On 08/09/94 at 01:48:00 mark.longridge@canrem.com said:

I leave it to Jerry or Dan to check my results. I checked up to 2
moves deep by hand and verified 10 different positions. What I don't
understand is how Jerry manages to look at so many cube positions:

On full 3x3x3 cube,   7     100,803,036     13.231  (new)

Using 10 bytes to store a single cube position would still
need over 1 billion bytes, or am I missing something?

Well, when I deal with the big problems I want to solve to the
bitter end, I use the M-conjugate and centerless cube tricks
I have described at much too great length in the past. This one
is a quick and dirty program using no conjugate tricks. The
only real "trick" is that I externalize the data. I decided
long ago that the problems I wanted to solve were too big to
keep in memory. Hence, I keep everything in simple (but large)
flat files and sort and merge the files like crazy. In this quick
and dirty program, the cube itself is 13 bytes and the level is
1 byte, for a total of 14 bytes per cube. I guess that makes
the file size about 1.4 gigabytes (10^9). I am
leery of using the word "billion" on E-mail forums because
E-mail is international and "billion" means 10^12 in some
countries.

 = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
Robert G. Bryan (Jerry Bryan)              (304) 293-5192
Associate Director, WVNET                  (304) 293-5540 fax
837 Chestnut Ridge Road                     BRYAN@WVNVM
Morgantown, WV 26505                        BRYAN@WVNVM.WVNET.EDU

If you don't have time to do it right today, what makes you think you are
going to have time to do it over again tomorrow?