Hi,
One more question below, but first, .....
Thanks Torsten, it did confuse me for a short while, especially because
the OPI was convinced it was in the line above?????????
But I found it eventually.
And thanks Raphael, I have it working now although it doesn't seem to
help with excessively large numbers. (Although I don't get the virtual
memory exhausted message any more, just a very,very slow system :) ). I
have a suspicion though that this may be due to the nature of the search
tree, not the parallel search itself. At the moment I'm just trying to
test it with problem sizes that a single search doesn't cope with.
I just have one more question if someone would be kind enough.......
I presume there is some activity that must be performed on a remote
machine to set up Mozart to accept incoming work requests.
I've read around the website (although I've probably missed the full
explanation) and have seen the computation server info. Is this the type
of thing that needs to be executed on a remote machine?
Also, I only use version 1.3.2, because 1.4.0 doesn't (or didn't) work
on the odd Vista machine I'm sometimes forced to use but mainly because
Ubuntu's Debian repository only has 1.3.2 and I'm too lazy to compile
from source! :)
Regards
Mark
Torsten Anders wrote:
Dear Mark,
Just in case: there is a little typo in Raphael's code, a missing
parenthesis
Xs={E one({Queens 20} $}
This should read
Xs={E one({Queens 20} $)}
I can confirm that this example works on Mozart 1.3.2, but please
remember that problems for parallel search have been reported for
Mozart 1.4.0.
Best,
Torsten
--
Torsten Anders
Interdisciplinary Centre for Computer Music Research
University of Plymouth
Office: +44-1752-586219
Private: +44-1752-558917
http://strasheela.sourceforge.net
http://www.torsten-anders.de
On 31 Oct 2008, at 11:23, Raphael Collet wrote:
Dear Mark,
The functor that you give to the search engine must export the script
of the problem. Here you export a function that creates the script!
You should remove the function from your functor, and keep the
procedure only:
declare
functor Queensdist
import FD
export Script
define
proc {Script Row}
L1N=...
...
end
end
You can even write a function that returns a functor. This makes it
possible to generate a script functor for a given size N of the
problem (no more hardcoded):
declare
fun {Queens N}
functor
import FD
export Script
define
proc {Script Row}
L1N={MakeTuple c N}
LM1N={MakeTuple c N}
in
{FD.tuple queens N 1#N Row}
for I in 1..N do
L1N.I=I LM1N.I=~I
end
{FD.distinct Row}
{FD.distinctOffset Row LM1N}
{FD.distinctOffset Row L1N}
{FD.distribute generic(value:mid) Row}
end
end
end
E={New Search.parallel init(localhost: 2)}
Xs={E one({Queens 20} $} % solves Queens for N=20
{Browse Xs}
Cheers,
raph
On Thu, Oct 30, 2008 at 11:50 PM, mark richardson
<[EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>> wrote:
Hi,
I'm still learning about the FD aspect of Mozart-Oz (amongst
other things) and I've been toying with using parallel search
with the n-queens script from the documentation, but with little
success. This is the code I'm using:
declare Queensdist E Xs M Host
functor Queensdist
import FD
export Script
define
fun {Script}
proc {$ Row}
L1N ={MakeTuple c 200}
LM1N={MakeTuple c 200}
in
{FD.tuple queens 200 1#200 Row}
{For 1 200 1 proc {$ I}
L1N.I=I LM1N.I=~I
end}
{FD.distinct Row}
{FD.distinctOffset Row LM1N}
{FD.distinctOffset Row L1N}
{FD.distribute generic(value:mid) Row}
end
end
end
[M]={Module.apply [Queensdist]}
E={New Search.parallel init(localhost:2)}
Xs={E one(Queensdist $)}
{Browse Xs}
Browsing Xs just gives me [<P/1>]. Can anyone offer any suggestions?
I'm not certain about the use of localhost as this isn't my host
name, but the actual hostname doesn't seem to be accepted at all
and the documentation says that localhost refers to the host that
spawned the oz process.
Before this version, I had kept the parametrised version which
worked on it's own with a {Browse {SearchOne {M.script 200}}} but
I amended the script to have a fixed value of 200 instead.
Needless to say, the script won't work either way now???
So, can this script still be parametrised like the original
n-queens script and if so how?
Regards
Mark
--
Mark Richardson
Final year undergraduate
University of Teesside
_________________________________________________________________________________
mozart-users mailing list
[email protected] <mailto:[email protected]>
http://www.mozart-oz.org/mailman/listinfo/mozart-users
<ATT00001.txt>
------------------------------------------------------------------------
_________________________________________________________________________________
mozart-users mailing list
[email protected]
http://www.mozart-oz.org/mailman/listinfo/mozart-users
--
Mark Richardson
Final year undergraduate
University of Teesside
_________________________________________________________________________________
mozart-users mailing list
[email protected]
http://www.mozart-oz.org/mailman/listinfo/mozart-users
