Re: SQL Performance

Tech-Archive recommends: Fix windows errors by optimizing your registry

From: Niall (morann_at_gmail.com)
Date: 11/03/04 

Date: Wed, 3 Nov 2004 15:46:10 -0000

I have changed the isolation level for the transaction being used and still
the same results I have also used the with (nolock) hint on the select
statement.

"JXStern" <JXSternChangeX2R@gte.net> wrote in message
news:79uho05dooorb0kaliv2vl8fsjq0cgrq4n@4ax.com...
> On Wed, 3 Nov 2004 15:15:08 -0000, <microsoft.public.sqlserver.misc>
> wrote:
> >I have a simple sql stress tool that creates threads to simulate
> >simultaneous connections. These threads execute a simple store procedure
> >call on a database ( reading data only ). With a single thread the sp
> >returns in about 80 ms. If I increase the threads to 5, the sp returns in
> >about 400 ms and so on. This linear progression does not seem right. Can
> >anybody suggest why this is happening
>
> If they're all returning the same memory-cached records there is no
> opportunity for overlap. That there isn't a couple of ms available
> for communications overlap is mildly surprising. Or you could be
> running in Serializable mode.
>
> J.
>