Clips 6.24 & Miss Manners issue

[Oskar45]

Hi,

I tried the benchmark available at http://osdir.com/ml/java.jess/2004/msg00713.html
but did run into a problem: everything works fine up to about "seat
123" but then it
starts to oscillate between seat 123 and seat 126.

Can anyone please confirm that the dataset has indeed a problem?

Thanks.

/oskar

[Peter Lin]

I've run manners in clips 6.2 and 6.3 just fine in the past.

peter lin

On Nov 2, 4:31 am, Oskar45 <oskar4...@yahoo.com> wrote:
> Hi,
>

> I tried the benchmark available athttp://osdir.com/ml/java.jess/2004/msg00713.html

[Alex_G]

Hi Peter

Do you still have the files with those MissManners benchmarks?

I tried my new 64-bit linux build with the ones that Oskar gave,
above, but they won't load.
I commented out the offending lines and did the rest of the load
manually, but then it went into a non-terminating loop, which I
stopped after the 500th (out of 128) seats was allocated! :)
Anyone fancy looking at that?

BTW, all the standard supplied examples work on my 64-bit Jaunty
build, so I'd suspect the supplied code before anything else,
especially since Oskar also reported a problem.

Thanks
Alex

[CLIPS Support]

There are other versions of the benchmark here:
http://clipsrules.svn.sourceforge.net/viewvc/clipsrules/examples/.
If you're compiling CLIPS 6.3 as a 64-bit, verify that you have the
latest source: http://clipsrules.svn.sourceforge.net/viewvc/clipsrules/core/

On Nov 2, 3:31 am, Oskar45 <oskar4...@yahoo.com> wrote:
> Hi,
>

> I tried the benchmark available athttp://osdir.com/ml/java.jess/2004/msg00713.html

[Alex_G]

Thanks for that MissManners link "CLIPS support", I'll check it out
ASAP.

<OT>BTW, the 64-bit build that I made was indeed with the latest
source, as per a previous message of yours.
Seems to work just fine, though it did say that the "system" command
is not yet implemented, and it would be nice if terminal "cursor
navigation" and "history" could be preserved.
Is this reasonably doable?</OT>

I'll report back on the MM benchmark soon as I have a chance.

[Alex_G]

For the sake of completeness:

I have just run the manners code from sourceforge.
Here's the output (of the 128 seat version)

<output>
8639 rules fired4762 mean number of facts (8953 maximum).
1 mean number of instances (1 maximum).
138 mean number of activations (9490 maximum).
</output>

I hope that's reasonable, I haven't checked details.
Time taken just over 5 seconds.
Machine PentiumD (HPxw4300) Ubuntu Jaunty. Compiled with gcc using
Anjuta.

[CLIPS Support]

Those are the same stats that I get.