->  DW> So I made that change, and did a few other little fiddles. I think 
->  DW> you'll find that this version runs to completion in 3 seconds, or 
->  DW> possibly even 2! 
 
-> Interesting !  It still took 4 seconds here, but I saw much more  
-> of the early solutions that I did in the earlier runs - which  
-> suggests its slowed up in the initial solving, and speeded up in  
-> the later stages... 
 
-> Make any sence ? 
  
To me, it makes more sense to think that the PRINTing is taking up a 
large fraction of the time, so speeding up the calculations is having 
little effect. If you disable the PRINT line, and make it calculate out 
to  A=45000 so as to take an appreciable time, you should see an 
improvement between the latest version and previous ones. 
  
Beyond about 45000, the thing bombs because "long" integers overflow. 
  
                                  dow 
--- Platinum Xpress/Win/WINServer v3.0pr5