[RC5] Has anyone else noticed this??

Paul paul.m.nz at gmail.com
Mon Apr 23 19:53:38 EDT 2007


> Never ever benchmark anything inside a virtual machine. Clock in the
> VM isn't accurate enough for measuring small quantities of time. You
> can feed dnetc some work and measure time with a "wall clock" (host OS
> clock, your watch, NTP? :) ). Then you can calculate the real keyrate.

This warning is also mentioned in the article below - I'll redo the
tests and use a wall clock.

In fact I've just done the tests and your quite correct. The results
reported by the the guest OS were completely wrong. When it stated it
took 6 minutes, it in the real world took 11 minutes - thanks for the
info.

> AFAIK, VMware's license forbids benchmarking at all. Can anyone
> correct me on this?

I wasn't actually trying to benchmark, I was simply trying to get the
best key rate possible (for my own reasons - not for public scrutiny).
I only posted to the list since the results seemed a little weird and
to check if any else had similar experience.

They (vmware) seem keen to get feedback on benchmarking performance:
http://blogs.vmware.com/console/2007/04/the_benchmarks_.html

/paul


-- 
Vista is "dramatically more secure than any other operating system
released" Bill Gates
Huh ??  Defender doesn't stop spyware (Webroot) ; firewall is only 50%
effective (Zdnet) ; UAC can be turned off and is annoying ; SP1 (incl
security reasons) due end of 2007...
-----
Try Torpark; a small portable, open-source, built on Firefox browser
that enables anonymous browsing. Requires no installation :
http://www.torrify.com/


More information about the rc5 mailing list