Skip to main content


Eclipse Community Forums
Forum Search:

Search      Help    Register    Login    Home
Home » Modeling » EMF "Technology" (Ecore Tools, EMFatic, etc)  » Problem with EMFT Compare
Problem with EMFT Compare [message #96798] Fri, 14 September 2007 21:01 Go to next message
MKTX is currently offline MKTXFriend
Messages: 5
Registered: July 2009
Junior Member
Hi,

We have a major interest in using EMFT Compare and EMFT Search (Whenever
it will be available) for our toolings set. When I tried the simple
examples with smaller ecore models the code (obtained from the EMF Compare
FAQ under Headless Model comparisons) seems to be working fine. However
when I tried our own models that are fairly big in sizes (roughly 300k)
the code seems to be taking much longer for the
"DifferencesServices().modelMatch()" method. Its taking roughly one hour
just to give a good MatchModel back. However the code is later crashing
before doing any difference with "Java Heap Size" exception. I have tried
going to the max. heap size but it did not help with either the extreme
slowness of the first method as well as the Java heap exception. Question
is the code is it suppose to work mostly for relatively smaller model, if
not than can anybody make suggestion to improve the code operation?

Regards.
Re: Problem with EMFT Compare [message #96828 is a reply to message #96798] Mon, 17 September 2007 08:07 Go to previous message
Laurent Goubet is currently offline Laurent GoubetFriend
Messages: 1902
Registered: July 2009
Senior Member
This is a multi-part message in MIME format.
--------------000402060106060003060905
Content-Type: text/plain; charset=ISO-8859-15; format=flowed
Content-Transfer-Encoding: 8bit

Hi,

The comparison time is quite dependent of the model's structure. A
"flat" model with all elements at the same hierarchy level will be way
longer to compare than a well structured model.

We have compared models of size exceeding 5MB with EMF Compare before
0.7 went RC; and though it was slow, it didn't fail with those dreaded
"java heap size" exceptions.

In order to be sure of this, I just created two models of 550k and
compared them both with 0.7RC3 and the current CVS version and didn't
notice any exceptions. Comparison time was about 40sec.

My computer has 2GB of RAM. I run EMF Compare under Eclipse Europa 3.3.0
I20070625-1500 with its default memory settings (I think it was -Xmx512m).

Yours could be models provoking bugs in EMF Compare. Are those Ecore or
UML models? Other possible problems could come from the Eclipse you're
using (we havent tested EMF Compare under 3.1). You'd probably want to
check the "default search window" preference for EMF Compare (under
Eclipse : Window => Preferences => EMF Compare) too and lower its value
to see if you get any improvement. If you cannot find the cause and your
models aren't confidential, I'd like to look further into this.

Cheers.
Laurent Goubet
Obeo

MK a
Re: Problem with EMFT Compare [message #609718 is a reply to message #96798] Mon, 17 September 2007 08:07 Go to previous message
Laurent Goubet is currently offline Laurent GoubetFriend
Messages: 1902
Registered: July 2009
Senior Member
This is a multi-part message in MIME format.
--------------000402060106060003060905
Content-Type: text/plain; charset=ISO-8859-15; format=flowed
Content-Transfer-Encoding: 8bit

Hi,

The comparison time is quite dependent of the model's structure. A
"flat" model with all elements at the same hierarchy level will be way
longer to compare than a well structured model.

We have compared models of size exceeding 5MB with EMF Compare before
0.7 went RC; and though it was slow, it didn't fail with those dreaded
"java heap size" exceptions.

In order to be sure of this, I just created two models of 550k and
compared them both with 0.7RC3 and the current CVS version and didn't
notice any exceptions. Comparison time was about 40sec.

My computer has 2GB of RAM. I run EMF Compare under Eclipse Europa 3.3.0
I20070625-1500 with its default memory settings (I think it was -Xmx512m).

Yours could be models provoking bugs in EMF Compare. Are those Ecore or
UML models? Other possible problems could come from the Eclipse you're
using (we havent tested EMF Compare under 3.1). You'd probably want to
check the "default search window" preference for EMF Compare (under
Eclipse : Window => Preferences => EMF Compare) too and lower its value
to see if you get any improvement. If you cannot find the cause and your
models aren't confidential, I'd like to look further into this.

Cheers.
Laurent Goubet
Obeo

MK a
Previous Topic:[EMF Compare] Release Schedule of 0.8.0
Next Topic:[EMF Compare] Release Schedule of 0.8.0
Goto Forum:
  


Current Time: Sun May 19 11:42:36 GMT 2024

Powered by FUDForum. Page generated in 0.04105 seconds
.:: Contact :: Home ::.

Powered by: FUDforum 3.0.2.
Copyright ©2001-2010 FUDforum Bulletin Board Software

Back to the top