In the past days, I've decided to play with Patsolve, which is Dr. Tom
Holroyd's solver for Freecell and several other variants of Card Solitaire.
Tom was a moderately active contributor to this mailing list, and he
corresponded with me and with Adrian Ettlinger, the chief programmer of
Freecell Pro. As a result, Patsolve was integrated into the Freecell Pro Solve
Evaluation Edition.
I remember that Patsolve-3.0 was made roughly as fast as Freecell Solver was
at the time. Since I incorporated many speed optimisations since then, I
decided to benchmark it again on this machine.
Here are the results for the first 32K MS Freecell Deals:
{{{
Started at 1242737477.395835071
Ended at 1242738863.377175298
Took: 1385.98134016991 seconds (23.08833392813 deals per second)
The UNIX time command says:
461.51user 19.50system 23:05.97elapsed 34%CPU (0avgtext+0avgdata
0maxresident)k
0inputs+258224outputs (0major+1277937minor)pagefaults 0swaps
}}}
I compiled patsolve using gcc and:
{{{
CFLAGS = -O3 -march=pentium4 -fomit-frame-pointer -I. $(GAMEOPTS) -Wall
}}}
This now makes Freecell Solver over 10 times faster than patsolve-3.0.
I should note that the patsolve-3.0.tgz build process is very quirky - it does
not compile out of the box, and "make clean" followed by "make" fails. I'd
like to release a patsolve-shlomif-3.0.1 with these problems sorted out.
Regards,
Shlomi Fish
--
-----------------------------------------------------------------
Shlomi Fish http://www.shlomifish.org/
Parody on "The Fountainhead" - http://xrl.us/bjria
God gave us two eyes and ten fingers so we will type five times as much as we
read.
Received on Tue May 19 2009 - 06:44:31 IDT