dotNum - .NET numerics

A free/open-source numeric library

The dotNum project complements the .NET class library with high performance numeric functions for math, engineering, scientific and financial applications. The class library can use highly optimized mathematical libraries such as ATLAS or the Intel Math Kernel Library. Without these optional libraries, dotNum uses its own CLS compliant implementations.

The code was originally part of QuantLib.NET, but was phased out to its own project due to general usefulness.

• Documentation

• Frequently asked questions

• Downloads

• Optimized libraries

• License

• General .NET Links

Documentation

Documentation for the Numerics and Random assemblies is now available.

The dotNum project consist of the class library and some optional modules, all available from CVS:

• dotNum

  This is the .NET class library. Documentation is currently not published, but will be included with the first release. For now, use CVS to get the code and generate the XML documentation yourself.

• dotNum\Native\MKL

  This project contains the necessary glue to use Intel's Math Kernel Library with dotNum. Modify the makefile to point to your MKL directory and copy the resulting DLL to your executable Make sure that the MKL-DLLs are available (from the same directory or in the PATH).

• dotNum\Native\ATLAS

  This project contains the necessary glue to use the Atlas Library with dotNum. Modify the makefile to point to your ATLAS directory and copy the the resulting DLL to your executable. (If you get linker errors, comment out all statements from atlas/interfaces/blas/F77/src/xerbla.f)

• dotNum\Tests

  To build and run the unit tests, you have to either checkout the NUnit module from CVS, or install a local copy of NUnit, which is available from Sourceforge.

Downloads

Download QuantLib.NET which contains the Stochastix.Numerics assembly.

Optimized Math Libraries

• Intel Math Kernel Library

  The MKL is composed of highly optimized mathematical functions for math, engineering, scientific and financial applications requiring high performance. It contains LAPACK, the basic linear algebra subprograms (BLAS), and a set of vectorized transcendental functions, offering both high performance and excellent accuracy compared to the libm functions for most of the processors.

• ATLAS

  The ATLAS (Automatically Tuned Linear Algebra Software) project is an ongoing research effort focusing on applying empirical techniques in order to provide portable performance. At present, it provides C and Fortran77 interfaces to a portably efficient BLAS implementation, as well as a few routines from LAPACK.

• NetLib BLAS / LAPACK

  The reference implementation from NetLib. You can get pre-compiled Win32 binaries from here.

• Compaq CXML

  A copy of this might be included with your Visual Fortran distribution.

Copyright and License

Copyright (c) 2001-2003 Jens Thiel (http://www.jens-thiel.de/) and others.
All rights reserved.

See http://stochastix.net/license for licensing details.