Open Source Lab: Beaver Challenge 2004

[Robert Garrett]

David Rhodus wrote:

> I do not see any clear specifications that they are testing for
> listed on the web site. I've sent an email to the address listed
> on the web site and have yet to receive a reply. Until I can
> have a clear specification of what the tests would be, I don't
> think the project could even consider this. Even if one is listed
> I don't think it is worth the resources to "compete" in is little
> game. DragonFly is developing the next generation operating
> system, not just building hacks upon an existing infrastructure
> to gain trivial performance/benchmark gains. We are still in the
> process of developing new technologies. Though most of what
> has been developed to this point could be considered extremely
> fast, but only under specific circumstances. There has not been any
> actual performance tuning on the new base technologies at this time.
> It would not be fair at this point to try and benchmark DragonFly in
> this specific context.
> 
> -DR

David,

    I agree that we shouldn't waste resourcess on this, I do believe
that as a whole not participating in a community wide event of this
kind would be bad for the project as a whole. There are some things
to be gained from this. One is simply wider recognition of our existence.


the benchmarks that are to be run are
A mysgl load test
bonnie++
lmbench

are the ones mentioned.


The more I think about it, the more I believe from at least an advocacy
viewpoint that we have to at least enter,  It is clear that where we 
are is not a polished endpoint. I also disagree that gcc3 needs to be
in before this happens.

I do agree whole heartedly that we shouldn't waste a lot of time trying 
to make things benchmark faster, it is the wrong time in our development 
to do that. 

Robert Garrett