Framework for blockstructured adaptive finite volume methods. Provides MPI-parallelized variant of the Berger-Oliger AMR algorithm for Beowulf-clusters. Uses Clawpack. Visualization and conversion tools for HDF4 files included.
Performance Timing Tools
Ichnaea is a set of tools that aid in collecting and tracking parameters and timings for parallel applications. The Performance Modelling Timing Module, PMTM, is a library that wraps system timing calls to abstract these from code developers and aid portability. It also has functionality to store parameters and print those, along with the timing information to a comma separated variable file. The Performance Modelling Analysis Tool, PMAT, is coming soon. This is able to read in and store the outputs from PMTM and produce some graphs based on user-defined criteria. Documentation found in the Wiki. As of March 2015, main development has been switched to the UK Mini-Application Consortium Github page. Work on PMTM from the unreleased version 2.6.0 can be found here: https://github.com/UK-MAC/PMTM Work on PMAT version 3 will be commencing shortly here: https://github.com/UK-MAC/PMAT Tarballs of releases will continue to be available here.
Open FMM ( http://www.openfmm.net ) is and open source EM software project. It consist of a fast parallel two-dimensional TM/TE solver called Nero2d.
The Parallel Seismic Earth Imaging System (PSEIS) is designed for geophysical research and parallel processing of large datasets. Interoperation and emulation are used to leverage other seismic processing software, existing skills and infrastructure.
In many parallel scientific applications the process of grid generation, decomposition, message passing, and performance measurements are of similar stucture. The projects intention is to develop a generic function library that will contain all these.
Pydusa is a package for parallel programming using Python. It contains a module for doing MPI programming in Python. We have added parallel solver packages such as Parallel SuperLU for solving sparse linear systems.
The STAR-MPI (Self-tuned adaptive routines for MPI collective routines) project is an on-going research and development effort to apply the adaptive software approach to develop efficient MPI collective routines across platforms and applications.
Users-Grid is a true virtualized infrastructure for Grid computing which provides computing power transparently. Its defining feature is automatic or seamless job submission. It enables existing software or applications to run on the grid.
An "Alternative IO" layer.
The alio library can be dynamically pre-linked to existing executables and replaces file IO (i.e. calls to glibc's IO functions). This allows to gather statistics about IO, create IO trace files, or even implement a different IO behaviour (e.g. use separate IO servers to handle IO).