I ran the Random-DFS scan again, this time with the updated test order of
[01][23456789]. What this means in practice is that tests 0 and 1 (moving
top cards to the foundations) will always be performed before all the
other tests. Here are the results:
* It took 131 minutes to run on a Pentium III 600 MHz. If we normalize it
to 667 MHz (which was the case for the machine of the previous test) we
get 117 minutes. The previous one took 145 minutes. In any case, that
makes it 273 boards per minute.
* 22 boards were reported as intractable (in comparison to 16 for
[012345678]). Some boards appear in both sets, but many are unique to one
of the two scans.
* The run took a total of 40,425,981 iterations (remember that there were
5 scans in parallel for each board). That's 1263 iterations in average for
every board. Dividing it by 5 we get an average of 252 iterations per
board per scan.
By downloading the latest development Freecell Solver you can run such
tests yourself. A call to the run program should be done as follows:
./mptest 1 32000 20 [ scan options] { --next-instance [scan options] {
--next-instance [scan options] ... } } } }
Where [scan options] are exactly the same as in fc-solve. (just don't
forget to pipe the output to a file, because everything is written to
stdout). If you manage to come up with a configuration that solves the
boards in a smaller number of iterations and/or has fewer intractables,
please let me know.
Regards,
Shlomi Fish
--
----------------------------------------------------------------------
Shlomi Fish shlomif_at_vipe.technion.ac.il
Home Page: http://t2.technion.ac.il/~shlomif/
Home E-mail: shlomif_at_techie.com
If:
1. A is A
2. A is not not-A
does it imply that
1. B is B
2. B is not not-B
Received on Sun Dec 02 2001 - 12:02:45 IST