FreeTDS Development Group

["Gad Hayisraeli" <gadh AT vmanage.com>]

> -----Original Message-----
> From: freetds-admin AT lists.ibiblio.org
> [mailto:freetds-admin AT lists.ibiblio.org]On Behalf Of Lowden, James K
> Sent: þã 05 ôáøåàø 2003 19:48þ
> To: 'freetds AT lists.ibiblio.org'
> Subject: RE: [freetds] slow freetds-sybase perfromance - thorough test
>
>
> > From: Steve Langasek [mailto:vorlon AT netexpress.net]
> > Sent: February 5, 2003 11:20 AM
> >
> > The two lines I see that directly compare FreeTDS
> > to the Windows driver without changing any other variables are these:
> >
> > linux       fastlinux       yes     odbc++ over freetds     1.587
> 16.38
> > windows     fastlinux       yes     odbc++ over ASE odbc    0.753   34
>
> Also interesting:
>
> fastlinux     fastlinux       local  odbc++ over freetds       1.57
> 16.56
> linux         fastlinux       yes     odbc++ over freetds      1.59    16.38
>
> Indicates that for this test FreeTDS is almost unaffected by the presence of
> a network, when using TDS 5.0.  Or is it 4.2, Gad?

Gad: 5.0

>
> If we accept that the FreeTDS linux client is 50% as fast as the the windows
> client, then this pair is very weird (which is how Gad got started, right?):
>
> linux         windows         yes     odbc++ over freetds      5.28
> 4.92
> windows       windows         local   odbc++ over ASE odbc     0.14
> 185.00
>
> Something is clearly, um, suboptimal here.  Microsoft has the opportunity to
> use shared memory (or a giant packet size) on the local server, and well
> might.  Even so, it's still hard to explain the difference.  One would
> suppose the Linux client could at least approach the 16 lps speed it had
> with the Linux server.

> I hope you can get FreeTDS working on your Win32 client, so we can see the
> apples-to-apples comparison.  If you have a nonlocal windows client
> available, that would help too.  That would reintroduce the network, and
> it's possible that Sybase on Win32 has poor network characteristics.

i'll try this

>
> Just to double check: Are you using FreeTDS 0.60?  If so, you really should
> get last night's snapshot (don't use rc1).  We discovered a bug some time
> ago that caused FreeTDS to write to the log file even when it should not.
> That would definitely affect performance, and is consistent with the "wide
> tables are slower" observation.  Also,

i tried with debug log and without it and it did not show any difference .

>
> > freetds trace  = yes
> > freetds trace level = 10
>
> If "freetds trace  = yes" means "dump file = /tmp/freetds.log" in
> freetds.conf, that's a problem, because you're logging your queries, which
> is very expensive.
>
> Thank you for the data, Gad.  Yours is the first objective measure I can
> recall.
>
> --jkl
>
>
> The information contained in this transmission may contain privileged and
> confidential information and is intended only for the use of the person(s)
> named above.  If you are not the intended recipient, or an employee or agent
> responsible for delivering this message to the intended recipient, any
> review, dissemination, distribution or duplication of this communication is
> strictly prohibited. If you are not the intended recipient, please contact
> the sender immediately by reply e-mail and destroy all copies of the
> original message. Please note that we do not accept account orders and/or
> instructions by e-mail, and therefore will not be responsible for carrying
> out such orders and/or instructions.
>
>
> _______________________________________________
> FreeTDS mailing list
> FreeTDS AT lists.ibiblio.org
> http://lists.ibiblio.org/mailman/listinfo/freetds
>
>