A Tool for Performance Modeling of Parallel Programs

Modeling Performance of Parallel Programs

The actual performance of parallel programs is often disappointing, especially in comparison to the peak performance ooered by the underlying hardware. There are many sources of performance degradation and understanding these sources is necessary to improve application performance. In this paper we discuss performance modeling, an approach to understanding the performance of parallel systems. W...

SCALEA: A Performance Analysis Tool for Distributed and Parallel Programs

A Performance Modeling System for Message-Passing Parallel Programs

We present a new performance modeling system for message-passing parallel programs that is based around a Performance Evaluating Virtual Parallel Machine (PEVPM). We explain how a PEVPM model for a particular message-passing program can be developed by inserting PEVPM performance directives that describe the serial segments of computation and the message-passing events that the program contains...

passert: A Tool for Debugging Parallel Programs

passert is a new debugging tool for parallel programs which allows programmers to express correctness criteria using a simple, expressive assertion language. We demonstrate how these parallel assertions allow the detection and diagnosis of real world concurrency bugs, detecting 14/17 bugs in an independently selected set of bugs from open source software. We describe a runtime checker which all...

Andes: a Performance Analyzer for Parallel Programs

ANDES is a performance monitor designed for MIMD distributed memory machines that inserts additional code in the program to be analyzed. ANDES determines the following metrics: speedup, eeciency, experimentally determined serial fraction, percentage of idle time per processor, load and communication balancing, synchronization time and percentage of cpu-communication overlapping. Within ANDES, a...