DOCSLIB.ORG

Cafe: Coarray Fortran Extensions for Heterogeneous Computing

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Cafe: Coarray Fortran Extensions for Heterogeneous Computing 2016 IEEE International Parallel and Distributed Processing Symposium Workshops CAFe: Coarray Fortran Extensions for Heterogeneous Computing Craig Rasmussen∗, Matthew Sottile‡, Soren Rasmussen∗ Dan Nagle† and William Dumas∗ ∗ University of Oregon, Eugene, Oregon †NCAR, Boulder, Colorado ‡Galois, Portland, Oregon Abstract—Emerging hybrid accelerator architectures are often simple tuning if the architecture is fundamentally different than proposed for inclusion as components in an exascale machine, previous ones. not only for performance reasons but also to reduce total power A number of options are available to an HPC programmer consumption. Unfortunately, programmers of these architectures face a daunting and steep learning curve that frequently requires beyond just using a serial language with MPI or OpenMP. learning a new language (e.g., OpenCL) or adopting a new New parallel languages have been under development for programming model. Furthermore, the distributed (and fre- over a decade, for example, Chapel and the PGAS languages quently multi-level) nature of the memory organization of clusters including Coarray Fortran (CAF). Options for heterogeneous of these machines provides an additional level of complexity. computing include usage of OpenACC, CUDA, CUDA For- This paper presents preliminary work examining how Fortran coarray syntax can be extended to provide simpler access to tran, and OpenCL for programming attached accelerator de- accelerator architectures. This programming model integrates the vices. Each of these choices provides the programmer with Partitioned Global Address Space (PGAS) features of Fortran abstractions over parallel systems that expose different levels with some of the more task-oriented constructs in OpenMP of detail about the specific systems to compile to, and as 4.0 and OpenACC. It also includes the potential for compiler- such, entail different degrees of effort to design for and realize based transformations targeting the Open Community Runtime (OCR) environment. We demonstrate these CoArray Fortran performance portability. extensions (CAFe) by implementing a multigrid Laplacian solver A reasonable solution given today’s language limitations is and transforming this high-level code to a mixture of standard to use MPI for distributed memory parallelism plus OpenMP coarray Fortran and OpenCL kernels. for on-chip parallelism. This hybrid approach is taken by two Index Terms —distributed memory parallelism, domain specific successful, large-scale AMR multiphysics code frameworks, language FLASH and Chombo; currently however, neither of these codes support hardware accelerators[1]. Other choices for I. INTRODUCTION expressing on-chip parallelism are OpenCL [2] and NVIDIA’s A goal of high-level programming languages should be CUDA. Unfortunately, achieving high performance using ei- to allow a programmer to develop software for a specific ther of these two languages can be a daunting challenge and hardware architecture, without actually exposing hardware there are no guarantees that either language is suitable for features. In fact, allowing low-level hardware details to be future hardware generations. controlled by the programmer is actually a barrier to software Another potential option is to abandon the structure of performance portability: the explicit use of low-level hardware message passing within a procedural programming language details leads to code becoming excessively specialized to a altogether. For example, the Open Community Runtime (OCR) given target system. The path forward for high-level program- project[3] provides an asynchronous task-based runtime de- ming languages should be to allow a programmer to provide signed for machines with high-core count[4], such as can more semantic information about their intent, thus allowing the be expected in any exascale hardware design. The project’s compiler more freedom to choose how to instantiate this intent goal is to create a framework and reference implementation in a specific implementation in order to retarget applications to help developers explore programming methods to improve to new architectures. the power efficiency, programmability, and reliability of HPC Performance portability (or, even simply portability) has applications while maintaining application performance. been a perpetual problem in computing, especially within the It cannot be emphasized too strongly that using the OpenCL high performance computing community in which relatively programming language and libraries (or programming for the new and immature emerging technologies are constantly being OCR environment) presents a high hurdle for the application tried and tested. As processor architectures hit limitations due developer creating real scientific applications. A goal of this to physical and manufacturing constraints, we are seeing a paper is to examine ways in which high-level syntax based trend towards more diversity in processor designs than were on Fortran coarrays can be utilized by compiler-based tools to experienced for the bulk of the 1990s and 2000s. Porta- target programming environments like OpenCL, thus freeing bility, especially performance portability, is becoming more the developer from the task of frequently porting the applica- challenging as moving to a new CPU may entail more than tion to new complex environments. The hope is that high-level /16978-1-5090-3682-0/16 $31.00 © 2016 IEEE $31.00 © 2016 IEEE 357 DOI 10.1109/IPDPSW.2016.140 syntax can hide the complexities of the underlying hardware II. COARRAY FORTRAN EXTENSIONS and software by delegating some optimization tasks to the Topics highlighted in this section are: 1. Distributed memory compiler instead of the programmer. array allocation; 2. Explicit memory placement; 3. Remote With the addition of coarrays in the 2008 standard, For- memory transfer; and 4. Remote execution and synchroniza- tran is now a parallel language. Originally called Coarray tion. This description is within the context of extensions Fortran (CAF) by its authors[5], parallelism in Fortran is to Fortran; as shorthand, these extensions are referred to similar to MPI in that all cooperating CAF processes (called as CAFe, for Coarray Fortran extensions. Please note that images) execute a Single Program but with different (Multiple) throughout this paper, text appearing in all caps (such as Data (SPMD). Key concepts in CAF are parallel execution, GET_SUBIMAGE) indicates new syntax or functionality ex- distributed memory allocation, and remote memory transfer. tended by CAFe. Parallel execution and memory allocation are straightforward A. Subimages as each program is replicated NP times with each program image allocating its own memory locally. Special syntax is We first introduce the important new concept of a CAFe provided to indicate which arrays are visible between dis- subimage. Fortran images are a collection of distributed mem- tributed memory processes and remote memory transfer is ory processes that all run the same program. CAFe extends indicated with new array syntax using square brackets, for the concept of a Fortran image by allowing images to be example, hierarchical; by this we mean that an image may host a subimage (or several subimages). It is important to note that U(1) = U(1)[3] a subimage is not visible to Fortran images other than its hosting image. Subimages also execute differently than normal copies memory from the first element of U residing on image images and may even execute on different non-homogeneous number 3 into the corresponding element of the local array hardware, e.g., an attached accelerator device. Subimages are U. The fundamental programming model is similar to that task based while images all execute a Single Program but with provided by message passing systems. different (Multiple) Data (SPMD). A task can be given to a subimage, but execution on the subimage terminates once Since each CAF process executes the same program, there the task is finished. Memory on a subimage is permanent, are no good avenues for exploiting task-based parallelism. however, and must be explicitly allocated and deallocated. Neither is there an adequate mechanism for dynamic coarray A programmer requests a subimage by executing the new memory allocation, nor a mechanism for the remote execu- CAFe function, tion of functions. In CAF, like MPI, the programmer often combines distributed memory processes running the same device = GET_SUBIMAGE(device_id) program image with a shared memory threaded programming where the integer argument represents an attached hardware model using OpenMP (or OpenACC). This model breaks down device (or perhaps a separate process or a team of threads; when dealing with architectures that are not isomorphic to it is the compilers responsibility to determine precisely what the model for which MPI and relatives were designed – a a subimage means in terms of the software and hardware distributed collection of homogenous processing elements in environment under its control). If the function fails (e.g., the which each processing element supports a small number of requested device is unavailable) it returns the local image num- parallel threads of execution. For example, an HPC system that ber of the process that is executing the current program, ob- makes use of traditional multicore CPUs as well as GPGPU tained by a call to the Fortran 2008 function this_image(). accelerators breaks this model and, to effectively program it, Returning the current image if the GET_SUBIMAGE call fails,
Load more

Recommended publications
  • Evaluation of the Coarray Fortran Programming Model on the Example of a Lattice Boltzmann Code
    Evaluation of the Coarray Fortran Programming Model on the Example of a Lattice Boltzmann Code Klaus Sembritzki Georg Hager Bettina Krammer Friedrich-Alexander University Friedrich-Alexander University Université de Versailles Erlangen-Nuremberg Erlangen-Nuremberg St-Quentin-en-Yvelines Erlangen Regional Computing Erlangen Regional Computing Exascale Computing Center (RRZE) Center (RRZE) Research (ECR) Martensstrasse 1 Martensstrasse 1 45 Avenue des Etats-Unis 91058 Erlangen, Germany 91058 Erlangen, Germany 78000 Versailles, France [email protected] [email protected] [email protected] Jan Treibig Gerhard Wellein Friedrich-Alexander University Friedrich-Alexander University Erlangen-Nuremberg Erlangen-Nuremberg Erlangen Regional Computing Erlangen Regional Computing Center (RRZE) Center (RRZE) Martensstrasse 1 Martensstrasse 1 91058 Erlangen, Germany 91058 Erlangen, Germany [email protected] [email protected] ABSTRACT easier to learn. The Lattice Boltzmann method is an explicit time-stepping sche- With the Fortran 2008 standard, coarrays became a native Fort- me for the numerical simulation of fluids. In recent years it has ran feature [2]. The “hybrid” nature of modern massively parallel gained popularity since it is straightforward to parallelize and well systems, which comprise shared-memory nodes built from multico- suited for modeling complex boundary conditions and multiphase re processor chips, can be addressed by the compiler and runtime flows. Starting from an MPI/OpenMP-parallel 3D prototype im- system by using shared memory for intra-node access to codimen- plementation of the algorithm in Fortran90, we construct several sions. Inter-node access may either by conducted using an existing coarray-based versions and compare their performance and requi- messaging layer such as MPI or SHMEM, or carried out natively red programming effort to the original code, demonstrating the per- using the primitives provided by the network infrastructure.
    [Show full text]
  • A Comparison of Coarray Fortran, Unified Parallel C and T
    Technical Report RAL-TR-2005-015 8 December 2005 Novel Parallel Languages for Scientific Computing ± a comparison of Co-Array Fortran, Unified Parallel C and Titanium J.V.Ashby Computational Science and Engineering Department CCLRC Rutherford Appleton Laboratory [email protected] Abstract With the increased use of parallel computers programming languages need to evolve to support parallelism. Three languages built on existing popular languages, are considered: Co-array Fortran, Universal Parallel C and Titanium. The features of each language which support paralleism are explained, a simple example is programmed and the current status and availability of the languages is reviewed. Finally we also review work on the performance of algorithms programmed in these languages. In most cases they perform favourably with the same algorithms programmed using message passing due to the much lower latency in their communication model. 1 1. Introduction As the use of parallel computers becomes increasingly prevalent there is an increased need for languages and programming methodologies which support parallelism. Ideally one would like an automatic parallelising compiler which could analyse a program written sequentially, identify opportunities for parallelisation and implement them without user knowledge or intervention. Attempts in the past to produce such compilers have largely failed (though automatic vectorising compilers, which may be regarded as one specific form of parallelism, have been successful, and the techniques are now a standard part of most optimising compilers for pipelined processors). As a result, interest in tools to support parallel programming has largely concentrated on two main areas: directive driven additions to existing languages and data-passing libraries.
    [Show full text]
  • Coarray Fortran Runtime Implementation in Openuh
    COARRAY FORTRAN RUNTIME IMPLEMENTATION IN OPENUH A Thesis Presented to the Faculty of the Department of Computer Science University of Houston In Partial Fulfillment of the Requirements for the Degree Master of Science By Debjyoti Majumder December 2011 COARRAY FORTRAN RUNTIME IMPLEMENTATION IN OPENUH Debjyoti Majumder APPROVED: Dr. Barbara Chapman, Chairman Dept. of Computer Science Dr. Jaspal Subhlok Dept. of Computer Science Dr. Terrence Liao TOTAL E&P Research & Technology USA, LLC Dean, College of Natural Sciences and Mathematics ii Acknowledgements I would like to express my gratitude to Dr. Barbara Chapman for providing financial support and for creating an excellent environment for research. I would like to thank my mentor Deepak Eachempati. Without his guidance, this work would not have been possible. I sincerely thank TOTAL S.A. for funding this project and Dr. Terrence Liao for his valuable inputs and help in performing experiments on TOTAL's hardware. I also thank the interns at TOTAL, Asma Farjallah, and France Boillod-Cerneux for the applications that I used for performance evaluation. I thank Hyoungjoon Jun for his earlier work on the CAF runtime. I thank Tony Curtis for his suggestions and excellent work on maintaining the infrastructure on which I ran my experiments. I thank Dr.Edgar Gabriel for answering all my questions and his help with using the shark cluster. I also thank the Berkeley Lab and PNNL's ARMCI developers for their guidance in using GASNet and ARMCI. Lastly, I would like to thank the Department of Computer Science for giving me the opportunity to be in this great country of USA and all my lab-mates and friends.
    [Show full text]
  • Parallel Computing
    Parallel Computing Benson Muite [email protected] http://math.ut.ee/˜benson https://courses.cs.ut.ee/2014/paralleel/fall/Main/HomePage 6 October 2014 Parallel Programming Models and Parallel Programming Implementations Parallel Programming Models • Shared Memory • Distributed Memory • Simple computation and communication cost models Parallel Programming Implementations • OpenMP • MPI • OpenCL – C programming like language that aims to be manufacturer independent and allow for wide range of multithreaded devices • CUDA (Compute Unified Device Architecture) – Used for Nvidia GPUs, based on C/C++, makes numerical computations easier • Pthreads – older threaded programming interface, some legacy codes still around, lower level than OpenMP but not as easy to use • Universal Parallel C (UPC) – New generation parallel programming language, distributed array support, consider trying it • Coarray Fortran (CAF) – New generation parallel programming language, distributed array support, consider trying it • Julia (http://julialang.org/) – New serial and parallel programming language, consider trying it • Parallel Virtual Machine (PVM) – Older, has been replaced by MPI • OpenACC – brief mention • Not covered: Java threads, C++11 threads, Python Threads, High Performance Fortran (HPF) Parallel Programming Models: Flynn’s Taxonomy • (Single Instruction Single Data) – Serial • Single Program Multiple Data • Multiple Program Multiple Data • Multiple Instruction Single Data • Master Worker • Server Client Parallel Programming Models • loop parallelizm;
    [Show full text]
  • CS 470 Spring 2018
    CS 470 Spring 2018 Mike Lam, Professor Parallel Languages Graphics and content taken from the following: http://dl.acm.org/citation.cfm?id=2716320 http://chapel.cray.com/papers/BriefOverviewChapel.pdf http://arxiv.org/pdf/1411.1607v4.pdf https://en.wikipedia.org/wiki/Cilk Parallel languages ● Writing efficient parallel code is hard ● We've covered two generic paradigms ... – Shared-memory – Distributed message-passing ● … and three specific technologies (but all in C!) – Pthreads – OpenMP – MPI ● Can we make parallelism easier by changing our language? – Similarly: Can we improve programmer productivity? Productivity Output ● Productivity= Economic definition: Input ● What does this mean for parallel programming? – How do you measure input? – How do you measure output? Productivity Output ● Productivity= Economic definition: Input ● What does this mean for parallel programming? – How do you measure input? ● Bad idea: size of programming team ● "The Mythical Man Month" by Frederick Brooks – How do you measure output? ● Bad idea: lines of code Productivity vs. Performance ● General idea: Produce better code faster – Better can mean a variety of things: speed, robustness, etc. – Faster generally means time/personnel investment ● Problem: productivity often trades off with performance – E.g., Python vs. C or Matlab vs. Fortran – E.g., garbage collection or thread management Why? Complexity ● Core issue: handling complexity ● Tradeoff: developer effort vs. system effort – Hiding complexity from the developer increases the complexity of the
    [Show full text]
  • Coarray Fortran: Past, Present, and Future
    Coarray Fortran: Past, Present, and Future John Mellor-Crummey Department of Computer Science Rice University [email protected] CScADS 2009 Workshop on Leadership Computing 1 Rice CAF Team • Staff – Bill Scherer – Laksono Adhianto – Guohua Jin – Fengmei Zhao • Alumni – Yuri Dotsenko – Cristian Coarfa 2 Outline • Partitioned Global Address Space (PGAS) languages • Coarray Fortran, circa 1998 (CAF98) • Assessment of CAF98 • A look at the emerging Fortran 2008 standard • A new vision for Coarray Fortran 3 Partitioned Global Address Space Languages • Global address space – one-sided communication (GET/PUT) simpler than msg passing • Programmer has control over performance-critical factors – data distribution and locality control lacking in OpenMP – computation partitioning HPF & OpenMP compilers – communication placement must get this right • Data movement and synchronization as language primitives – amenable to compiler-based communication optimization 4 Outline • Partitioned Global Address Space (PGAS) languages • Coarray Fortran, circa 1998 (CAF98) • motivation & philosophy • execution model • co-arrays and remote data accesses • allocatable and pointer co-array components • processor spaces: co-dimensions and image indexing • synchronization • other features and intrinsic functions • Assessment of CAF98 • A look at the emerging Fortran 2008 standard • A new vision for Coarray Fortran 5 Co-array Fortran Design Philosophy • What is the smallest change required to make Fortran 90 an effective parallel language? • How can this change be expressed so that it is intuitive and natural for Fortran programmers? • How can it be expressed so that existing compiler technology can implement it easily and efficiently? 6 Co-Array Fortran Overview • Explicitly-parallel extension of Fortran 95 – defined by Numrich & Reid • SPMD parallel programming model • Global address space with one-sided communication • Two-level memory model for locality management – local vs.
    [Show full text]
  • The Cascade High Productivity Programming Language
    The Cascade High Productivity Programming Language Hans P. Zima University of Vienna, Austria and JPL, California Institute of Technology, Pasadena, CA ECMWF Workshop on the Use of High Performance Computing in Meteorology Reading, UK, October 25, 2004 Contents 1 IntroductionIntroduction 2 ProgrammingProgramming ModelsModels forfor HighHigh ProductivityProductivity ComputingComputing SystemsSystems 3 CascadeCascade andand thethe ChapelChapel LanguageLanguage DesignDesign 4 ProgrammingProgramming EnvironmentsEnvironments 5 ConclusionConclusion Abstraction in Programming Programming models and languages bridge the gap between ““reality”reality” and hardware – at differentdifferent levels ofof abstractionabstraction - e.g., assembly languages general-purpose procedural languages functional languages very high-level domain-specific languages Abstraction implies loss of information – gain in simplicity, clarity, verifiability, portability versus potential performance degradation The Emergence of High-Level Sequential Languages The designers of the very first highhigh level programming langulanguageage were aware that their successsuccess dependeddepended onon acceptableacceptable performance of the generated target programs: John Backus (1957): “… It was our belief that if FORTRAN … were to translate any reasonable scientific source program into an object program only half as fast as its hand-coded counterpart, then acceptance of our system would be in serious danger …” High-level algorithmic languages became generally accepted standards for
    [Show full text]
  • Locally-Oriented Programming: a Simple Programming Model for Stencil-Based Computations on Multi-Level Distributed Memory Architectures
    Locally-Oriented Programming: A Simple Programming Model for Stencil-Based Computations on Multi-Level Distributed Memory Architectures Craig Rasmussen, Matthew Sottile, Daniel Nagle, and Soren Rasmussen University of Oregon, Eugene, Oregon, USA [email protected] Abstract. Emerging hybrid accelerator architectures for high perfor- mance computing are often suited for the use of a data-parallel pro- gramming model. Unfortunately, programmers of these architectures face a steep learning curve that frequently requires learning a new language (e.g., OpenCL). Furthermore, the distributed (and frequently multi-level) nature of the memory organization of clusters of these machines provides an additional level of complexity. This paper presents preliminary work examining how programming with a local orientation can be employed to provide simpler access to accelerator architectures. A locally-oriented programming model is especially useful for the solution of algorithms requiring the application of a stencil or convolution kernel. In this pro- gramming model, a programmer codes the algorithm by modifying only a single array element (called the local element), but has read-only ac- cess to a small sub-array surrounding the local element. We demonstrate how a locally-oriented programming model can be adopted as a language extension using source-to-source program transformations. Keywords: domain specific language, stencil compiler, distributed mem- ory parallelism 1 Introduction Historically it has been hard to create parallel programs. The responsibility (and difficulty) of creating correct parallel programs can be viewed as being spread between the programmer and the compiler. Ideally we would like to have paral- lel languages that make it easy for the programmer to express correct parallel arXiv:1502.03504v1 [cs.PL] 12 Feb 2015 programs — and conversely it should be difficult to express incorrect parallel programs.
    [Show full text]
  • Coarray Fortran 2.0
    Coarray Fortran 2.0 John Mellor-Crummey, Karthik Murthy, Dung Nguyen, Sriraj Paul, Scott Warren, Chaoran Yang Department of Computer Science Rice University DEGAS Retreat 4 June 2013 Coarray Fortran (CAF) P0 P1 P2 P3 A(1:50)[0] A(1:50)[1] A(1:50)[2] A(1:50)[3] Global view B(1:40) B(1:40) B(1:40) B(1:40) Local view • Global address space SPMD parallel programming model —one-sided communication • Simple, two-level memory model for locality management —local vs. remote memory • Programmer has control over performance critical decisions —data partitioning —data movement —synchronization • Adopted in Fortran 2008 standard 2 Coarray Fortran 2.0 Goals • Exploit multicore processors • Enable development of portable high-performance programs • Interoperate with legacy models such as MPI • Facilitate construction of sophisticated parallel applications and parallel libraries • Support irregular and adaptive applications • Hide communication latency • Colocate computation with remote data • Scale to exascale 3 Coarray Fortran 2.0 (CAF 2.0) • Teams: process subsets, like MPI communicators CAF 2.0 Features —formation using team_split —collective communication (two-sided) Fortran —barrier synchronization 2008 • Coarrays: shared data allocated across processor subsets —declaration: double precision :: a(:,:)[*] —dynamic allocation: allocate(a(n,m)[@row_team]) —access: x(:,n+1) = x(:,0)[mod(team_rank()+1, team_size())] • Latency tolerance —hide: asynchronous copy, asynchronous collectives —avoid: function shipping • Synchronization —event variables: point-to-point
    [Show full text]
  • IBM Power Systems Compiler Strategy
    Software Group Compilation Technology IBM Power Systems Compiler Strategy Roch Archambault IBM Toronto Laboratory [email protected] ASCAC meeting | August 11, 2009 @ 2009 IBM Corporation Software Group Compilation Technology IBM Rational Disclaimer © Copyright IBM Corporation 2008. All rights reserved. The information contained in these materials is provided for informational purposes only, and is provided AS IS without warranty of any kind, express or implied. IBM shall not be responsible for any damages arising out of the use of, or otherwise related to, these materials. Nothing contained in these materials is intended to, nor shall have the effect of, creating any warranties or representations from IBM or its suppliers or licensors, or altering the terms and conditions of the applicable license agreement governing the use of IBM software. References in these materials to IBM products, programs, or services do not imply that they will be available in all countries in which IBM operates. Product release dates and/or capabilities referenced in these materials may change at any time at IBM’s sole discretion based on market opportunities or other factors, and are not intended to be a commitment to future product or feature availability in any way. IBM, the IBM logo, Rational, the Rational logo, Telelogic, the Telelogic logo, and other IBM products and services are trademarks of the International Business Machines Corporation, in the United States, other countries or both. Other company, product, or service names may be trademarks or service
    [Show full text]
  • Selecting the Right Parallel Programming Model
    Selecting the Right Parallel Programming Model perspective and commitment James Reinders, Intel Corp. 1 © Intel 2012, All Rights Reserved Selecting the Right Parallel Programming Model TACC‐Intel Highly Parallel Computing Symposium April 10th‐11th, 2012 Austin, TX 8:30‐10:00am, Tuesday April 10, 2012 Invited Talks: • "Selecting the right Parallel Programming Model: Intel's perspective and commitments," James Reinders • "Many core processors at Intel: lessons learned and remaining challenges," Tim Mattson James Reinders, Director, Software Evangelist, Intel Corporation James Reinders is a senior engineer who joined Intel Corporation in 1989 and has contributed to projects including systolic arrays systems WARP and iWarp, and the world's first TeraFLOPS supercomputer (ASCI Red), and the world's first TeraFLOPS microprocessor (Knights Corner), as well as compilers and architecture work for multiple Intel processors and parallel systems. James has been a driver behind the development of Intel as a major provider of software development products, and serves as their chief software evangelist. James is the author of a book on VTune (Intel Press) and Threading Building Blocks (O'Reilly Media) as well as a co‐author of Structured Parallel Programming (Morgan Kauffman) due out in June 2012. James is currently involved in multiple efforts at Intel to bring parallel programming models to the industry including for the Intel MIC architecture. 2 © Intel 2012, All Rights Reserved No Widely Used Programming Language was designed as a Parallel Programming
    [Show full text]
  • Evaluation of the Coarray Fortran Programming Model on the Example of a Lattice Boltzmann Code
    Evaluation of the Coarray Fortran Programming Model on the Example of a Lattice Boltzmann Code Masters Thesis in Computational Engineering submitted by Klaus Sembritzki born 16.05.1986 in Erlangen Written at the Friedrich-Alexander-University of Erlangen-Nurnberg¨ Advisor: Prof. Dr. Gerhard Wellein, Dr.-Ing. Jan Treibig Started: 1.8.2011 Finished: 31.1.2012 ii iii Ich versichere, dass ich die Arbeit ohne fremde Hilfe und ohne Benutzung anderer als der angegebenen Quellen angefertigt habe und dass die Arbeit in gleicher oder ahnlicher¨ Form noch keiner anderen Prufungsbeh¨ orde¨ vorgelegen hat und von dieser als Teil einer Prufungsleistung¨ angenommen wurde. Alle Ausfuhrungen,¨ die wortlich¨ oder sinngemaߨ ubernommen¨ wurden, sind als solche gekennzeichnet. Erlangen, den 31. Januar 2012 iv v Ubersicht¨ Die Suche nach Programmierparadigmen, welche den Quelltext paralleler Programme robuster, leser- licher und wartbarer gestalten, hat zur Entwicklung sogenannter PGAS (Partitioned Global Address Space) Sprachen gefuhrt.¨ Typische Beispiele sind UPC (Unified Parallel C) und CAF (Coarray Fortran). Bisher haben diese Sprachen jedoch noch nicht den Weg aus dem akademischen Umfeld in Produktionsumge- bungen gefunden. Mit dem Fortran 2008 Standard sind Coarrays ein nativer Bestandteil von Fortran geworden. Die Popularitat¨ von Fortran im wissenschaftlichen Rechnen und die große Fortran Code Basis lasst¨ darauf hoffen, dass fur¨ die Parallelisierung von bereits existierendem Fortran Code und fur¨ das Schreiben von neuem Quelltext von Entwicklern die native Sprachvariante einer MPI/OpenMP Parallelisierung vorgezo- gen wird. Dies ist jedoch nur moglich,¨ wenn sowohl der Sprachstandard fur¨ das wissenschaftliche An- wendungsfeld ausreichend ist als auch die zur Verfugung¨ stehenden Compiler und Entwicklungsumgebun- gen den Anforderungen des wissenschaftlichen Rechnens nach Performance und Zuverlassigkeit¨ gerecht werden.
    [Show full text]