Hi Norman,

Because of the huge discrepancy I tried it again. It took 1 second to
run it for a 40,000 element array; 2 seconds for 80,000; 4 or 5 seconds
for 200,000; and 10 or 11 seconds for 300,000.

It would be interesting to see how fast it runs on Dana DeLouis's machine.

Alan Beban

Norman Jones wrote:

Hi Alan,

On the slowest machine that I could conveniently find, I ran Dana's routine
(amended for the OP's 600k array) in slightly over 3.2 seconds - well under
one tenth of the time you report!

---
Regards,
Norman


"Alan Beban" wrote in message
...

With the 20's replaced by 600000, the size of the OP's array, it took 39
seconds to run on my machine. I know I don't always bow to efficiency,
but that's a lot even for me :-)

Alan Beban