Eclipse Community Forums
Forum Search:

Search      Help    Register    Login    Home
Home » Modeling » EMF "Technology" (Ecore Tools, EMFatic, etc)  » EMF Compare with large files
EMF Compare with large files [message #536533] Fri, 28 May 2010 15:21 Go to next message
Maxime Jaouen is currently offline Maxime JaouenFriend
Messages: 10
Registered: March 2010
Junior Member
Hello,
I develop an Eclipse RCP application and i use EMF Compare to compare ecore files into this application. But i must compare very large ecore files (between 10Mo and 20Mo), and when those files are larger than 10Mo the comparaison has never end. To optimize a little the comparaison, i use this code to launch comparaison (found in FAQ EMF Compare) :
EObject model2 = ModelUtils.load(origine, resourceSet);
EObject model1 = ModelUtils.load(modif, resourceSet);
MatchModel match = MatchService.doMatch(model1, model2, Collections.<String, Object> emptyMap());
DiffModel diff = DiffService.doDiff(match, false);
With the DiffModel, i can obtains all differences and generate a report. I want optimize EMF Compare for compare my large files. I read an other topic (http:// www.eclipse.org/forums/index.php?t=msg&th=167686&sta rt=0&) and i think the IDs system can help me to optimize the comparaison, but i don't understand how IDs are determinate. (for example, if i generate 2 same ecore, have it the same IDs ?)

Please, can anybody explain me the ID system in ecore files? or give me a hint to solve this problem ?

Thanks a lot,

Maxime




Re: EMF Compare with large files [message #536874 is a reply to message #536533] Mon, 31 May 2010 10:59 Go to previous messageGo to next message
Laurent Goubet is currently offline Laurent GoubetFriend
Messages: 1639
Registered: July 2009
Senior Member
This is a multi-part message in MIME format.
--------------050706000504050801090707
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit

Hi Maxime,

Please use the EMF newsgroup for questions about EMF Compare.

That being said, I don't think that comparing through IDs will help in
any way : if your comparison "never ends" it mostly is a memory issue.
ID comparison only helps with time performances. As for size, we
successfully used EMF Compare with models of a size > 250Mo, so 10Mo
seems "small" even if you use a greatly compressed format.

I only see two things you can do; first things first : use the 1.1
version of EMF Compare if you are not already, we greatly improved the
performances.

Other than that, you could increase the memory settings of your Eclipse
(I usually use -Xmx512m and -XX:MaxPermSize=256m in the eclipse.ini file).

If none of these helps, could you provide us with the problematic
models? There might be something specific we do not handle well.

Laurent Goubet
Obeo

Maxime Jaouen wrote:
> Hello,
> I develop an Eclipse RCP application and i use EMF Compare to compare
> ecore files into this application. But i must compare very large ecore
> files (between 10Mo and 20Mo), and when those files are larger than 10Mo
> the comparaison has never end. To optimize a little the comparaison, i
> use this code to launch comparaison (found in FAQ EMF Compare) : EObject
> model2 = ModelUtils.load(origine, resourceSet);
> EObject model1 = ModelUtils.load(modif, resourceSet);
> MatchModel match = MatchService.doMatch(model1, model2,
> Collections.<String, Object> emptyMap());
> DiffModel diff = DiffService.doDiff(match, false); With the DiffModel, i
> can obtains all differences and generate a report. I want optimize EMF
> Compare for compare my large files. I read an other topic
> ( http://www.eclipse.org/forums/index.php?t=msg&th=167686& amp;start=0&) and i
> think the IDs system can help me to optimize the comparaison, but i
> don't understand how IDs are determinate. (for example, if i generate 2
> same ecore, have it the same IDs ?)
>
> Please, can anybody explain me the ID system in ecore files? or give me
> a hint to solve this problem ?
>
> Thanks a lot,
>
> Maxime
>
>
>
>
>


--------------050706000504050801090707
Content-Type: text/x-vcard; charset=utf-8;
name="laurent_goubet.vcf"
Content-Transfer-Encoding: base64
Content-Disposition: attachment;
filename="laurent_goubet.vcf"

YmVnaW46dmNhcmQNCmZuOkxhdXJlbnQgR291YmV0DQpuOkdvdWJldDtMYXVy ZW50DQpvcmc6
PGEgaHJlZj0iaHR0cDovL3d3dy5vYmVvLmZyIj5PYmVvPC9hPg0KZW1haWw7 aW50ZXJuZXQ6
bGF1cmVudC5nb3ViZXRAb2Jlby5mcg0KdXJsOmh0dHA6Ly93d3cub2Jlby5m cg0KdmVyc2lv
bjoyLjENCmVuZDp2Y2FyZA0KDQo=
--------------050706000504050801090707--
Re: EMF Compare with large files [message #537761 is a reply to message #536874] Thu, 03 June 2010 14:59 Go to previous messageGo to next message
Maxime Jaouen is currently offline Maxime JaouenFriend
Messages: 10
Registered: March 2010
Junior Member
Hi Laurent,

Thanks a lot for your answer.

My version of EMF Compare was 1.0.1, i installed 1.1 and this new version is more effecient. The comparison is faster. I increase the memory settings of Eclipse too like you said me. Now i can use the graphical comparison (i don't use code to launch comparison).

Now my very long comparison (with my 20mo files and with many differences between us) takes 5 min. Do you think it's a normal time ? I observe too when i run the same comparison a second time, the second comparison is slower than the first. Is it a normal thing ?

Thanks a lot,

Maxime
Re: EMF Compare with large files [message #537775 is a reply to message #537761] Thu, 03 June 2010 15:26 Go to previous message
Laurent Goubet is currently offline Laurent GoubetFriend
Messages: 1639
Registered: July 2009
Senior Member
This is a multi-part message in MIME format.
--------------080600000503000107000204
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit

Hi Maxime,

See my comments inlined below.

Laurent Goubet
Obeo

Maxime Jaouen wrote:
> Hi Laurent,
>
> Thanks a lot for your answer.
> My version of EMF Compare was 1.0.1, i installed 1.1 and this new
> version is more effecient. The comparison is faster. I increase the
> memory settings of Eclipse too like you said me. Now i can use the
> graphical comparison (i don't use code to launch comparison).

I didn't understand that you were using graphical comparison. That
invalidates my mentionning of the files greater than 250mo, these
comparisons were "classic" comparisons.

> Now my very long comparison (with my 20mo files and with many
> differences between us) takes 5 min. Do you think it's a normal time ? I

Comparison times greatly depend on the models/diagrams being compared :
whether they're flat or deeply hierarchic, the number of other resources
they depend on... 5 minutes seem rather long, but I don't really know
standard comparison times for diagrams.

> observe too when i run the same comparison a second time, the second
> comparison is slower than the first. Is it a normal thing ?

Doesn't seem normal to me; it might on the contrary denote a memory leak
:/. Do your memory usage / comparison time increase on each comparison?

>
> Thanks a lot,
>
> Maxime


--------------080600000503000107000204
Content-Type: text/x-vcard; charset=utf-8;
name="laurent_goubet.vcf"
Content-Transfer-Encoding: base64
Content-Disposition: attachment;
filename="laurent_goubet.vcf"

YmVnaW46dmNhcmQNCmZuOkxhdXJlbnQgR291YmV0DQpuOkdvdWJldDtMYXVy ZW50DQpvcmc6
PGEgaHJlZj0iaHR0cDovL3d3dy5vYmVvLmZyIj5PYmVvPC9hPg0KZW1haWw7 aW50ZXJuZXQ6
bGF1cmVudC5nb3ViZXRAb2Jlby5mcg0KdXJsOmh0dHA6Ly93d3cub2Jlby5m cg0KdmVyc2lv
bjoyLjENCmVuZDp2Y2FyZA0KDQo=
--------------080600000503000107000204--
Re: EMF Compare with large files [message #622958 is a reply to message #536533] Mon, 31 May 2010 10:59 Go to previous message
Laurent Goubet is currently offline Laurent GoubetFriend
Messages: 1639
Registered: July 2009
Senior Member
This is a multi-part message in MIME format.
--------------050706000504050801090707
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit

Hi Maxime,

Please use the EMF newsgroup for questions about EMF Compare.

That being said, I don't think that comparing through IDs will help in
any way : if your comparison "never ends" it mostly is a memory issue.
ID comparison only helps with time performances. As for size, we
successfully used EMF Compare with models of a size > 250Mo, so 10Mo
seems "small" even if you use a greatly compressed format.

I only see two things you can do; first things first : use the 1.1
version of EMF Compare if you are not already, we greatly improved the
performances.

Other than that, you could increase the memory settings of your Eclipse
(I usually use -Xmx512m and -XX:MaxPermSize=256m in the eclipse.ini file).

If none of these helps, could you provide us with the problematic
models? There might be something specific we do not handle well.

Laurent Goubet
Obeo

Maxime Jaouen wrote:
> Hello,
> I develop an Eclipse RCP application and i use EMF Compare to compare
> ecore files into this application. But i must compare very large ecore
> files (between 10Mo and 20Mo), and when those files are larger than 10Mo
> the comparaison has never end. To optimize a little the comparaison, i
> use this code to launch comparaison (found in FAQ EMF Compare) : EObject
> model2 = ModelUtils.load(origine, resourceSet);
> EObject model1 = ModelUtils.load(modif, resourceSet);
> MatchModel match = MatchService.doMatch(model1, model2,
> Collections.<String, Object> emptyMap());
> DiffModel diff = DiffService.doDiff(match, false); With the DiffModel, i
> can obtains all differences and generate a report. I want optimize EMF
> Compare for compare my large files. I read an other topic
> ( http://www.eclipse.org/forums/index.php?t=msg&th=167686& amp;start=0&) and i
> think the IDs system can help me to optimize the comparaison, but i
> don't understand how IDs are determinate. (for example, if i generate 2
> same ecore, have it the same IDs ?)
>
> Please, can anybody explain me the ID system in ecore files? or give me
> a hint to solve this problem ?
>
> Thanks a lot,
>
> Maxime
>
>
>
>
>


--------------050706000504050801090707
Content-Type: text/x-vcard; charset=utf-8;
name="laurent_goubet.vcf"
Content-Transfer-Encoding: base64
Content-Disposition: attachment;
filename="laurent_goubet.vcf"

YmVnaW46dmNhcmQNCmZuOkxhdXJlbnQgR291YmV0DQpuOkdvdWJldDtMYXVy ZW50DQpvcmc6
PGEgaHJlZj0iaHR0cDovL3d3dy5vYmVvLmZyIj5PYmVvPC9hPg0KZW1haWw7 aW50ZXJuZXQ6
bGF1cmVudC5nb3ViZXRAb2Jlby5mcg0KdXJsOmh0dHA6Ly93d3cub2Jlby5m cg0KdmVyc2lv
bjoyLjENCmVuZDp2Y2FyZA0KDQo=
--------------050706000504050801090707--
Re: EMF Compare with large files [message #622983 is a reply to message #536874] Thu, 03 June 2010 14:59 Go to previous message
Maxime Jaouen is currently offline Maxime JaouenFriend
Messages: 10
Registered: March 2010
Junior Member
Hi Laurent,

Thanks a lot for your answer.

My version of EMF Compare was 1.0.1, i installed 1.1 and this new version is more effecient. The comparison is faster. I increase the memory settings of Eclipse too like you said me. Now i can use the graphical comparison (i don't use code to launch comparison).

Now my very long comparison (with my 20mo files and with many differences between us) takes 5 min. Do you think it's a normal time ? I observe too when i run the same comparison a second time, the second comparison is slower than the first. Is it a normal thing ?

Thanks a lot,

Maxime
Re: EMF Compare with large files [message #622984 is a reply to message #622983] Thu, 03 June 2010 15:26 Go to previous message
Laurent Goubet is currently offline Laurent GoubetFriend
Messages: 1639
Registered: July 2009
Senior Member
This is a multi-part message in MIME format.
--------------080600000503000107000204
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit

Hi Maxime,

See my comments inlined below.

Laurent Goubet
Obeo

Maxime Jaouen wrote:
> Hi Laurent,
>
> Thanks a lot for your answer.
> My version of EMF Compare was 1.0.1, i installed 1.1 and this new
> version is more effecient. The comparison is faster. I increase the
> memory settings of Eclipse too like you said me. Now i can use the
> graphical comparison (i don't use code to launch comparison).

I didn't understand that you were using graphical comparison. That
invalidates my mentionning of the files greater than 250mo, these
comparisons were "classic" comparisons.

> Now my very long comparison (with my 20mo files and with many
> differences between us) takes 5 min. Do you think it's a normal time ? I

Comparison times greatly depend on the models/diagrams being compared :
whether they're flat or deeply hierarchic, the number of other resources
they depend on... 5 minutes seem rather long, but I don't really know
standard comparison times for diagrams.

> observe too when i run the same comparison a second time, the second
> comparison is slower than the first. Is it a normal thing ?

Doesn't seem normal to me; it might on the contrary denote a memory leak
:/. Do your memory usage / comparison time increase on each comparison?

>
> Thanks a lot,
>
> Maxime


--------------080600000503000107000204
Content-Type: text/x-vcard; charset=utf-8;
name="laurent_goubet.vcf"
Content-Transfer-Encoding: base64
Content-Disposition: attachment;
filename="laurent_goubet.vcf"

YmVnaW46dmNhcmQNCmZuOkxhdXJlbnQgR291YmV0DQpuOkdvdWJldDtMYXVy ZW50DQpvcmc6
PGEgaHJlZj0iaHR0cDovL3d3dy5vYmVvLmZyIj5PYmVvPC9hPg0KZW1haWw7 aW50ZXJuZXQ6
bGF1cmVudC5nb3ViZXRAb2Jlby5mcg0KdXJsOmh0dHA6Ly93d3cub2Jlby5m cg0KdmVyc2lv
bjoyLjENCmVuZDp2Y2FyZA0KDQo=
--------------080600000503000107000204--
Previous Topic:EMF Compare NPE
Next Topic:Mwe2.xtext:15 type reference ecore
Goto Forum:
  


Current Time: Sun Dec 21 13:40:32 GMT 2014

Powered by FUDForum. Page generated in 0.02519 seconds
.:: Contact :: Home ::.

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