Maude-help mailing list

[Steven Eker <eker AT csl.sri.com>]

Cool. I always need examples like that. Maude 3 (don't hold your breath - 
it's 
not due out for at least a year) will have a new rewrite engine with based on 
a radically new design that should make cred irrelevant. But extreme 
equational term rewriting systems (1000000s of equations, equations with 
100000s of function symbols occurrences, signatures with 1000s of sorts per 
component and 1000s of subsort overloaded operators, 1000000s of operators 
total, billions of rewrites to normal form etc) are _very_ valuable for 
optimizing the design.

So by all means send it. It should just be a file that I can load and see 
stuff 
happening without having to understand what is happening - I'm really only 
interested in monitoring what is happening in Maude's critical loops. I don't 
want to look at the code, and especially don't want to have to uncomment 
reduce statements etc.

Steven

On Wednesday 11 May 2011 01:10:58 am Marc Boyer wrote:
> Le 09/05/2011 22:39, Steven Eker a écrit :
> > If you want to post your Maude program I can take a quick look at it for
> > obvious inefficiencies. Actually I'm always interested in seeing pure
> > 'reduce' code that runs slowly - we are planned a brand new rewrite
> > engine for Maude 3 and it's good to see what features need to be
> > accelerated.
> 
> What did you call "running slowly" ?
> 
> I have a pure reduce code that goes from 14 to 1670000 rewrites per
> second, with an average of 400 000, on a Intel(R) Core(TM)2 Duo CPU
> running at 1.4GHz.
> 
> It takes about 24h to make 30 000 reductions (the input file size is 
>  1,3M).
> 
> Marc Boyer
>