Distributed benchmarking network

[John Nilsson <john milsson nu>]

Hi,

I was thinking about an optimization problem today an had an idea of a
benchmarking network. So I thought I'd pass it your way to determine
it's usefulness and if deemed useful its feasibility.

So, the idea is simply to have a distributed computing network that can
be used to run benchmarks.

Some benefits as I can see.

* GNOME runs on a wide variety of platforms with an equally wide variety
of performance characteristics. This could be a way to get data from all
of those platforms.

* Some benchmarks will have a large number of candidate solutions. This
network would be a way to distribute the benchmarking load to get
quicker results.

* People without the skills needed to contribute with patches could
still find a way to help out by simply running a benchmarking server.

* This could be a feasible way of applying genetic algorithms to some
optimization problems.

What do you think?

Regards,
John Nilsson