Industrial Data Science
in C# and .NET:
Simple. Fast. Reliable.
 
 

ILNumerics - Technical Computing

Modern High Performance Tools for Technical

Computing and Visualization in Industry and Science

tgt

Performance Limits

We do not claim to be faster than [paste any known language here] in general. The advantages of ILNumerics become most obvious for large data. As a rule of thumb the speed of ILNumerics for many common situations will be in range [0.5 ... 2] compared to equivalent C implementations. But the performance of ILNumerics will further improve over time, since we are working on even more automated parallelizations (namely SIMD support and vector extensions).

 

However, if you are dealing with tight loops in small kernels which run for hours you might be better off with FORTRAN or Assembler still. ILNumerics and .NETs P/Invoke make it really simple and convenient to interface and exchange data with native modules efficiently. That way you can still profit from the excellent framework support and easily integrate native computing libraries into your GUI application.

 

Most users of ILNumerics find themselves doing more and more processing directly in C# over time, leveraging ILNumerics.