Hello, I have sent the following mail yesterday but it did non appear in the archive, therefore I resend it.
Let me add that the barrier for working/non-working is 1728/1727. With the larger value the computation leads to the correct result. Interestingly, if I increase the value from 500 in steps to 1728 then the number of solutions found increses until the correct value is reached, too. Cheers, Dirk --- Yesterdays mail --- i have a problem with the function fd_set_vector_max. My CSP needs four variables (A-D) with are constrained to take values from 0 to 563. If I set the fd_set_vector_max value to 564, then the system finds only about 3000 solutions. If I increase fd_set_vector_max dramatically, e.g. to 5640, then all solutions (>160000) are found. I checked with a different method that the >160000 solutions are correct. Please run the attached file and change the lines marked "Fails"/"Works" to see the effect. My question is: Is this a misunderstanding of fd_set_vector_max, my values are 0--N, therefore using N+1 should be ok, or is it a bug in the implementation? Please contact me for further questions. Cheers, Dirk This is my configuration: Running on Ubuntu 6.10 with the provided gprolog packages. Linux nw-08 2.6.17-11-generic #2 SMP Tue Mar 13 23:32:38 UTC 2007 i686 GNU/Linux Prolog top-Level (GNU Prolog) 1.2.18 By Daniel Diaz Copyright (C) 1999-2004 Daniel Diaz GNU Prolog comes with ABSOLUTELY NO WARRANTY. You may redistribute copies of GNU Prolog under the terms of the GNU General Public License. For more information about these matters, see the files named COPYING. configure options are not available. $gprolog [example].
example.pl
Description: Perl program
_______________________________________________ Bug-prolog mailing list [email protected] http://lists.gnu.org/mailman/listinfo/bug-prolog
