The article presents the software architecture and implementation of the problem-solving environment (PSE) PELLPACK for modeling physical objects described by partial differential equations (PDEs). The scope of this PSE is broad, as PELLPACK incorporates many PDE solving systems, and some of these, in turn, include several specific PDE solving methods. Its coverage for 1D, 2D. and 3D elliptic or parabolic problems is quite broad, and it handles some hyperbolic problems, Since a PSE should provided complete support for the problem-solving process, PELLPACK also contains a large amount of code to support graphical user interfaces, analystic tools, user help, domain or mesh partitioning, machine and data selection, visualization, and various other tasks. Its total size is well over 1 million lines code. Its open-ended software architecture consists of several software layers. The top layer is an interactive graphical interface for specifying the PDE model and its solution framework This interface saves that results of the user specification in the form of a very high level PDE language which is an alternative interface to the PELLPACK system. This language also allows a user to specify the PDE problem and its solution framework textually in a natural form. The PELLPACK language preprocessor generates a Fortran control program with the interfaces, calls to specified components and libraries of the PDE solution framework, and functions defining the PDE problem. The PELLPACK program execution is supported by a hign-level tool where the virtual parallel system is defined, where the compilation, loading, and execution are controlled. Finally, the PELLPACK PSE integrates several PDE libraries and PDE systems based on the decomposition of discrete geometric data to map sparse PDE computations to parallel machines. An instance of the system is available as aWeb server(WebPELLPACK) for public use at http://pellpack.cs.purdue.edu.

We describe pHPF, an research prototype HPF compiler for the IBM SP series parallel machines. The compiler accepts as input Fortran 90 and Fortran 77 programs, augmented with HPF directives; sequential loops are automatically parallelized. The compiler supports symbolic analysis of expressions. This allows parameters such as the number of processors to be unknown at compile-time without significantly affecting performance. Communication schedules and computation guards are generated in a parameterized form at compile-time. Several novel optimizations and improved versions of well-known optimizations have been implemented in pHPF to exploit parallelism and reduce communication costs. These optimizations include elimination of redundant communication using data-availability analysis; using collective communication; new techniques for mapping scalar variables; coarse-grain wavefronting; and communication reduction in multi-dimensional shift communications. We present experimenta...

Optimizing compilers (particularly parallel compilers) are constrained by their ability to predict performance consequences of the transformations they apply. Many factors, such as unknowns in control structures, dynamic behavior of programs, and complexity of the underlying hardware, make it very difficult for compilers to estimate the performance of the transformations accurately and efficiently. In this paper, we present a performance prediction framework that combines several innovative approaches to solve this problem. First, the framework employs a detailed, architecture-specific, but portable, cost model that can be used to estimate the cost of straight line code efficiently. Second, aggregated costs of loops and conditional statements are computed and represented symbolically. This avoids unnecessary, premature guesses and preserves the precision of the prediction. Third, symbolic comparison allows compilers to choose the best transformation dynamically and systematically. Some methodologies for applying the framework to optimizing parallel compilers to support automatic, performance-guided program restructuring are discussed.