Speed 2003 VS 97

[Guest]

I am very dissapointed with Access 2003. I just recently upgraded to 2003
from 97. Of course I also upgraded my previous DBs. Man can we say SLOW!!!
Processing time of methods as well as records is just horrible in comparrison.

Can anyone tell me why, or perhaps how to improve performance?

Any suggestions will be appreciated.

Thanks

 

[John W. Vinson]

I am very dissapointed with Access 2003. I just recently upgraded to 2003
from 97. Of course I also upgraded my previous DBs. Man can we say SLOW!!!
Processing time of methods as well as records is just horrible in comparrison.

Can anyone tell me why, or perhaps how to improve performance?

Check out Tony's suggestions:

http://www.granite.ab.ca/access/performancefaq.htm

The website refers to A2000, but the suggested solutions work (and benefit)
later versionis as well.

John W. Vinson [MVP]

 

[Guest]

Thanks for the input, but I wasn't able to get to the site. Each attempt
resulted in a page not found message.

 

[Allen Browne]

It does look like Tony Toews's site is down at the moment.

It is worth persuing though: he has a very good check list covering Name
AutoCorrect, subdatasheets, long path names, record-level locking, JET
service packs, and a bunch of other things that are each significant and
collectively make a huge impact.

Once you deal with those issues, you should find that A2003 is almost as
fast as A97.

However, if you have not converted the back end yet, that will make a
difference too. When an A2003 front end is connected to an A97 back end, JET
4 must thunk everything to JET 3.5. If you have many records (hundreds of
thousands) the difference is perceivable.

 

[Guest]

Allen,

Thanks too for your input. I have converted both backend and frontend DBs.
I am seeing signifcant performance issues just in the proccessing time of my
methods. Computations and such are taking a drastic amount of time in
contrast to the performance offered when operating as 97 versions. Subforms
requery much slower and so on... Queries of course perform just as poorly.
As far as records go, I am only working with about 85,000 records in one
particular case. Again, everything was lightning fast in 97, as apposed to
the 2003 version.

Thanks again,

MPM

 

[Guest]

Allen and John,

Is it possible for both Access 2003 and 97 to reside and operate on the same
machine?

 

[Allen Browne]

Yes, it's fine to have both versions. (The machine I'm answering from has
A97, 2000, 2002, 2003, and 2007 installed.)

Things like subdatasheets can really slow you down.

If you have not read this article, it might help:
Converting from Access 97 to 2000 and later
at:
http://allenbrowne.com/ser-48.html

 

[John W. Vinson]

Is it possible for both Access 2003 and 97 to reside and operate on the same
machine?

I'm running both... along with 2.0, 2000 and 2002. Haven't installed 2007 yet,
because by all accounts it REALLY doesn't "play nice" with older versions.

John W. Vinson [MVP]

 

[David W. Fenton]

It is worth persuing though: he has a very good check list
covering Name AutoCorrect, subdatasheets, long path names,
record-level locking, JET service packs, and a bunch of other
things that are each significant and collectively make a huge
impact.

Once you deal with those issues, you should find that A2003 is
almost as fast as A97.

I find that the apps I've designed are just as fast in both, simply
because today's machines are so fast and have so much RAM.

However, if you have not converted the back end yet, that will
make a difference too. When an A2003 front end is connected to an
A97 back end, JET 4 must thunk everything to JET 3.5. If you have
many records (hundreds of thousands) the difference is
perceivable.

I don't think this is actually true. I ran one of my clients' apps
with an A97 back end used by an A2K front end for a couple of years
before converting the back end to A2K. It slowed down with the
conversion. I did all the usual things you mention that Tony
suggests (since we were all learning about them together back in the
early days of A2K). I really don't think there's much "thunking"
required, since Jet 4 is really a superset of Jet 3.5, rather than
being fundamentally different. That is, it has a few different data
types, but includes all the data types that Jet 3.5 had. It also has
a few new table properties, but it doesn't omit or alter any of the
properties that Access 97/Jet 3.5 had. The only real difference is
Unicode support, but my bet is that would work in Jet 3.5's favor.

And I'm not talking about a trivial app. It had 15 simultaneous
users back then with 3 main tables over 350K records.

 

[Allen Browne]

My comment was not based on measurements, David.

It was based on an app where we converted the back end to JET 4 for other
reasons, and the client responded with, "Thanks, it's quite a bit faster as
well." Half a dozen users. 100mbit LAN. 200k records in some tables. Some
forms have about 10 subforms (on tab control.)

The JET 4 dll does in fact have to convert calls to JET 3.x, so the
perceived performance difference makes sense.

(This does not imply that JET 4 in general is faster than JET 3.x; merely
that translation is slower.)

 

[Tony Toews [MVP]]

Thanks for the input, but I wasn't able to get to the site. Each attempt
resulted in a page not found message.

The web server was down for a while over night. We kicked it and it's
fine now.

Tony
--
Tony Toews, Microsoft Access MVP
Please respond only in the newsgroups so that others can
read the entire thread of messages.
Microsoft Access Links, Hints, Tips & Accounting Systems at
http://www.granite.ab.ca/accsmstr.htm

 

[David W. Fenton]

My comment was not based on measurements, David.

It was based on an app where we converted the back end to JET 4
for other reasons, and the client responded with, "Thanks, it's
quite a bit faster as well." Half a dozen users. 100mbit LAN. 200k
records in some tables. Some forms have about 10 subforms (on tab
control.)

Exactly the opposite of my experience. Perhaps it had something to
do with the fact that the app was originally Access 2.

The JET 4 dll does in fact have to convert calls to JET 3.x, so
the perceived performance difference makes sense.

But my point is that the translation is *trivial*, as the vast
majority of calls will be identical. It's definitely using the Jet 4
DLL to do it, as you can access data in earlier versions of Jet
without needing the DLLs for the earlier versions.

So, I really think there is very little overhead involved.

(This does not imply that JET 4 in general is faster than JET 3.x;
merely that translation is slower.)

I don't think it's significantly slower at all, if it's even slower.