shutting off Spotlight

matt neuburg wrote:
> Denominator <flash DeleteThis @invalid.invalid> wrote:
>
>> Xbench tells me my Mini's disk is 20% slower than other Minis of the
>> same type with the same drive.
>>
>> I think I should run the test with Spotlight off. The only way I know
>> to turn it off is to put the HD icon in Privacy. In that case I'll have
>> to let it reindex. Is that the best approach?
>
> No. If you want to turn Spotlight off you should *really* turn it off.
> Say
>
> mdutil -i off /

Thanks. Some figures within the disk test changed a lot. Overall my
score was lower with it off.
>
> and then restart the computer.

Uh-oh, I didn't restart. I created a folder on the desktop and
spotlight didn't find it, so I assumed spotlight was off.

However, I doubt that this is going to
> account for the problem. My guess is that the disk is very full /
> fragmented, and/or this Mini lacks sufficient RAM compared to the
> others. m.
>
I'm using 17 g of a 40 g disk. I don't know how to check for
fragmentation.

I've got 512MB memory. On the list, the ones with 1024 don't seem to do
any better on disk scores.

One reason I ran the test was to see how much a firewire drive might
improve performance. My drive score might more than double, but my
overall score would rise about 15%.

It looks as if something called q2d could improve performance a lot. I
wonder what q2d is.

This message is not archived

Ilgaz Öcal wrote:
> On 2007-01-27 21:04:35 +0200, Denominator <flash.RemoveThis@invalid.invalid> said:
>
>> Xbench tells me my Mini's disk is 20% slower than other Minis of the
>> same type with the same drive.
>>
>> I think I should run the test with Spotlight off. The only way I know
>> to turn it off is to put the HD icon in Privacy. In that case I'll
>> have to let it reindex. Is that the best approach?
>
> XBench's disk benchmark is known to be a bit problematic. In fact, as
> all benchmark programs it should write random pattern or a pattern of
> zeroes as a file and there is no way Spotlight will try to index the
> file. Perhaps couple of mini seconds until it figures it is not a file
> which could be indexed.
>
> Another thing is, Spotlight index services run at "idle" (lowest)
> priority so they can't normally effect really used applications.
>
> I don't think your Mini's slowness (if it is real?) comes from
> spotlight. If it is real, it could be coming from a memory resident
> utility such as virus checker.
>
> Unless Xbench creates a huge PDF, Rich Text, Text file in your home
> directory, Spotlight can't be the problem.
>
> A Mac site created own simple benchmark just involves copying big files
> in Finder and it proved to be much more effective. Problem is, I forgot
> the site
>
> Ilgaz
>
>
You were right about Spotlight, and I couldn't find that other site. I
don't have a virus checker. I wish I could identify the processes in
Activity Monitor.

It turns out scores like mine are common for Minis like mine with drives
like mine. I wonder why some are faster with no difference I can see.