Hello Miles.

10 Aug 04 05:18, you wrote to David Williams:

 MM> G'morning David,

 DW>> So I made that change, and did a few other little fiddles. I
 DW>> think you'll find that this version runs to completion in 3
 DW>> seconds, or possibly even 2!

 MM> Interesting !  It still took 4 seconds here, but I saw much more
 MM> of the early solutions that I did in the earlier runs - which
 MM> suggests its slowed up in the initial solving, and speeded up in
 MM> the later stages...

 MM> Make any sence ?

Yes, we've been timing it with the printing disabled, display output is the
biggest slow-down. by speeding up the algorithm the results would come
faster at the end where the algorithm is most sluggish but at the start it
will still seem the same speed (as printing is the bottleneck)





 MM> ___ MultiMail/MS-DOS v0.45

 MM> --- Maximus/2 3.01

Jasen

--- GoldED+/LNX 1.1.4.7