Second-generation supercompilers should comprise performance analysis tools for creating an information feed-back to improve parallel programs. We have developed a technique and performance analysis tool, PEPSY, embed...
详细信息
parallel programs, generated by the supercompiler, VFCS (Vienna Fortran Compilation System), are an application area for the tool, PEPSY (PErformance Prediction System), which we have developed recently. PEPSY derives...
详细信息
This paper presents a Parameter based Performance Prediction Tool (PPPT) which is part of the Vienna Fortran Compilation System (VFCS), a compiler that automatically translates Fortran programs into message passing pr...
详细信息
Current virtual memory systems provided for scalable computer systems typically offer poor performance for scientiFIc applications when an application's working data set does not fit in main memory. As a result, p...
详细信息
Current compilation systems for distributed memory computers have to integrate new techniques to support the highly complex task of producing efficient programs for parallelsystems. Two techniques, program comprehens...
详细信息
Language extensions of Fortran are being developed which permit the user to map data structures to the individual processors of distributed memory machines. These languages allow a programming style in which global da...
详细信息
For an increasing number of data intensive scientific applications, parallel I/O concepts are a major performance issue. Tackling this issue, we provide an outline of an input/output system designed for highly efficie...
详细信息
parallel programs, generated by the supercompiler, VFCS (Vienna Fortran Compilation System), are an application area for the tool, PEPSY (PErformance Prediction SYstem), which we have developed recently. PEPSY derives...
详细信息
parallel programs, generated by the supercompiler, VFCS (Vienna Fortran Compilation System), are an application area for the tool, PEPSY (PErformance Prediction SYstem), which we have developed recently. PEPSY derives a performance model from an internal representation of the parallel program in the compiler, automatically, and makes performance analysis by discrete-event simulation. Several monitoring modes of PEPSY enable to analyze a parallel program behavior at the detailed and global level, evaluating static and dynamic performance indices. Performance prediction is carried out before the final target parallel code is generated, or before a real run is executed on the parallel computer, in order to optimize the parallelization strategy for VPCS. This paper is focused on the prediction of a new subset of performance indices for the parallel programs, evaluated by PEPSY, and illustrating its new monitoring facilities suitable for comparison of different versions of the program.< >
Performance prediction should be included in the compilation process, by second-generation supercompilers, for creating an information feedback to improve parallel programs. However, application programs might consist...
详细信息
Performance prediction should be included in the compilation process, by second-generation supercompilers, for creating an information feedback to improve parallel programs. However, application programs might consist of a number of procedures calling one another and creating a complex call structure. Every procedure consists of several parallel processes with more or less intensive interprocess communications. This parallel system represents a very time consuming workload for the performance analysis tool. We have designed a hierarchical strategy for analysis of the program call structure, which significantly reduces the total time necessary for the performance prediction of the whole parallel program.
暂无评论