[Mono-list] Re: Time problems on Mono

shudo@computer.org shudo@computer.org
Sun, 12 Dec 2004 21:42:33 +0900 (JST)

> From: Mauro Parra Miranda <mauro@novell.com>
> Date: Sat, 11 Dec 2004 23:35:36 -0600

> > Perhaps it would be nice to port LINPACK test to C# and see the differences
> > between runtimes
> There is a work arround about this: 
> http://www.shudo.net/jit/perf/

Yes.  I have ported the Java version of Linpack benchmark to C#.
The porting work was reasonably easy due to similarity between the two

Now, the comparison page has the source code.

Note that the ported C# version still uses jagged 2-dimentional arrays
in the same way as the original Java version, not rectangular arrays.
The Java version achieves the pivoting (row exchange) by an exchange
of pointer values in an array.  Pivoting by a pointer exchange is
possible because the arrays are jagged arrays.  In C#, arrays can be
rectangular arrays (similar to C and FORTRAN), and performance of
array accesses will be improved.  On the other hand, pivoting gets
slow a little with rectangular arrays.  I expect the C# version will
be improved with rectangular arrays.

  Kazuyuki Shudo
  Grid Technology Research Center
  National Institute of Advanced Industrial Science and Technology (AIST)