High Performance Fortran Compiler Research Articles

Analysis of two-level data mapping in an HPF compiler for distributed-memory machines

Application programs in scientific and technological fields have grown increasingly large and complex. Thus, it is becoming more difficult to parallelize these programs by hand using message-passing libraries. To reduce this difficulty, we are researching the compilation technology for an industry-standard data-parallel language: High Performance Fortran (HPF). This paper proposes new data-mapping analysis applicable to HPF data-remapping directives more general than those previous research has coped with. This analysis allows programmers to develop highly productive HPF programs. We implemented the analysis in our HPF compiler and evaluated using some benchmark programs. As a result, both programs with general HPF directives and with restricted ones were transformed to similar codes and these codes ran in similar processing speed on a Hitachi SR2201 supercomputer.

Domain interaction patterns to coordinate HPF tasks

This paper describes domain interaction patterns, a pattern-based, high level coordination language, which provides a new way of integrating task and data parallelism. Coordination patterns are used to express task parallelism among a collection of data parallel High Performance Fortran (HPF) tasks. Patterns specify the interaction among domains involved in the application along with the processor and data layouts. The use of domains, i.e. regions together with some interaction information, improves pattern reusability. Data distribution belonging to the different HPF tasks is known at the coordination level. This is the key for both computational code reutilization and an efficient implementation of the communication among tasks. Besides that, our system implementation requires no change to the runtime system support of the HPF compiler used. In addition, a set of different implementation templates is provided in order to ease the programmer task. The suitability, expressiveness and efficiency of the language are shown by means of some examples.

Techniques for compiling and implementing all NAS parallel benchmarks in HPF

AbstractThe NAS parallel benchmarks (NPB) are a well‐known benchmark set for high‐performance machines. Much effort has been made to implement them in High‐Performance Fortran (HPF). In previous attempts, however, the HPF versions did not include the complete set of benchmarks, and the performance was not always good. In this study, we implement all eight benchmarks of the NPB in HPF, and parallelize them using an HPF compiler that we have developed. This report describes the implementation techniques and compiler features necessary to achieve good performance. We evaluate the HPF version on the Hitachi SR2201, a distributed‐memory parallel machine. With 16 processors, the execution time of the HPF version is within a factor of 1.5 of the hand‐parallelized version of the NPB 2.3 beta. Copyright © 2002 John Wiley & Sons, Ltd.

Efficient parallel programming on scalable shared memory systems with High Performance Fortran

AbstractOpenMP offers a high‐level interface for parallel programming on scalable shared memory (SMP) architectures. It provides the user with simple work‐sharing directives while it relies on the compiler to generate parallel programs based on thread parallelism. However, the lack of language features for exploiting data locality often results in poor performance since the non‐uniform memory access times on scalable SMP machines cannot be neglected. High Performance Fortran (HPF), the de‐facto standard for data parallel programming, offers a rich set of data distribution directives in order to exploit data locality, but it has been mainly targeted towards distributed memory machines.In this paper we describe an optimized execution model for HPF programs on SMP machines that avails itself with mechanisms provided by OpenMP for work sharing and thread parallelism, while exploiting data locality based on user‐specified distribution directives. Data locality does not only ensure that most memory accesses are close to the executing threads and are therefore faster, but it also minimizes synchronization overheads, especially in the case of unstructured reductions. The proposed shared memory execution model for HPF relies on a small set of language extensions, which resemble the OpenMP work‐sharing features. These extensions, together with an optimized shared memory parallelization and execution model, have been implemented in the ADAPTOR HPF compilation system and experimental results verify the efficiency of the chosen approach. Copyright © 2002 John Wiley & Sons, Ltd.

Compatibility comparison and performance evaluation for Japanese HPF compilers using scientific applications

AbstractThe lack of compatibility of High‐Performance Fortran (HPF) between vender implementations has been disheartening scientific application users so as to hinder the development of portable programs. Thus parallel computing is still unpopular in the computational science community, even though parallel programming is common to the computer science community. As users would like to run the same source code on parallel machines with different architectures as fast as possible, we have investigated the compatibility of source codes for Japanese HPF compilers (NEC, Fujitsu and Hitachi) with two real‐world applications: a 3D fluid code and a 2D particle code. We have found that the source‐level compatibility between Japanese HPF compilers is almost preserved, but more effort will be needed to sustain complete compatibility. We have also evaluated parallel performance and found that HPF can achieve good performance for the 3D fluid code with almost the same source code. For the 2D particle code, good results have also been obtained with a small number of processors, but some changes in the original source code and the addition of interface blocks is required. Copyright © 2002 John Wiley & Sons, Ltd.

SPiDER—An advanced symbolic debugger for Fortran 90/HPF programs

AbstractDebuggers play an important role in developing parallel applications. They are used to control the state of many processes, to present distributed information in a concise and clear way, to observe the execution behavior, and to detect and locate programming errors. More sophisticated debugging systems also try to improve understanding of global execution behavior and intricate details of a program. In this paper we describe the design and implementation of SPiDER, which is an interactive source‐level debugging system for both regular and irregular High‐Performance Fortran (HPF) programs. SPiDER combines a base debugging system for message‐passing programs with a high‐level debugger that interfaces with an HPF compiler. SPiDER, in addition to conventional debugging functionality, allows a single process of a parallel program to be expected or the entire program to be examined from a global point of view. A sophisticated visualization system has been developed and included in SPiDER to visualize data distributions, data‐to‐processor mapping relationships, and array values. SPiDER enables a programmer to dynamically change data distributions as well as array values. For arrays whose distribution can change during program execution, an animated replay displays the distribution sequence together with the associated source code location. Array values can be stored at individual execution points and compared against each other to examine execution behavior (e.g. convergence behavior of a numerical algorithm). Finally, SPiDER also offers limited support to evaluate the performance of parallel programs through a graphical load diagram. SPiDER has been fully implemented and is currently being used for the development of various real‐world applications. Several experiments are presented that demonstrate the usefulness of SPiDER. Copyright © 2002 John Wiley & Sons, Ltd.

HPF Library and Compiler Support for Halos in Data Parallel Irregular Computations

On distributed memory architectures data parallel compilers emulate the global address space by distributing the data onto the processors according to the mapping directives of the user and by generating automatically explicit inter-processor communication. A shadow is additionally allocated local memory to keep on one processor also non-local values of the data that is accessed or defined by this processor. While shadow edges are already well studied for structured grids, this paper focuses on its use for applications with unstructured grids where updates on the shadow edges involve unstructured communication with complex communication schedules. The use of shadow edges is considered for High Performance Fortran (HPF) as the de facto standard language for writing data parallel programs in Fortran. A library with a HPF binding provides the explicit control of unstructured shadows and their communication schedules, also called halos. This halo library allows writing HPF programs with a performance close to hand-coded message-passing versions but where the user is freed of the burden to calculate shadow sizes and communication schedules and to do the exchanging of data with explicit message passing commands. In certain situations, the HPF compiler can create and use halos automatically. This paper shows the advantages and also the limits of this approach. The halo library and an automatic support of halos have been implemented within the ADAPTOR HPF compilation system. The performance results verify the effectiveness of the chosen approach.

A generic interface for parallel access to large data sets from HPF applications

A critical performance issue for a number of scientific and engineering applications is the efficient transfer of data to secondary storage. Languages such as High Performance Fortran (HPF) have been introduced to allow programming distributed-memory systems at a relatively high level of abstraction. However, the present version of HPF does not provide appropriate constructs for controlling the parallel I/O capabilities of these systems. In this paper, constructs to specify parallel I/O operations on multidimensional arrays in the context of HPF are proposed. The paper also presents implementation concepts that are based on the HPF compiler VFC and the parallel I/O run-time system Panda. Experimental performance results are discussed in the context of financial management and traffic simulation applications.

HPF LIBRARY AND COMPILER SUPPORT FOR HALOS IN DATA PARALLEL IRREGULAR COMPUTATIONS

On distributed memory architectures data parallel compilers emulate the global address space by distributing the data onto the processors according to the mapping directives of the user and by generating automatically explicit inter-processor communication. A shadow is additionally allocated local memory to keep on one processor also non-local values of the data that is accessed or defined by this processor. While shadow edges are already well studied for structured grids, this paper focuses on its use for applications with unstructured grids where updates on the shadow edges involve unstructured communication with complex communication schedules. The use of shadow edges is considered for High Performance Fortran (HPF) as the de facto standard language for writing data parallel programs in Fortran. A library with a HPF binding provides the explicit control of unstructured shadows and their communication schedules, also called halos. This halo library allows writing HPF programs with a performance close to hand-coded message-passing versions but where the user is freed of the burden to calculate shadow sizes and communication schedules and to do the exchanging of data with explicit message passing commands. In certain situations, the HPF compiler can create and use halos automatically. This paper shows the advantages and also the limits of this approach. The halo library and an automatic support of halos have been implemented within the ADAPTOR HPF compilation system. The performance results verify the effectiveness of the chosen approach.

HPFBench

The high performance Fortran (HPF) benchmark suite HPFBench is designed for evaluating the HPF language and compilers on scalable architectures. The functionality of the benchmarks covers scientific software library functions and application kernels that reflect the computational structure and communication patterns in fluid dynamic simulations, fundamental physics, and molecular studies in chemistry and biology. The benchmarks are characterized in terms of FLOP count, memory usage, communication pattern, local memory accesses, array allocation mechanism, as well as operation and communication counts per iteration. The benchmarks output performance evaluation metrics in the form of elapsed times, FLOP rates, and communication time breakdowns. We also provide a benchmark guide to aid the choice of subsets of the benchmarks for evaluating particular aspects of an HPF compiler. Furthermore, we report an evaluation of an industry-leading HPF compiler from the Portland Group Inc. using the HPFBench benchmarks on the distributed-memory IBM SP2

Compiling High Performance Fortran for distributed-memory architectures

High Performance Fortran (HPF) is a data-parallel language that provides a high-level interface for programming scientific applications, while delegating to the compiler the task of generating explicitly parallel message-passing programs. This paper provides an overview of HPF compilation and runtime technology for distributed-memory architectures, and deals with a number of topics in some detail. In particular, we discuss distribution and alignment processing, the basic compilation scheme and methods for the optimization of regular computations. A separate section is devoted to the transformation and optimization of independent loops with irregular data accesses. The paper concludes with a discussion of research issues and outlines potential future development paths of the language.

High Performance Fortran for practical scientific algorithms: An up-to-date evaluation

A suite of High Performance Fortran (HPF) coding examples of practical scientific algorithms are examined in detail, with the idea that on these simple but non-trivial examples, we can fairly well understand issues related to different data distributions, different parallel constructs, and different programming styles (static versus dynamic allocations). Coding examples include 2D stencils solution of PDEs, N-body problem, LU factorization, several vector/matrix library routines, 2D and 3D array redistribution. Performances of HPF codes are compared to hand-written Fortran codes with message passing libraries. From 1997 to 1998, HPF compilers are improved significantly such that HPF codes perform as well as Fortran+MPI codes for all the examples investigated here. However, many important peculiarities of HPF coding still exist.

VFC: The Vienna Fortran Compiler

High Performance Fortran (HPF) offers an attractive high‐level language interface for programming scalable parallel architectures providing the user with directives for the specification of data distribution and delegating to the compiler the task of generating an explicitly parallel program. Available HPF compilers can handle regular codes quite efficiently, but dramatic performance losses may be encountered for applications which are based on highly irregular, dynamically changing data structures and access patterns. In this paper we introduce the Vienna Fortran Compiler (VFC), a new source‐to‐source parallelization system for HPF+, an optimized version of HPF, which addresses the requirements of irregular applications. In addition to extended data distribution and work distribution mechanisms, HPF+ provides the user with language features for specifying certain information that decisively influence a program’s performance. This comprises data locality assertions, non‐local access specifications and the possibility of reusing runtime‐generated communication schedules of irregular loops. Performance measurements of kernels from advanced applications demonstrate that with a high‐level data parallel language such as HPF+ a performance close to hand‐written message‐passing programs can be achieved even for highly irregular codes.

DAQV: Distributed array query and visualization framework

This paper describes the design and implementation of the distributed array query and visualization (DAQV) framework, a project sponsored by the Parallel Tools Consortium. DAQV's reference implementation targets High Performance Fortran (HPF) and leverages the HPF language, compiler, and runtime system to address the general problem of providing high-level access to distributed data structures. DAQV supports a framework in which visualization and analysis clients connect to a distributed array server (i.e., the HPF application with DAQV control) for program-level access to array values. Implementing key components of DAQV in the parallel language (e.g., HPF) itself has led to a robust solution in which clients do not need to know how data is distributed. The framework is highly portable to different computing environments as well as other languages where distributed data is involved.

Parallel iterative solvers for irregular sparse matrices in High Performance Fortran

Writing efficient iterative solvers for irregular sparse matrices in High Performance Fortran (HPF) is hard. The locality in the computations is unclear, and for efficiency we use storage schemes that obscure any structure in the matrix. Moreover, the limited capabilities of HPF to distribute and align data structures make it hard to implement the desired distributions, or to indicate these such that the compiler recognizes the efficient implementation. We propose techniques to handle these problems. We combine strategies that have become popular in message-passing parallel programming, like mesh partitioning and splitting the matrix in local submatrices, with the functionality of HPF and HPF compilers, like the implicit handling of communication and distribution. The implementation of these techniques in HPF is not trivial, and we describe in detail how we propose to solve the problems. Our results demonstrate that very efficient implementations are possible. We indicate how some of the ‘approved extensions’ of HPF-2.0 can be used, but they do not solve all problems. For comparison we show the results for regular sparse matrices.

High performance Fortran for highly irregular problems

We present a general data parallel formulation for highly irregular problems in High Performance Fortran (HPF). Our formulation consists of(1) a method for linearizing irregular data structures (2) a data parallel implementation (in HPF) of graph partitioning algorithms applied to the linearized data structure, (3) techniques for expressing irregular communication and nonuniform computations associated with the elements of linearized data structures.We demonstrate and evaluate our formulation on a parallel, hierarchical N--body method for the evaluation of potentials and forces of nonuniform particle distributions. Our experimental results demonstrate that efficient data parallel (HPF) implementations of highly nonuniform problems are feasible with the proper language/compiler/runtime support. Our data parallel N--body code provides a much needed "benchmark" code for evaluating and improving HPF compilers.

Performance Issues in High Performance Fortran Implementations of Sensor-Based Applications

Applications that get their inputs from sensors are an important and often overlooked application domain for High Performance Fortran (HPF). Such sensor-based applications typically perform regular operations on dense arrays, and often have latency and through put requirements that can only be achieved with parallel machines. This article describes a study of sensor-based applications, including the fast Fourier transform, synthetic aperture radar imaging, narrowband tracking radar processing, multibaseline stereo imaging, and medical magnetic resonance imaging. The applications are written in a dialect of HPF developed at Carnegie Mellon, and are compiled by the Fx compiler for the Intel Paragon. The main results of the study are that (1) it is possible to realize good performance for realistic sensor-based applications written in HPF and (2) the performance of the applications is determined by the performance of three core operations: independent loops (i.e., loops with no dependences between iterations), reductions, and index permutations. The article discusses the implications for HPF implementations and introduces some simple tests that implementers and users can use to measure the efficiency of the loops, reductions, and index permutations generated by an HPF compiler.

PGHPF – An Optimizing High Performance Fortran Compiler for Distributed Memory Machines

High Performance Fortran (HPF) is the first widely supported, efficient, and portable parallel programming language for shared and distributed memory systems. HPF is realized through a set of directive-based extensions to Fortran 90. It enables application developers and Fortran end-users to write compact, portable, and efficient software that will compile and execute on workstations, shared memory servers, clusters, traditional supercomputers, or massively parallel processors. This article describes a production-quality HPF compiler for a set of parallel machines. Compilation techniques such as data and computation distribution, communication generation, run-time support, and optimization issues are elaborated as the basis for an HPF compiler implementation on distributed memory machines. The performance of this compiler on benchmark programs demonstrates that high efficiency can be achieved executing HPF code on parallel architectures.

Experiences in Data-Parallel Programming

To efficiently parallelize a scientific application with a data-parallel compiler requires certain structural properties in the source program, and conversely, the absence of others. A recent parallelization effort of ours reinforced this observation and motivated this correspondence. Specifically, we have transformed a Fortran 77 version of GROMOS, a popular dusty-deck program for molecular dynamics, into Fortran D, a data-parallel dialect of Fortran. During this transformation we have encountered a number of difficulties that probably are neither limited to this particular application nor do they seem likely to be addressed by improved compiler technology in the near future. Our experience with GROMOS suggests a number of points to keep in mind when developing software that may at some time in its life cycle be parallelized with a data-parallel compiler. This note presents some guidelines for engineering data-parallel applications that are compatible with Fortran D or High Performance Fortran compilers.

Scientific Programming with High Performance Fortran: A Case Study Using the xHPF Compiler

Recently, the first commercial High Performance Fortran (HPF) subset compilers have appeared. This article reports on our experiences with the xHPF compiler of Applied Parallel Research, version 1.2, for the Intel Paragon. At this stage, we do not expect very High Performance from our HPF programs, even though performance will eventually be of paramount importance for the acceptance of HPF. Instead, our primary objective is to study how to convert large Fortran 77 (F77) programs to HPF such that the compiler generates reasonably efficient parallel code. We report on a case study that identifies several problems when parallelizing code with HPF; most of these problems affect current HPF compiler technology in general, although some are specific for the xHPF compiler. We discuss our solutions from the perspective of the scientific programmer, and presenttiming results on the Intel Paragon. The case study comprises three programs of different complexity with respect to parallelization. We use the dense matrix-matrix product to show that the distribution of arrays and the order of nested loops significantly influence the performance of the parallel program. We use Gaussian elimination with partial pivoting to study the parallelization strategy of the compiler. There are various ways to structure this algorithm for a particular data distribution. This example shows how much effort may be demanded from the programmer to support the compiler in generating an efficient parallel implementation. Finally, we use a small application to show that the more complicated structure of a larger program may introduce problems for the parallelization, even though all subroutines of the application are easy to parallelize by themselves. The application consists of a finite volume discretization on a structured grid and a nested iterative solver. Our case study shows that it is possible to obtain reasonably efficient parallel programs with xHPF, although the compiler needs substantial support from the programmer.