>
>
> Here's an example on my old AMDk6-450 (old ide drives
> :-).
>
> File length of text file: 3.1 MB
> record count: 100.000
> insert into bla select * from tmp_table takes _20_
> seconds. I don't think, that this is bad.
> select (*) from tmp_table
> Result: 100000
>

A similar example (10M file of 100000 records) on my PII 333

Db2 v8.1 loads in 3s
Postgres 7.3.3 copy's in 9s

I suspect that our machines are similar enough for the results to be
interesting
( I have quite good ide drives, which might help me a bit, but my cpu is
slower...)

If you guys want to have a go at the files I am using see

http://sourceforge.net/projects/benchw

and get the latest cvs (it makes the generated files more interesting)

In addition it hopefully explains why I am(was) interested in a bulk
load api - I would like to add Firebird to the list of db managers in
the test suite.

Cheers

Mark