DOCSLIB.ORG

Message Passing Interface for ROOT

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Message Passing Interface for ROOT Message Passing Interface for ROOT Omar Zapata Gerardo Gutiérrez http://oproject.org/ROOTMpi Outline Message Passing Interface ROOTMpi MPI Current status and future Implementation with ROOT Conclusions 2 Message Passing Interface Standard of communication for HPC/Grid Computing Widely used in the scientific community. Implementations OpenMPI, MPICH, IBM, Intel etc.. Support for: RMA (Remote Memory Access) Shared memory Checkpointing P2P and collective communication Fault tolerance under development http://mpi-forum.org/mpi-40 3 ROOTMpi Integration of MPI and ROOT technologies in a framework for parallel computing. Motivation: Communicate ROOT objects through processes. Implement MPI with a better design for ROOT. Create an interface that uses the new C++ features to write parallel code. Implement ROOT algorithms in parallel for HPC/Grid systems. 4 ROOTMpi Prototype is going under development with the next features already implemented Peer to Peer communication Some collective operation implemented Gather/AllGtaher Scatter/AllScatter Redude/AllReduce Bcast Communicators IntraCommunicator InterCommunicator Blocking and non-blocking communication. Tested with OpenMPI and MPICH 5 ROOTMpi Design 6 ROOTMpi Design 7 ROOTMpi Design Comparison with C Pointer Size Predefined Message MPI Tag id to void of the array MPI datatype destination Communicator 8 ROOTMpi Design Comparison with C Send in ROOTMpi Global Template Any serializable Message Communicator Method object destination Tag id 9 ROOTMpi Design 10 Example Peer to Peer 11 Example Peer to Peer output 12 Current Status Feature Description Status TCommunicator Base class for all DONE communcators, to perform p2p and collective TIntraComminicator Class to do Almost DONE communication in the same context TInterCommunicator Class to perform DONE communication between different contexts. TCartCommunicator Cartesian TODO Communicator TGraphComminicator Graph communicator TODO 13 Current Status Feature Description Status Request Classes Classes to handle non- DONE blocking communication TEnvironment Class initialize/finalize DONE communication system TInterCommunicator Class to perform DONE communication between different contexts. Checkpointing Feature to save the TODO need design state of the execution and restore it. Fault Tolerance Feature to continue TODO MPI-4 comming execution in case of soon. fault. 14 Current Status FeatureCl Description Status Serialization for Made with the classes DONE communication TmpiMessage and TMpiMessageInfo Error Handling MPI error are very hard TODO need design to understant also for MPI experts, improbe the error handling. MPI Files Class manage TODO need design distribute file Memory Window Class to shared regions TODO need design of memory with Remote Access Memory 15 Conclusions MPI standard have all need HPC/Grid computing Shared/Distributed memory Checkpointing Fault tolerance under development ROOTMpi is A modern interface for MPI that uses powerful C++ design A great communication system through serialization. An easy way to parallelize codes in ROOT for HPC/Grid computing. 16 More Information Website http://oproject.org 17 To finish Thanks ! 18.
Load more

Recommended publications
  • 2.5 Classification of Parallel Computers
    52 // Architectures 2.5 Classification of Parallel Computers 2.5 Classification of Parallel Computers 2.5.1 Granularity In parallel computing, granularity means the amount of computation in relation to communication or synchronisation Periods of computation are typically separated from periods of communication by synchronization events. • fine level (same operations with different data) ◦ vector processors ◦ instruction level parallelism ◦ fine-grain parallelism: – Relatively small amounts of computational work are done between communication events – Low computation to communication ratio – Facilitates load balancing 53 // Architectures 2.5 Classification of Parallel Computers – Implies high communication overhead and less opportunity for per- formance enhancement – If granularity is too fine it is possible that the overhead required for communications and synchronization between tasks takes longer than the computation. • operation level (different operations simultaneously) • problem level (independent subtasks) ◦ coarse-grain parallelism: – Relatively large amounts of computational work are done between communication/synchronization events – High computation to communication ratio – Implies more opportunity for performance increase – Harder to load balance efficiently 54 // Architectures 2.5 Classification of Parallel Computers 2.5.2 Hardware: Pipelining (was used in supercomputers, e.g. Cray-1) In N elements in pipeline and for 8 element L clock cycles =) for calculation it would take L + N cycles; without pipeline L ∗ N cycles Example of good code for pipelineing: §doi =1 ,k ¤ z ( i ) =x ( i ) +y ( i ) end do ¦ 55 // Architectures 2.5 Classification of Parallel Computers Vector processors, fast vector operations (operations on arrays). Previous example good also for vector processor (vector addition) , but, e.g. recursion – hard to optimise for vector processors Example: IntelMMX – simple vector processor.
    [Show full text]
  • Massively Parallel Computing with CUDA
    Massively Parallel Computing with CUDA Antonino Tumeo Politecnico di Milano 1 GPUs have evolved to the point where many real world applications are easily implemented on them and run significantly faster than on multi-core systems. Future computing architectures will be hybrid systems with parallel-core GPUs working in tandem with multi-core CPUs. Jack Dongarra Professor, University of Tennessee; Author of “Linpack” Why Use the GPU? • The GPU has evolved into a very flexible and powerful processor: • It’s programmable using high-level languages • It supports 32-bit and 64-bit floating point IEEE-754 precision • It offers lots of GFLOPS: • GPU in every PC and workstation What is behind such an Evolution? • The GPU is specialized for compute-intensive, highly parallel computation (exactly what graphics rendering is about) • So, more transistors can be devoted to data processing rather than data caching and flow control ALU ALU Control ALU ALU Cache DRAM DRAM CPU GPU • The fast-growing video game industry exerts strong economic pressure that forces constant innovation GPUs • Each NVIDIA GPU has 240 parallel cores NVIDIA GPU • Within each core 1.4 Billion Transistors • Floating point unit • Logic unit (add, sub, mul, madd) • Move, compare unit • Branch unit • Cores managed by thread manager • Thread manager can spawn and manage 12,000+ threads per core 1 Teraflop of processing power • Zero overhead thread switching Heterogeneous Computing Domains Graphics Massive Data GPU Parallelism (Parallel Computing) Instruction CPU Level (Sequential
    [Show full text]
  • Parallel Computer Architecture
    Parallel Computer Architecture Introduction to Parallel Computing CIS 410/510 Department of Computer and Information Science Lecture 2 – Parallel Architecture Outline q Parallel architecture types q Instruction-level parallelism q Vector processing q SIMD q Shared memory ❍ Memory organization: UMA, NUMA ❍ Coherency: CC-UMA, CC-NUMA q Interconnection networks q Distributed memory q Clusters q Clusters of SMPs q Heterogeneous clusters of SMPs Introduction to Parallel Computing, University of Oregon, IPCC Lecture 2 – Parallel Architecture 2 Parallel Architecture Types • Uniprocessor • Shared Memory – Scalar processor Multiprocessor (SMP) processor – Shared memory address space – Bus-based memory system memory processor … processor – Vector processor bus processor vector memory memory – Interconnection network – Single Instruction Multiple processor … processor Data (SIMD) network processor … … memory memory Introduction to Parallel Computing, University of Oregon, IPCC Lecture 2 – Parallel Architecture 3 Parallel Architecture Types (2) • Distributed Memory • Cluster of SMPs Multiprocessor – Shared memory addressing – Message passing within SMP node between nodes – Message passing between SMP memory memory nodes … M M processor processor … … P … P P P interconnec2on network network interface interconnec2on network processor processor … P … P P … P memory memory … M M – Massively Parallel Processor (MPP) – Can also be regarded as MPP if • Many, many processors processor number is large Introduction to Parallel Computing, University of Oregon,
    [Show full text]
  • Cluster, Grid and Cloud Computing: a Detailed Comparison
    The 6th International Conference on Computer Science & Education (ICCSE 2011) August 3-5, 2011. SuperStar Virgo, Singapore ThC 3.33 Cluster, Grid and Cloud Computing: A Detailed Comparison Naidila Sadashiv S. M Dilip Kumar Dept. of Computer Science and Engineering Dept. of Computer Science and Engineering Acharya Institute of Technology University Visvesvaraya College of Engineering (UVCE) Bangalore, India Bangalore, India [email protected] [email protected] Abstract—Cloud computing is rapidly growing as an alterna- with out any prior reservation and hence eliminates over- tive to conventional computing. However, it is based on models provisioning and improves resource utilization. like cluster computing, distributed computing, utility computing To the best of our knowledge, in the literature, only a few and grid computing in general. This paper presents an end-to- end comparison between Cluster Computing, Grid Computing comparisons have been appeared in the field of computing. and Cloud Computing, along with the challenges they face. This In this paper we bring out a complete comparison of the could help in better understanding these models and to know three computing models. Rest of the paper is organized as how they differ from its related concepts, all in one go. It also follows. The cluster computing, grid computing and cloud discusses the ongoing projects and different applications that use computing models are briefly explained in Section II. Issues these computing models as a platform for execution. An insight into some of the tools which can be used in the three computing and challenges related to these computing models are listed models to design and develop applications is given.
    [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]
  • A Review of Multicore Processors with Parallel Programming
    International Journal of Engineering Technology, Management and Applied Sciences www.ijetmas.com September 2015, Volume 3, Issue 9, ISSN 2349-4476 A Review of Multicore Processors with Parallel Programming Anchal Thakur Ravinder Thakur Research Scholar, CSE Department Assistant Professor, CSE L.R Institute of Engineering and Department Technology, Solan , India. L.R Institute of Engineering and Technology, Solan, India ABSTRACT When the computers first introduced in the market, they came with single processors which limited the performance and efficiency of the computers. The classic way of overcoming the performance issue was to use bigger processors for executing the data with higher speed. Big processor did improve the performance to certain extent but these processors consumed a lot of power which started over heating the internal circuits. To achieve the efficiency and the speed simultaneously the CPU architectures developed multicore processors units in which two or more processors were used to execute the task. The multicore technology offered better response-time while running big applications, better power management and faster execution time. Multicore processors also gave developer an opportunity to parallel programming to execute the task in parallel. These days parallel programming is used to execute a task by distributing it in smaller instructions and executing them on different cores. By using parallel programming the complex tasks that are carried out in a multicore environment can be executed with higher efficiency and performance. Keywords: Multicore Processing, Multicore Utilization, Parallel Processing. INTRODUCTION From the day computers have been invented a great importance has been given to its efficiency for executing the task.
    [Show full text]
  • Cloud Computing Over Cluster, Grid Computing: a Comparative Analysis
    Journal of Grid and Distributed Computing Volume 1, Issue 1, 2011, pp-01-04 Available online at: http://www.bioinfo.in/contents.php?id=92 Cloud Computing Over Cluster, Grid Computing: a Comparative Analysis 1Indu Gandotra, 2Pawanesh Abrol, 3 Pooja Gupta, 3Rohit Uppal and 3Sandeep Singh 1Department of MCA, MIET, Jammu 2Department of Computer Science & IT, Jammu Univ, Jammu 3Department of MCA, MIET, Jammu e-mail: [email protected], [email protected], [email protected], [email protected], [email protected] Abstract—There are dozens of definitions for cloud Virtualization is a technology that enables sharing of computing and through each definition we can get the cloud resources. Cloud computing platform can become different idea about what a cloud computing exacting is? more flexible, extensible and reusable by adopting the Cloud computing is not a very new concept because it is concept of service oriented architecture [5].We will not connected to grid computing paradigm whose concept came need to unwrap the shrink wrapped software and install. into existence thirteen years ago. Cloud computing is not only related to Grid Computing but also to Utility computing The cloud is really very easier, just to install single as well as Cluster computing. Cloud computing is a software in the centralized facility and cover all the computing platform for sharing resources that include requirements of the company’s users [1]. software’s, business process, infrastructures and applications. Cloud computing also relies on the technology II. CLUSTER COMPUTING of virtualization. In this paper, we will discuss about Grid computing, Cluster computing and Cloud computing i.e.
    [Show full text]
  • “Grid Computing”
    VISHVESHWARAIAH TECHNOLOGICAL UNIVERSITY S.D.M COLLEGE OF ENGINEERING AND TECHNOLOGY A seminar report on “Grid Computing” Submitted by Nagaraj Baddi (2SD07CS402) 8th semester DEPARTMENT OF COMPUTER SCIENCE ENGINEERING 2009-10 1 VISHVESHWARAIAH TECHNOLOGICAL UNIVERSITY S.D.M COLLEGE OF ENGINEERING AND TECHNOLOGY DEPARTMENT OF COMPUTER SCIENCE ENGINEERING CERTIFICATE Certified that the seminar work entitled “Grid Computing” is a bonafide work presented by Mr. Nagaraj.M.Baddi, bearing USN 2SD07CS402 in a partial fulfillment for the award of degree of Bachelor of Engineering in Computer Science Engineering of the Vishveshwaraiah Technological University Belgaum, during the year 2009-10. The seminar report has been approved as it satisfies the academic requirements with respect to seminar work presented for the Bachelor of Engineering Degree. Staff in charge H.O.D CSE (S. L. DESHPANDE) (S. M. JOSHI) Name: Nagaraj M. Baddi USN: 2SD07CS402 2 INDEX 1. Introduction 4 2. History 5 3. How Grid Computing Works 6 4. Related technologies 8 4.1 Cluster computing 8 4.2 Peer-to-peer computing 9 4.3 Internet computing 9 5. Grid Computing Logical Levels 10 5.1 Cluster Grid 10 5.2 Campus Grid 10 5.3 Global Grid 10 6. Grid Architecture 11 6.1 Grid fabric 11 6.2 Core Grid middleware 12 6.3 User-level Grid middleware 12 6.4 Grid applications and portals. 13 7. Grid Applications 13 7.1 Distributed supercomputing 13 7.2 High-throughput computing 14 7.3 On-demand computing 14 7.4 Data-intensive computing 14 7.5 Collaborative computing 15 8. Difference: Grid Computing vs Cloud Computing 15 9.
    [Show full text]
  • Vector Vs. Scalar Processors: a Performance Comparison Using a Set of Computational Science Benchmarks
    Vector vs. Scalar Processors: A Performance Comparison Using a Set of Computational Science Benchmarks Mike Ashworth, Ian J. Bush and Martyn F. Guest, Computational Science & Engineering Department, CCLRC Daresbury Laboratory ABSTRACT: Despite a significant decline in their popularity in the last decade vector processors are still with us, and manufacturers such as Cray and NEC are bringing new products to market. We have carried out a performance comparison of three full-scale applications, the first, SBLI, a Direct Numerical Simulation code from Computational Fluid Dynamics, the second, DL_POLY, a molecular dynamics code and the third, POLCOMS, a coastal-ocean model. Comparing the performance of the Cray X1 vector system with two massively parallel (MPP) micro-processor-based systems we find three rather different results. The SBLI PCHAN benchmark performs excellently on the Cray X1 with no code modification, showing 100% vectorisation and significantly outperforming the MPP systems. The performance of DL_POLY was initially poor, but we were able to make significant improvements through a few simple optimisations. The POLCOMS code has been substantially restructured for cache-based MPP systems and now does not vectorise at all well on the Cray X1 leading to poor performance. We conclude that both vector and MPP systems can deliver high performance levels but that, depending on the algorithm, careful software design may be necessary if the same code is to achieve high performance on different architectures. KEYWORDS: vector processor, scalar processor, benchmarking, parallel computing, CFD, molecular dynamics, coastal ocean modelling All of the key computational science groups in the 1. Introduction UK made use of vector supercomputers during their halcyon days of the 1970s, 1980s and into the early 1990s Vector computers entered the scene at a very early [1]-[3].
    [Show full text]
  • Computer Systems Architecture
    CS 352H: Computer Systems Architecture Topic 14: Multicores, Multiprocessors, and Clusters University of Texas at Austin CS352H - Computer Systems Architecture Fall 2009 Don Fussell Introduction Goal: connecting multiple computers to get higher performance Multiprocessors Scalability, availability, power efficiency Job-level (process-level) parallelism High throughput for independent jobs Parallel processing program Single program run on multiple processors Multicore microprocessors Chips with multiple processors (cores) University of Texas at Austin CS352H - Computer Systems Architecture Fall 2009 Don Fussell 2 Hardware and Software Hardware Serial: e.g., Pentium 4 Parallel: e.g., quad-core Xeon e5345 Software Sequential: e.g., matrix multiplication Concurrent: e.g., operating system Sequential/concurrent software can run on serial/parallel hardware Challenge: making effective use of parallel hardware University of Texas at Austin CS352H - Computer Systems Architecture Fall 2009 Don Fussell 3 What We’ve Already Covered §2.11: Parallelism and Instructions Synchronization §3.6: Parallelism and Computer Arithmetic Associativity §4.10: Parallelism and Advanced Instruction-Level Parallelism §5.8: Parallelism and Memory Hierarchies Cache Coherence §6.9: Parallelism and I/O: Redundant Arrays of Inexpensive Disks University of Texas at Austin CS352H - Computer Systems Architecture Fall 2009 Don Fussell 4 Parallel Programming Parallel software is the problem Need to get significant performance improvement Otherwise, just use a faster uniprocessor,
    [Show full text]
  • Strategies for Managing Business Disruption Due to Grid Computing
    Strategies for managing business disruption due to Grid Computing by Vidyadhar Phalke Ph.D. Computer Science, Rutgers University, New Jersey, 1995 M.S. Computer Science, Rutgers University, New Jersey, 1992 B.Tech. Computer Science, Indian Institute of Technology, Delhi, 1989 Submitted to the MIT Sloan School of Management in Partial Fulfillment of the Requirements for the Degree of Master of Science in the Management of Technology at the Massachusetts Institute of Technology June 2003 © 2003 Vidyadhar Phalke All Rights Reserved The author hereby grants to MIT permission to reproduce and to distribute publicly paper and electronic copies of this thesis document in whole or in part Signature of Author: MIT Sloan School of Management 9 May 2003 Certified By: Starling D. Hunter III Theodore T. Miller Career Development Assistant Professor Thesis Supervisor Accepted By: David A. Weber Director, Management of Technology Program 2 Strategies for managing business disruption due to Grid Computing by Vidyadhar Phalke Submitted to the MIT Sloan School of Management on May 9 2003 in partial fulfillment of the requirements for the degree of Master of Science in the Management of Technology ABSTRACT In the technology centric businesses disruptive technologies displace incumbents time and again, sometimes to the extent that incumbents go bankrupt. In this thesis we would address the issue of what strategies are essential to prepare for and to manage disruptions for the affected businesses and industries. Specifically we will look at grid computing that is poised to disrupt (1) certain Enterprise IT departments, and (2) the software industry in the high-performance and web services space.
    [Show full text]
  • Parallel Processing! 1! CSE 30321 – Lecture 23 – Introduction to Parallel Processing! 2! Suggested Readings! •! Readings! –! H&P: Chapter 7! •! (Over Next 2 Weeks)!
    CSE 30321 – Lecture 23 – Introduction to Parallel Processing! 1! CSE 30321 – Lecture 23 – Introduction to Parallel Processing! 2! Suggested Readings! •! Readings! –! H&P: Chapter 7! •! (Over next 2 weeks)! Lecture 23" Introduction to Parallel Processing! University of Notre Dame! University of Notre Dame! CSE 30321 – Lecture 23 – Introduction to Parallel Processing! 3! CSE 30321 – Lecture 23 – Introduction to Parallel Processing! 4! Processor components! Multicore processors and programming! Processor comparison! vs.! Goal: Explain and articulate why modern microprocessors now have more than one core andCSE how software 30321 must! adapt to accommodate the now prevalent multi- core approach to computing. " Introduction and Overview! Writing more ! efficient code! The right HW for the HLL code translation! right application! University of Notre Dame! University of Notre Dame! CSE 30321 – Lecture 23 – Introduction to Parallel Processing! CSE 30321 – Lecture 23 – Introduction to Parallel Processing! 6! Pipelining and “Parallelism”! ! Load! Mem! Reg! DM! Reg! ALU ! Instruction 1! Mem! Reg! DM! Reg! ALU ! Instruction 2! Mem! Reg! DM! Reg! ALU ! Instruction 3! Mem! Reg! DM! Reg! ALU ! Instruction 4! Mem! Reg! DM! Reg! ALU Time! Instructions execution overlaps (psuedo-parallel)" but instructions in program issued sequentially." University of Notre Dame! University of Notre Dame! CSE 30321 – Lecture 23 – Introduction to Parallel Processing! CSE 30321 – Lecture 23 – Introduction to Parallel Processing! Multiprocessing (Parallel) Machines! Flynn#s
    [Show full text]