Good morning Bob;

It is always nice to sort out the variables and there are more than a few that I can think of.

1. Is the exe on the Client or Server?
2. If the exe is on the Client are you using the same Client to run two instances of the application?
3. Is the temp file used by VFP on the Client or Server?

Most of my VFP applications have 15 to 20 users. One has 250 users at one site and we notice no slow downs as you described.

It is good to optimize your applications ability to handle data. You may find that the experts are not always correct about how to do something.

Have you tried a dbc with stored procedures?

Could you post your code? A final thought is that a particular test may not emulate “real life”. Coming from the electronics engineering world I tend to keep an open mind and try things for myself.

By the way we were in Oregon for a week recently and enjoyed the East, Central, and West Coastal areas. Eugene is a very nice area and a location we are considering for retirement.

Tom - in "Lovely Silicon Valley".


>I seem to be seeing a large degradation of performance in a FoxPro application as more and more workstations load the executable and
>run it over the network.
>
>The degradation becomes severe when two or more workstations are running some data-intensive function.
>
>It is easy to get some part of the program to slow down by a factor of 10 in this way.
>
>Experimentation suggests that some of this is due to FoxPro checking the status of record and file locks, as the degradation increases with the number of users and especially with instructions that either perform a lock or update a table (REPLACE, APPEND,INSERT, DELETE, and TABLEUPDATE).
>
>I wrote a test application in FoxPro that executes a SEEK, gets a record lock in a free table and REPLACEs a value on that record.
>It does this 30,000 times in a loop, using the same record each time.
>
>When I run this executable over the network with two users, each user SEEKing, LOCKing and REPLACEing on a different record,
>the performance typically slows down by a factor of 10. Since there is no contention for record locks, I am surprised by this performance.
>
>I was under the impression that FoxPro has been used with great success in large, data-intensive applications with hundreds of users.
>If so, I am at a loss as to why I am seeing this kind of performance.
>
>Has anyone had experience with this sort of scenario? Is there any way to optimize performance in these circumstances?