Zoltan is a software toolkit for parallel partitioning, load balancing, and data management services. Zoltan also contains implementations of parallel algorithms for graph coloring and matrix ordering. Zoltan 3.0 was released in May 2007, and version 3.1 in September 2008. Zoltan v3.1 is available both as a stand-alone and as a package in Trilinos.
Isorropia is primarily a matrix-based interface to Zoltan, and provides tight integration with Trilinos via Epetra matrices. It is distributed as a package in Trilinos.
OpenAD is an open source framework for automatic differentiation of numerical computer programs. It implements the source transformation paradigm via a language-independent abstract intermediate representation. It has support for source codes written in C and Fortran.
ADIC is a source-transformation based tool for the automatic differentiation of programs written in ANSI C. Uses the OpenAD framework.
OpenAD/F is a source-transformation based tool for the automatic differentiation of programs written in Fortran 77. Uses the OpenAD framework.
ADIFOR is a source-transformation based tool for the automatic differentiation of programs written in Fortran 77.
Rapsodia is a code generator for the efficient computation of high-order derivative tensors. Can generate C++ or Fortran code that is accessed by using operator overloading.
ColPack is a software package consisting of C++ implementations of serial algorithms for a variety of graph coloring and related problems in sparse derivative matrix computation. First version released in Oct 2008. Parallel implementations of some of its coloring algorithms are made available in Zoltan 3.1.
MatchBox is a C++ software package consisting of sequential exact as well as approximation algorithms for a variety of graph matching problems. The package is currently under development. A trial copy can be obtained by request.
MatchBoxP is a software package currently being written. It will consist of parallel approximation algorithms for various matching problems implemented in C++ and using MPI.
The Data_N_Comp_Reorder package includes various sparse matrix reordering algorithms that aim to improve the data locality and therefore the performance of computations that iterate over the sparse matrix data structure.