Subject Re: FB 2.0.1 CpuAffinity
Author jack_engleman
I run my development on a AMD Athon Chip Compaq portable with 1 GB of
Memory and a slow disk 60 GB 4200 drive. But in running FB 1.5, I can
routinely run faster that 2.8 Ghertz Pentinium with 2 GB of Main
Memory and that have raid 5 - 3 disk drive arrays. I believe that the
writes to the array can be slower than writing to my local drive on my

Sometimes the disk arrays can run much slower like almost 5 or 6 times
slower, I am not sure why, but these systems can handle eighty
terminals and the Firebird database with a 1.5 GBYTE database in a
production environment and keep up with the online environment and the
load on the database. Much of my testing is running database
maintenance routines and updates, which are normally only done when
all users are off of the customers servers at night.

I have found that having the system restore turned on will slow my
development down about 10 times in Delphi and Firebird so I keep it
off on my portable and I have my customers run their servers with it
off also.

So I am running into the same situation as you have noticed.

These computers always make sense.


--- In, "technisoft2005" <preber@...>
> --- In, Svein Erling Tysvær
> <svein.erling.tysvaer@> wrote:
> >
> > > --- In, "technisoft2005" wrote:
> > > I have concluded that the customers problem is a major hardware
> or
> > > software problem orginating somehwere else as certain operations
> on my
> > > 2Ghz notebook run about 10 times faster than on his 3.2Ghz
> machine with
> > > otherwise similar specs. Hyperthreading is also disabled in the
> > Hi Peter!
> >
> > The speed difference between your notebook and his server could be
> due to different PLANs (just prepare statements to compare) or one or
> more long running transactions (check the statistics and see if next
> transaction is close to oldest (active) transaction). Transferring
> over a network could of course also be a bottleneck.
> >
> > Just to name a few possibilities if you haven't checked them
> already.
> >
> Thank you Set. The gap was rather large but doing the necessary
> things to reduce the gap, I received no comments from the customer. I
> interpret this as the speed increase being too small to be noticeable.
> Peter
> Technisoft