Distributed-memory multiprocessing systems (DMS), such as Intel's hypercubes, the Paragon, Thinking Machine's CM-5, and the Meiko Computing Surface, have rapidly gained user acceptance and promise to deliver the computing power required to solve the grand challenge problems of Science and Engineering. These machines are relatively inexpensive to build, and are potentially scalable to large numbers of processors. However, they are difficult to program: the non-uniformity of the memory which makes local accesses much faster than the transfer of non-local data via message-passing operations implies that the locality of algorithms must be exploited in order to achieve acceptable performance. The management of data, with the twin goals of both spreading the computational workload and minimizing the delays caused when a processor has to wait for non-local data, becomes of paramount importance. When a code is parallelized by hand, the programmer must distribute the program's work and data to the processors which will execute it. One of the common approaches to do so makes use of the regularity of most numerical computations. This is the so-called Single Program Multiple Data (SPMD) or data parallel model of computation. With this method, the data arrays in the original program are each distributed to the processors, establishing an ownership relation, and computations defining a data item are performed by the processors owning the data.

Compiling for parallelism is a longstanding topic of compiler research. This book describes the fundamental principles of compiling "regular" numerical programs for parallelism. We begin with an explanation of analyses that allow a compiler to understand the interaction of data reads and writes in different statements and loop iterations during program execution. These analyses include dependence analysis, use-def analysis and pointer analysis. Next, we describe how the results of these analyses are used to enable transformations that make loops more amenable to parallelization, and discuss transformations that expose parallelism to target shared memory multicore and vector processors. We then discuss some problems that arise when parallelizing programs for execution on distributed memory machines. Finally, we conclude with an overview of solving Diophantine equations and suggestions for further readings in the topics of this book to enable the interested reader to delve deeper into the field. Table of Contents: Introduction and overview / Dependence analysis, dependence graphs and alias analysis / Program parallelization / Transformations to modify and eliminate dependences / Transformation of iterative and recursive constructs / Compiling for distributed memory machines / Solving Diophantine equations / A guide to further reading

The automatic generation of parallel code from high level sequential description is of key importance to the wide spread use of high performance machine architectures. This text considers (in detail) the theory and practical realization of automatic mapping of algorithms generated from systems of uniform recurrence equations (do-lccps) onto fixed size architectures with defined communication primitives. Experimental results of the mapping scheme and its implementation are given.

I Unidimensional Problems.- 1 Scheduling DAGs without Communications.- 2 Scheduling DAGs with Communications.- 3 Cyclic Scheduling.- II Multidimensional Problems.- 4 Systems of Uniform Recurrence Equations.- 5 Parallelism Detection in Nested Loops.

Welcome to the proceedings of APPT 2005: the 6th International Workshop on Advanced Parallel Processing Technologies. APPT is a biennial workshop on parallel and distributed processing. Its scope covers all aspects of parallel and distributed computing technologies, including architectures, software systems and tools, algorithms, and applications. APPT originated from collaborations by researchers from China and Germany and has evolved to be an international workshop. APPT 2005 was the sixth in the series. The past ?ve workshops were held in Beijing, Koblenz, Changsha, Ilmenau, and Xiamen, respectively. The Program Committee is pleased to present the proceedings for APPT 2005. This year, APPT 2005 received over 220 submissions from researchers all over the world. All the papers were peer reviewed by two to three Program Committee members on their relevance, originality, signi?cance, technical qu- ity, and presentation. Based on the review result, 55 high-quality papers were selected to be included in the proceedings. The papers in this volume represent the forefront of research on parallel processing and related ?elds by researchers from China, Germany, USA, Korea, India, and other countries. The papers - cepted cover a wide range of exciting topics, including architectures, software, networking, and applications.

Both algorithms and the software . and hardware of automatic computers have gone through a rapid development in the past 35 years. The dominant factor in this development was the advance in computer technology. Computer parameters were systematically improved through electron tubes, transistors and integrated circuits of ever-increasing integration density, which also influenced the development of new algorithms and programming methods. Some years ago the situation in computers development was that no additional enhancement of their performance could be achieved by increasing the speed of their logical elements, due to the physical barrier of the maximum transfer speed of electric signals. Another enhancement of computer performance has been achieved by parallelism, which makes it possible by a suitable organization of n processors to obtain a perform ance increase of up to n times. Research into parallel computations has been carried out for several years in many countries and many results of fundamental importance have been obtained. Many parallel computers have been designed and their algorithmic and program ming systems built. Such computers include ILLIAC IV, DAP, STARAN, OMEN, STAR-100, TEXAS INSTRUMENTS ASC, CRAY-1, C mmp, CM*, CLIP-3, PEPE. This trend is supported by the fact that: a) many algorithms and programs are highly parallel in their structure, b) the new LSI and VLSI technologies have allowed processors to be combined into large parallel structures, c) greater and greater demands for speed and reliability of computers are made.

Multicore Application Programming is a comprehensive, practical guide to high-performance multicore programming that any experienced developer can use.

This book provides basic and practical techniques of parallel computing and related methods of numerical analysis for researchers who conduct numerical calculation and simulation. Although the techniques provided in this book are field-independent, these methods can be used in fields such as physics, chemistry, biology, earth sciences, space science, meteorology, disaster prevention, and manufacturing. In particular, those who develop software code in these areas will find this book useful. The contents are suitable for graduate students and researchers in computational science rather than novices at programming or informed experts in computer science. Starting with an introduction to the recent trends in computer architecture and parallel processing, Chapter 1 explains the basic knowledge of speedup programs with simple examples of numerical computing. Chapters 2 – 4 detail the basics of parallel programming, the message passing interface (MPI), and OpenMP and discuss hybrid parallelization techniques. Showing an actual example of adaptation, Chapter 5 gives an overview of performance tuning and communication optimizations. To deal with dense matrix calculations, Chapter 6 details the basics and practice of linear algebra calculation libraries BLAS and LAPACK, including some examples that can be easily reproduced by readers using free software. Focusing on sparse matrix calculations, Chapter 7 explains high performance algorithms for numerical linear algebra. Chapter 8 introduces the fast Fourier transform in large-scale systems from the basics. Chapter 9 explains optimization and related topics such as debug methods and version control systems. Chapter 10 discusses techniques for increasing computation accuracy as an essential topic in numerical calculation. This is the first of the two volumes that grew out of a series of lectures in the K computer project in Japan. The second volume will focus on advanced techniques and examples of applications in materials science.

This book constitutes the thoroughly refereed post-conference proceedings of the 28th International Workshop on Languages and Compilers for Parallel Computing, LCPC 2015, held in Raleigh, NC, USA, in September 2015. The 19 revised full papers were carefully reviewed and selected from 44 submissions. The papers are organized in topical sections on programming models, optimizing framework, parallelizing compiler, communication and locality, parallel applications and data structures, and correctness and reliability.

This book constitutes the thoroughly refereed post-workshop proceedings of the First and the Second International Workshop on OpenMP, IWOMP 2005 and IWOMP 2006, held in Eugene, OR, USA, and in Reims, France, in June 2005 and 2006 respectively. The first part of the book presents 16 revised full papers carefully reviewed and selected from the IWOMP 2005 program and organized in topical sections on performance tools, compiler technology, run-time environment, applications, as well as the OpenMP language and its evaluation. In the second part there are 19 papers of IWOMP 2006, fully revised and grouped thematically in sections on advanced performance tuning aspects of code development applications, and proposed extensions to OpenMP.