Hi,
Im using Firebird 2.1.
I have a table that has about 8 fields of prices and a date/timestamp field.
There are about 7 million records.
There is only one index, and thats on the time stamp field.
I have a routine that does a simple sql select and orders on the date/time
field (That has an index)

I then loop through the entire table, and do some complex math etc on the
data.. top to bottom..
Then, I do it again and again, testing some modeling parameters or "What if"
stuff. I will litterally loop through it 500 times..

No, there is no way to do this in a stored proc..

Anyway, It works fine, but I was doing it on my pentium 4, took about 5
minutes for each iteration.
memory usage was small, and CPU was only about 30 %.

I went out and bought a quad core , with more memory, much faster cpu and
bus, faster drives etc..

Its only a tiny bit faster now.. Im so discouraged with it.. I can imagine
why its not many times faster..

Any idea what the bottle neck might be? I am doing a unidirection query, to
save memory, that helps, Im not using "Fieldbyname", getting the field
directly by TempQuery.Fields.Fields[12].AsFloat.

I dont know what else to do..

Gordon