scientific article; zbMATH DE number 1737489
A performance simulator tool for parallel systems is presented. The simulator is embedded into an integrated parallel program development environment in which parallel applications based on the message-passing programming paradigm can be designed by graphical means. It is a discrete event simulator that uses a synthetic description of a parallel application and that of a parallel architecture. Synthetic model of the application is extracted from the real application code automatically although, some of the required parameters must be supplied by the user manually. The simulator also takes into account a simplified model of the underlying operating system. The output of the simulation is a tracefile based on the NPICL format can be visualised and analyzed with standard tools like Paragraph. After introducing the model and the architecture of their simulation environment, the authors presents some validation results obtained during practical experiments with an image convolution application.
- Performance limitations of parallel simulations
- PAPS -- A testbed for performance prediction of parallel applications
- Realistic parallel performance estimation
- Reducing parallel program simulation complexity by static analysis
- A New Strategy to Improve the Performance of PDP-Systems Simulators
- Parallel system simulation
- Hybrid Performance Modeling and Prediction of Large-Scale Parallel Systems
- Performance prediction model for parallel computing on network of workstation
- An Integrated Simulation Environment for Parallel and Distributed System Prototying
- scientific article; zbMATH DE number 1737490 (Why is no real title available?)
- EDEPPS: An integrated graphical toolset for the design and performance evaluation of portable parallel software
- Modeling cost/performance of a parallel computer simulator
- Early prediction of MPP performance: The SP2, T3D, and Paragon experiences
- A trace based performance evaluation tool for parallel real time systems
- Compiler-optimized simulation of large-scale applications on high performance architectures
- Interconnection network simulation using traces of MPI applications
This page was built for publication:
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4529593)
