Re: Performance difference between 2 machines

"Fabrizio" <fabrizio.magni_at_mycontinent.com> wrote in message news:42b9123c$0$21079$892e7fe2_at_authen.white.readfreenews.net...
> DA Morgan wrote:
>>
>> The reason for the 16K block size? Perhaps someone has been reading too
>> many books by authors that haven't backed up their theoretical
>> projections with actual benchmarking.
>>
>> You should be using 4K blocks with AIX.
>
> I could be considered heretic for claiming this...
>
> ... but maybe they knew what were doing.
> If using async i/o (having a good readahead setting) than a 16k is not a
> bad choice for a DWH.
> I would even think of larger block size.
>
> And on post-8i versions you can exploit direct i/o on filesystem so the fs
> block size "constraint" is greatly relaxed.
>
> --
> Fabrizio Magni
>
> fabrizio.magni_at_mycontinent.com
>
> replace mycontinent with europe

Mmm ... this is getting more complicated than I thought ;-) Yes we are using async i/o, but no direct i/o (although the filesystem buffer cache has been set pretty low, so this might have the same effect as activating direct i/o ?).

Now what about RHLES 3.0 on VMWare 5.0 on Windows 2000 ? ;-)

Matthias Received on Wed Jun 22 2005 - 13:55:00 CDT