These benchmarks are just to give an idea of the relative performance. I am not trying to prove anything.
I tried to make sure that the benchmarks were optimized for speed in the same condition (esp. for C++ and CrossNet).

We are comparing Crossnet and .NET 2.0, and where it makes sense with VC++ 2005.
All C++ / CrossNet benchmarks are done with latest version of dlmalloc (used in Crossnet for unmanaged allocations like with STL).

All benchmarks are run a first time before the actual measure to reduce the impact of previous cache state or JITTING.

The tested configuration is a bi-Pentium D 3.0 GHz - 2 GB of RAM on Windows XP SP2.

Arithmetic Benchmark
Flow Benchmark
.NET features Benchmark
Garbage Collector benchmark

So in summary, nothing to be ashamed of.

Last edited Sep 6, 2007 at 4:45 AM by OlivierNallet, version 3


No comments yet.