Re: Benchmarking

[Ivan Voras <ivoras@xxxxxxxxxxxxxx>]

Brian Reichert wrote:

> On Mon, Apr 05, 2004 at 11:44:13PM +0200, Ivan Voras wrote:
>
>>Well, It's over at last. The results in a "world readable" format can be
>>fetched as http://geri.cc.fer.hr/~ivoras/osbench_h.html. There are a few
>>surprises, and I'd like to hear some comments :)
>
>
> I'm not familar with the benchmarks in question.  Would it be
> possible to have the title of the benchmark (eg. 'UBench') link to
> some description of the benchmark?  Basically, for various benchmarks,
> I don't know if big numbers are neccesarily better than small
> numbers...

> Oh, and you misspelled 'Sumarry'. :)
>
> What hardware did you run this on?
>
> Yay!  Collated data...


Yes, sorry for the mess - this is my worksheet in the literal sense - 
everything just thrown together during the work, and english is not my 
native tongue :)

The article to be written will explain all these things and provide 
comments and interpretation of results, but for now:

* bigger is better for ALL numbers
* the machine is P3/933MHz, single SCSI drive, 1GB RAM
* google can be asked about benchmark programs details :)