Re: Benchmark: NetBSD 2.0 beats FreeBSD 5.3

Robert Ryan wrote:
> Fellow FreeBSD developers,
> 
> I hate to say I told you but it was inevitable.
> 
> Check this out: http://www.feyrer.de/NetBSD/gmcgarry/


Well, even though this link was submitted by a troll, the benchmarks are 
actually pretty interesting and worth reading.  Just keep the following 
in mind:

1. Although NetBSD did perform better on many of the benchmarks (kudos 
to them), but the difference is usually not dramatic.
2. The benchmarks are strictly uniprocessor benchmarks.
3. The author does admit at the end of the article that NetBSD still 
uses a big, giant lock around the kernel and the benchmarks might be 
very different on a multiprocessor system.
4. The benchmarks are only NetBSD versus FreeBSD, so it's hard to judge 
where the performance of each system fits in the grand scheme of things. 
  It could be that both systems are performing very well.  The 
benchmarks need to include at least one non-BSD system (usually Linux) 
in order to get some perspective.
5. It would have been nice if FreeBSD 4.10 and NetBSD 1.6.2 were also 
include, so we could see the relative progress (or lack) of each system.

But it's worth reading.  Data is always a good thing.  Just don't get 
hung up on them.  Hopefully, it will inspire more comprehensive tests.

Richard Coleman
rcoleman_at_criticalmagic.com