[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [henshin-user] LHS Match takes very long.

Hi Jürgen,

thanks for your interesting question, for convenience I'm attaching a screenshot of your rules.

The answer is yes, this behavior is expected here. The "find" rule is an example of a Performance Bad Smell called "LHS not connected" [1]. The matcher basically computes all possible combinations of A and B nodes (25000000 in your example), which can take pretty long. A more efficient strategy of handling such situations might be possible, but we currently fall short of providing one.

[1] See page 4 in https://pdfs.semanticscholar.org/1d11/051c5fa03749ae51b488b6e9fe3ad131e916.pdf

Kind regards,
Daniel

Am 07.09.2017 um 13:39 schrieb Juergen Freund:
Dear Henshin users and developers,


attached you will find a project where I try to find 2 independent classes within a rather big graph (10000 nodes).


The rule I mean is called "find". If I execute it on the before mentioned graph on my machine it takes excessive long for the match.

But if a delete one of the classes of the rule it is pretty fast for the match.

I just wanted to ask, if this is normal behavior.


There is another rule "createMuch" which is used to build this big graph.


Kind regards,
Jürgen



-- 
Dipl.-Ing. Jürgen Freund
Arbeitsgruppe Ähnlichkeitsmechanik

Telefon: +49 (0)711 685 69530

Postadresse:
Universität Stuttgart
Institut für Statik und Dynamik der Luft- und Raumfahrtkonstruktionen
Pfaffenwaldring 27 - Zimmer 01.002
70569 Stuttgart


_______________________________________________
henshin-user mailing list
henshin-user@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/henshin-user

Attachment: rules.png
Description: PNG image