My Results for the Rook Paths Puzzle

The number of rook paths on an MxN sized grid is called W(M,N) by Steven Finch. I will adopt this notion here.

The tables for W(3,*) up to W(12,*) were computed at 9-Jul-1994 by me (Heiner Marxen) on an Apollo HP9000-710. Computation stopped, when the number length exceeded 63 decimal places. In brackets is given floor(log10(W)) (i.e. length-1). Up to W(13,13) the results have been confirmed by Andreas Gammel.
The table for W(13,*) has been computed on an SGI IRIX workstation (R4400) at 20-May-1995.
The table for W(14,*) and W(15,*) have been computed on an HP9000/819 at 10-Sep-1996 and 11-Sep-1996.

W(1,N) is not given, it is proven to be 1.
W(2,N) is not given, it is proven to be 2(N-1).
Some timing info:

    W( 3,*)  <  1 sec
    W( 4,*)  <  1 sec
    W( 5,*)  <  1 sec
    W( 6,*)  <  1 sec
    W( 7,*)  <  1 sec
    W( 8,*)  <  3 sec
    W( 9,*)  < 11 sec
    W(10,*)  < 51 sec,   using  0.7 MB
    W(11,*)  <  4.5 min, using  3.2 MB
    W(12,*)  < 28   min, using 14.7 MB

    W(13,*)  < 103  min, using  4.45 MB
    W(14,*)  < 211  min, using 12.09 MB
    W(15,*) ca. 16.5 hours, using 33.36 MB

The results have been split up in order to allow for partial transfer (44 KB total).


Back to the Rook Paths Page.
Back to the home page of Heiner Marxen.
$Date: 2003/06/30 15:30:29 $