unbalanced, 2. General approach for distributed algorithms of tree- multi-phase distribution systems, including both radial and based analysis looped systems. Previous works employed the sparse/ Jacobin matrix to solve load flow or load flow based 2.1. Controller and workers algorithms for transmission systems [3Á/9,11Á/14] and radial distribution systems [10]. This work utilizes a tree- The distributed algorithm defines two types of based approach to solve the unbalanced, multi-phase processes, controllers and workers, which play different load flow for radial and looped distribution systems, roles in the message passing scheme. A controller is a based on the previous works [15Á/19], especially the process with the following responsibilities: previous work [19]. Compared with the matrix approach, the tree-base . To accept user input approach may reduce the complexity of software . To assign initial data (job) to workers . To invoke workers to execute tasks implementation for distribution systems [20Á/22]. The reason is that the matrix approach needs to model nodes . To send/receive intermediate data to/from workers and edges for the system topology, while the tree-based . To do system-wide, non-intensive calculations approach only needs to model edges. In addition, with . To terminate the algorithm at the appropriate time the tree-based approach, the looped load flow can be and notify workers to stop. viewed as inherited from the radial load flow. Hence, A worker is a process with the following respon- development and implementation works can be reduced sibilities: in practice. Since the distributed algorithms presented in . To receive initial data from a controller . To do intensive calculations upon a controller’s request . To send/receive intermediate data to/from a control- ler . To stop when notified. In the approach here there is only one controller, while there are multiple workers. The controller is Fig. 1. An Ethernet LAN with multiple workstations. mainly responsible for coordinating all workers and F. Li, R.P. Broadwater / Electric Power Systems Research 65 (2003) 169Á/177 171 maybe some non-intensive computations, while workers . Assign (p/1)th feeder to pth processor, (p/2)th are responsible for intensive computations. The con- feeder to (p/1)th processor,...,(2p)th feeder to 1st troller has system-wide knowledge, while a worker has processor. knowledge of only a part of the system, i.e., the job . Assign (2p/1)th feeder to 1st processor, (2p/2)th assigned by the controller. In this sense, a controller feeder to 2nd processor, and so on. behaves like a manager in a company, while a worker . Stop when all k feeders are assigned. behaves like a salesman or an engineer. 2.2. General approach 3. Distributed algorithm for radial load flow Fig. 2 illustrates the above steps using unified model- ing language (UML) notation [23], a standard for software design. In this figure, CController is the controller class and CWorker is the worker class. 3.1. Sequential algorithm based on tree approach Also, c is an object of class CController,andw is an object of class CWorker. Each directed line represents a The term sequential is used here to indicate one message or an activity, ordered by its sequence in time. processor. The radial load flow algorithm used here is The name of each message or activity is self-explana- based on tree traverses [15Á/20]. Compared with tradi- tory. Also, an asterisk (*) indicates a repeated activity. tional load flow algorithms like GaussÁ/Seidel and NewtonÁ/Raphson, this approach suits power distribu- tion systems. First, it is designed for unbalanced, multi- 2.3.
Recommended publications
  • Distributed & Cloud Computing: Lecture 1
    Distributed & cloud computing: Lecture 1 Chokchai “Box” Leangsuksun SWECO Endowed Professor, Computer Science Louisiana Tech University [email protected] CTO, PB Tech International Inc. [email protected] Class Info ■" Class Hours: noon-1:50 pm T-Th ! ■" Dr. Box’s & Contact Info: ●"www.latech.edu/~box ●"Phone 318-257-3291 ●"Email: [email protected] Class Info ■" Main Text: ! ●" 1) Distributed Systems: Concepts and Design By Coulouris, Dollimore, Kindberg and Blair Edition 5, © Addison-Wesley 2012 ●" 2) related publications & online literatures Class Info ■" Objective: ! ●"the theory, design and implementation of distributed systems ●"Discussion on abstractions, Concepts and current systems/applications ●"Best practice on Research on DS & cloud computing Course Contents ! •" The Characterization of distributed computing and cloud computing. •" System Models. •" Networking and inter-process communication. •" OS supports and Virtualization. •" RAS, Performance & Reliability Modeling. Security. !! Course Contents •" Introduction to Cloud Computing ! •" Various models and applications. •" Deployment models •" Service models (SaaS, PaaS, IaaS, Xaas) •" Public Cloud: Amazon AWS •" Private Cloud: openStack. •" Cost models ( between cloud vs host your own) ! •" ! !!! Course Contents case studies! ! •" Introduction to HPC! •" Multicore & openMP! •" Manycore, GPGPU & CUDA! •" Cloud-based EKG system! •" Distributed Object & Web Services (if time allows)! •" Distributed File system (if time allows)! •" Replication & Disaster Recovery Preparedness! !!!!!! Important Dates ! ■" Apr 10: Mid Term exam ■" Apr 22: term paper & presentation due ■" May 15: Final exam Evaluations ! ■" 20% Lab Exercises/Quizzes ■" 20% Programming Assignments ■" 10% term paper ■" 20% Midterm Exam ■" 30% Final Exam Intro to Distributed Computing •" Distributed System Definitions. •" Distributed Systems Examples: –" The Internet. –" Intranets. –" Mobile Computing –" Cloud Computing •" Resource Sharing. •" The World Wide Web. •" Distributed Systems Challenges. Based on Mostafa’s lecture 10 Distributed Systems 0.
    [Show full text]
  • Scalability and Performance Management of Internet Applications in the Cloud
    Hasso-Plattner-Institut University of Potsdam Internet Technology and Systems Group Scalability and Performance Management of Internet Applications in the Cloud A thesis submitted for the degree of "Doktors der Ingenieurwissenschaften" (Dr.-Ing.) in IT Systems Engineering Faculty of Mathematics and Natural Sciences University of Potsdam By: Wesam Dawoud Supervisor: Prof. Dr. Christoph Meinel Potsdam, Germany, March 2013 This work is licensed under a Creative Commons License: Attribution – Noncommercial – No Derivative Works 3.0 Germany To view a copy of this license visit http://creativecommons.org/licenses/by-nc-nd/3.0/de/ Published online at the Institutional Repository of the University of Potsdam: URL http://opus.kobv.de/ubp/volltexte/2013/6818/ URN urn:nbn:de:kobv:517-opus-68187 http://nbn-resolving.de/urn:nbn:de:kobv:517-opus-68187 To my lovely parents To my lovely wife Safaa To my lovely kids Shatha and Yazan Acknowledgements At Hasso Plattner Institute (HPI), I had the opportunity to meet many wonderful people. It is my pleasure to thank those who sup- ported me to make this thesis possible. First and foremost, I would like to thank my Ph.D. supervisor, Prof. Dr. Christoph Meinel, for his continues support. In spite of his tight schedule, he always found the time to discuss, guide, and motivate my research ideas. The thanks are extended to Dr. Karin-Irene Eiermann for assisting me even before moving to Germany. I am also grateful for Michaela Schmitz. She managed everything well to make everyones life easier. I owe a thanks to Dr. Nemeth Sharon for helping me to improve my English writing skills.
    [Show full text]
  • Scalability and Optimization Strategies for GPU Enhanced Neural Networks (Genn)
    Scalability and Optimization Strategies for GPU Enhanced Neural Networks (GeNN) Naresh Balaji1, Esin Yavuz2, Thomas Nowotny2 {T.Nowotny, E.Yavuz}@sussex.ac.uk, [email protected] 1National Institute of Technology, Tiruchirappalli, India 2School of Engineering and Informatics, University of Sussex, Brighton, UK Abstract: Simulation of spiking neural networks has been traditionally done on high-performance supercomputers or large-scale clusters. Utilizing the parallel nature of neural network computation algorithms, GeNN (GPU Enhanced Neural Network) provides a simulation environment that performs on General Purpose NVIDIA GPUs with a code generation based approach. GeNN allows the users to design and simulate neural networks by specifying the populations of neurons at different stages, their synapse connection densities and the model of individual neurons. In this report we describe work on how to scale synaptic weights based on the configuration of the user-defined network to ensure sufficient spiking and subsequent effective learning. We also discuss optimization strategies particular to GPU computing: sparse representation of synapse connections and occupancy based block-size determination. Keywords: Spiking Neural Network, GPGPU, CUDA, code-generation, occupancy, optimization 1. Introduction The computational performance of traditional single-core CPUs has steadily increased since its arrival, primarily due to the combination of increased clock frequency, process technology advancements and compiler optimizations. The clock frequency was pushed higher and higher, from 5 MHz to 3 GHz in the years from 1983 to 2002, until it reached a stall a few years ago [1] when the power consumption and dissipation of the transistors reached peaks that could not compensated by its cost [2].
    [Show full text]
  • Grid Computing: What Is It, and Why Do I Care?*
    Grid Computing: What Is It, and Why Do I Care?* Ken MacInnis <[email protected]> * Or, “Mi caja es su caja!” (c) Ken MacInnis 2004 1 Outline Introduction and Motivation Examples Architecture, Components, Tools Lessons Learned and The Future Questions? (c) Ken MacInnis 2004 2 What is “grid computing”? Many different definitions: Utility computing Cycles for sale Distributed computing distributed.net RC5, SETI@Home High-performance resource sharing Clusters, storage, visualization, networking “We will probably see the spread of ‘computer utilities’, which, like present electric and telephone utilities, will service individual homes and offices across the country.” Len Kleinrock (1969) The word “grid” doesn’t equal Grid Computing: Sun Grid Engine is a mere scheduler! (c) Ken MacInnis 2004 3 Better definitions: Common protocols allowing large problems to be solved in a distributed multi-resource multi-user environment. “A computational grid is a hardware and software infrastructure that provides dependable, consistent, pervasive, and inexpensive access to high-end computational capabilities.” Kesselman & Foster (1998) “…coordinated resource sharing and problem solving in dynamic, multi- institutional virtual organizations.” Kesselman, Foster, Tuecke (2000) (c) Ken MacInnis 2004 4 New Challenges for Computing Grid computing evolved out of a need to share resources Flexible, ever-changing “virtual organizations” High-energy physics, astronomy, more Differing site policies with common needs Disparate computing needs
    [Show full text]
  • Parallel System Performance: Evaluation & Scalability
    ParallelParallel SystemSystem Performance:Performance: EvaluationEvaluation && ScalabilityScalability • Factors affecting parallel system performance: – Algorithm-related, parallel program related, architecture/hardware-related. • Workload-Driven Quantitative Architectural Evaluation: – Select applications or suite of benchmarks to evaluate architecture either on real or simulated machine. – From measured performance results compute performance metrics: • Speedup, System Efficiency, Redundancy, Utilization, Quality of Parallelism. – Resource-oriented Workload scaling models: How the speedup of a parallel computation is affected subject to specific constraints: 1 • Problem constrained (PC): Fixed-load Model. 2 • Time constrained (TC): Fixed-time Model. 3 • Memory constrained (MC): Fixed-Memory Model. • Parallel Performance Scalability: For a given parallel system and a given parallel computation/problem/algorithm – Definition. Informally: – Conditions of scalability. The ability of parallel system performance to increase – Factors affecting scalability. with increased problem size and system size. Parallel Computer Architecture, Chapter 4 EECC756 - Shaaban Parallel Programming, Chapter 1, handout #1 lec # 9 Spring2013 4-23-2013 Parallel Program Performance • Parallel processing goal is to maximize speedup: Time(1) Sequential Work Speedup = < Time(p) Max (Work + Synch Wait Time + Comm Cost + Extra Work) Fixed Problem Size Speedup Max for any processor Parallelizing Overheads • By: 1 – Balancing computations/overheads (workload) on processors
    [Show full text]
  • Multiprocessing and Scalability
    Multiprocessing and Scalability A.R. Hurson Computer Science and Engineering The Pennsylvania State University 1 Multiprocessing and Scalability Large-scale multiprocessor systems have long held the promise of substantially higher performance than traditional uni- processor systems. However, due to a number of difficult problems, the potential of these machines has been difficult to realize. This is because of the: Fall 2004 2 Multiprocessing and Scalability Advances in technology ─ Rate of increase in performance of uni-processor, Complexity of multiprocessor system design ─ This drastically effected the cost and implementation cycle. Programmability of multiprocessor system ─ Design complexity of parallel algorithms and parallel programs. Fall 2004 3 Multiprocessing and Scalability Programming a parallel machine is more difficult than a sequential one. In addition, it takes much effort to port an existing sequential program to a parallel machine than to a newly developed sequential machine. Lack of good parallel programming environments and standard parallel languages also has further aggravated this issue. Fall 2004 4 Multiprocessing and Scalability As a result, absolute performance of many early concurrent machines was not significantly better than available or soon- to-be available uni-processors. Fall 2004 5 Multiprocessing and Scalability Recently, there has been an increased interest in large-scale or massively parallel processing systems. This interest stems from many factors, including: Advances in integrated technology. Very
    [Show full text]
  • Pascal Viewer: a Tool for the Visualization of Parallel Scalability Trends
    PaScal Viewer: a Tool for the Visualization of Parallel Scalability Trends 1st Anderson B. N. da Silva 2nd Daniel A. M. Cunha Pro-Reitoria´ de Pesquisa, Inovac¸ao˜ e Pos-Graduac¸´ ao˜ Dept. de Eng. de Comp. e Automac¸ao˜ Instituto Federal da Paraiba Univ. Federal do Rio Grande do Norte Joao Pessoa, Brazil Natal, Brazil [email protected] [email protected] 3st Vitor R. G. Silva 4th Alex F. de A. Furtunato 5th Samuel Xavier de Souza Dept. de Eng. de Comp. e Automac¸ao˜ Diretoria de Tecnologia da Informac¸ao˜ Dept. de Eng. de Comp. e Automac¸ao˜ Univ. Federal do Rio Grande do Norte Instituto Federal do Rio Grande do Norte Univ. Federal do Rio Grande do Norte Natal, Brazil Natal, Brazil Natal, Brazil [email protected] [email protected] [email protected] Abstract—Taking advantage of the growing number of cores particular environment. The configuration of this environment in supercomputers to increase the scalabilty of parallel programs includes the number of cores, their operating frequency, and is an increasing challenge. Many advanced profiling tools have the size of the input data or the problem size. Among the var- been developed to assist programmers in the process of analyzing data related to the execution of their program. Programmers ious collected information, the elapsed time in each function can act upon the information generated by these data and make of the code, the number of function calls, and the memory their programs reach higher performance levels. However, the consumption of the program can be cited to name just a few information provided by profiling tools is generally designed [2].
    [Show full text]
  • Computer Architecture: Parallel Processing Basics
    Computer Architecture: Parallel Processing Basics Onur Mutlu & Seth Copen Goldstein Carnegie Mellon University 9/9/13 Today What is Parallel Processing? Why? Kinds of Parallel Processing Multiprocessing and Multithreading Measuring success Speedup Amdhal’s Law Bottlenecks to parallelism 2 Concurrent Systems Embedded-Physical Distributed Sensor Claytronics Networks Concurrent Systems Embedded-Physical Distributed Sensor Claytronics Networks Geographically Distributed Power Internet Grid Concurrent Systems Embedded-Physical Distributed Sensor Claytronics Networks Geographically Distributed Power Internet Grid Cloud Computing EC2 Tashi PDL'09 © 2007-9 Goldstein5 Concurrent Systems Embedded-Physical Distributed Sensor Claytronics Networks Geographically Distributed Power Internet Grid Cloud Computing EC2 Tashi Parallel PDL'09 © 2007-9 Goldstein6 Concurrent Systems Physical Geographical Cloud Parallel Geophysical +++ ++ --- --- location Relative +++ +++ + - location Faults ++++ +++ ++++ -- Number of +++ +++ + - Processors + Network varies varies fixed fixed structure Network --- --- + + connectivity 7 Concurrent System Challenge: Programming The old joke: How long does it take to write a parallel program? One Graduate Student Year 8 Parallel Programming Again?? Increased demand (multicore) Increased scale (cloud) Improved compute/communicate Change in Application focus Irregular Recursive data structures PDL'09 © 2007-9 Goldstein9 Why Parallel Computers? Parallelism: Doing multiple things at a time Things: instructions,
    [Show full text]
  • Parallel Programming with Openmp
    Parallel Programming with OpenMP OpenMP Parallel Programming Introduction: OpenMP Programming Model Thread-based parallelism utilized on shared-memory platforms Parallelization is either explicit, where programmer has full control over parallelization or through using compiler directives, existing in the source code. Thread is a process of a code is being executed. A thread of execution is the smallest unit of processing. Multiple threads can exist within the same process and share resources such as memory OpenMP Parallel Programming Introduction: OpenMP Programming Model Master thread is a single thread that runs sequentially; parallel execution occurs inside parallel regions and between two parallel regions, only the master thread executes the code. This is called the fork-join model: OpenMP Parallel Programming OpenMP Parallel Computing Hardware Shared memory allows immediate access to all data from all processors without explicit communication. Shared memory: multiple cpus are attached to the BUS all processors share the same primary memory the same memory address on different CPU's refer to the same memory location CPU-to-memory connection becomes a bottleneck: shared memory computers cannot scale very well OpenMP Parallel Programming OpenMP versus MPI OpenMP (Open Multi-Processing): easy to use; loop-level parallelism non-loop-level parallelism is more difficult limited to shared memory computers cannot handle very large problems An alternative is MPI (Message Passing Interface): require low-level programming; more difficult programming
    [Show full text]
  • Challenges for the Message Passing Interface in the Petaflops Era
    Challenges for the Message Passing Interface in the Petaflops Era William D. Gropp Mathematics and Computer Science www.mcs.anl.gov/~gropp What this Talk is About The title talks about MPI – Because MPI is the dominant parallel programming model in computational science But the issue is really – What are the needs of the parallel software ecosystem? – How does MPI fit into that ecosystem? – What are the missing parts (not just from MPI)? – How can MPI adapt or be replaced in the parallel software ecosystem? – Short version of this talk: • The problem with MPI is not with what it has but with what it is missing Lets start with some history … Argonne National Laboratory 2 Quotes from “System Software and Tools for High Performance Computing Environments” (1993) “The strongest desire expressed by these users was simply to satisfy the urgent need to get applications codes running on parallel machines as quickly as possible” In a list of enabling technologies for mathematical software, “Parallel prefix for arbitrary user-defined associative operations should be supported. Conflicts between system and library (e.g., in message types) should be automatically avoided.” – Note that MPI-1 provided both Immediate Goals for Computing Environments: – Parallel computer support environment – Standards for same – Standard for parallel I/O – Standard for message passing on distributed memory machines “The single greatest hindrance to significant penetration of MPP technology in scientific computing is the absence of common programming interfaces across various parallel computing systems” Argonne National Laboratory 3 Quotes from “Enabling Technologies for Petaflops Computing” (1995): “The software for the current generation of 100 GF machines is not adequate to be scaled to a TF…” “The Petaflops computer is achievable at reasonable cost with technology available in about 20 years [2014].” – (estimated clock speed in 2004 — 700MHz)* “Software technology for MPP’s must evolve new ways to design software that is portable across a wide variety of computer architectures.
    [Show full text]
  • Scalable SIMD-Efficient Graph Processing on Gpus
    2015 International Conference on Parallel Architecture and Compilation Scalable SIMD-Efficient Graph Processing on GPUs Farzad Khorasani Rajiv Gupta Laxmi N. Bhuyan Computer Science and Engineering Department University of California Riverside, CA, USA {fkhor001, gupta, bhuyan}@cs.ucr.edu Abstract—The vast computing power of GPUs makes them In this paper we present techniques that maximize the an attractive platform for accelerating large scale data parallel scalability and performance of vertex-centric graph process- computations such as popular graph processing applications. ing on multi-GPU systems by fully exploiting the available However, the inherent irregularity and large sizes of real- world power law graphs makes effective use of GPUs a resources as follows: major challenge. In this paper we develop techniques that SIMD hardware – The irregular nature of power law greatly enhance the performance and scalability of vertex- graphs makes it difficult to balance load across threads centric graph processing on GPUs. First, we present Warp leading to underutilization of SIMD resources. We address Segmentation, a novel method that greatly enhances GPU the device underutilization problem of a GPU by developing device utilization by dynamically assigning appropriate number of SIMD threads to process a vertex with irregular-sized Warp Segmentation that dynamically assigns appropriate neighbors while employing compact CSR representation to number of SIMD threads to process a vertex with irregular- maximize the graph size that can be kept inside the GPU sized neighbors. Our experiments show that the warp exe- global memory. Prior works can either maximize graph sizes cution efficiency of warp segmentation exceeds 70% while (VWC [11] uses the CSR representation) or device utilization for the well known VWC [11] technique it is around 40%.
    [Show full text]
  • Computing at Massive Scale: Scalability and Dependability Challenges
    This is a repository copy of Computing at massive scale: Scalability and dependability challenges. White Rose Research Online URL for this paper: http://eprints.whiterose.ac.uk/105671/ Version: Accepted Version Proceedings Paper: Yang, R and Xu, J orcid.org/0000-0002-4598-167X (2016) Computing at massive scale: Scalability and dependability challenges. In: Proceedings - 2016 IEEE Symposium on Service-Oriented System Engineering, SOSE 2016. 2016 IEEE Symposium on Service-Oriented System Engineering (SOSE), 29 Mar - 02 Apr 2016, Oxford, United Kingdom. IEEE , pp. 386-397. ISBN 9781509022533 https://doi.org/10.1109/SOSE.2016.73 (c) 2016 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other users, including reprinting/ republishing this material for advertising or promotional purposes, creating new collective works for resale or redistribution to servers or lists, or reuse of any copyrighted components of this work in other works. Reuse Unless indicated otherwise, fulltext items are protected by copyright with all rights reserved. The copyright exception in section 29 of the Copyright, Designs and Patents Act 1988 allows the making of a single copy solely for the purpose of non-commercial research or private study within the limits of fair dealing. The publisher or other rights-holder may allow further reproduction and re-use of this version - refer to the White Rose Research Online record for this item. Where records identify the publisher as the copyright holder, users can verify any specific terms of use on the publisher’s website. Takedown If you consider content in White Rose Research Online to be in breach of UK law, please notify us by emailing [email protected] including the URL of the record and the reason for the withdrawal request.
    [Show full text]