IntBLAS

1.0

Introduction

Ever since Ray Moore proposed interval arithmetic in the early of 1970's, there have been numerous results of interval computing in various application areas. To meet the needs for interval basic linear algebra subroutines (BLAS) as software development tool, the Basic Linear Algebra Subprograms Technical (BLAST) Forum proposed an interval BLAS standard as a part of its new BLAS standard. This is an implementation, named IntBLAS, of the interval BLAS standard. Implemented with ISO/ANSI standard C++, IntBLAS is cross-platform and architecture independent. By taking the advantages of object-oriented features of C++, IntBLAS treats intervals, interval vectors and matrices like other regular basic numerical types. Hence, we overload the operators in the standard mathematics library <cmath> wherever appropriate for seamless integration. For example, the operators + and - can be used to add and subtract two intervals, interval vectors, or interval matrices. Only interval specific operations are defined with newly named class member functions. IntBLAS uses directed rounding for rigor and double precision as default to record real intervals.

Installation

To install this package you first must download the source and unpack it. If you are using Linux/UNIX and you have gcc then you simply need to run make install inside the gcc directory. To install using Visual Studio .Net 2003, simply run the install.bat file in the ./vs/bin directory. For more detailed instructions, see the README.html

Usage

The IntBlas library is easy to use, it is simply a static library. Sample C++ programs in the sample directory are ready to be compiled and run.
Generated on Wed Apr 26 16:12:03 2006 for IntBLAS by  doxygen 1.4.4