Skip to main content


Eclipse Community Forums
Forum Search:

Search      Help    Register    Login    Home
Home » Modeling » M2T (model-to-text transformation) » [Acceleo] standalon vs plugin execution time
[Acceleo] standalon vs plugin execution time [message #1817593] Tue, 26 November 2019 11:20 Go to next message
Paolo Bocciarelli is currently offline Paolo BocciarelliFriend
Messages: 1
Registered: November 2019
Junior Member
Hi! I wrote an Acceleo M2T transformation and I noticed something I cannot understand.
If I run the transformation inside Eclipse or as a standalone java application, it works fine but the transformation execution requires a very long time (more than 10 minute).
I' ve also exported the transformation as a Plugin and I have installed the jars (both the transformation and the generated UI) into Eclipse.
The execution of the *same* transformation on the *same* input file produces the *same* output but in few seconds instead of more that 10 minutes.
Any idea or suggestion to understand the reason for the above-mentioned behavior??
The input model is quite complex sysml model (generated from magicdraw) which makes use of some profiles. The size of the main .uml file is more than 20MB and it includes more that 200k xml rows....
I can understand that a complex transformation applied on a large file might require long time to be executed. But I cannot figure out why the execution time as a plugin is different from the execution time as standalone application.
Any suggestion is very welcome.

Paolo
Re: [Acceleo] standalon vs plugin execution time [message #1817707 is a reply to message #1817593] Thu, 28 November 2019 06:40 Go to previous message
Ed Willink is currently offline Ed WillinkFriend
Messages: 6538
Registered: July 2009
Senior Member
Hi

My guess is that initialization is a problem causing an optimization option to be missed, but it might be a bloat of some facility in one context provoking long searches.

The only sensible way to tell is to use a profiler such as JVisualVM or YourKit to measure.

Please report back so that others can see what can be done well/badly. Maybe the developers can then discourage the 'bad' approach.

Regards

Ed Willink
Previous Topic:How to convter msg to vcf
Next Topic:Access to target folder from within Acceleo generation
Goto Forum:
  


Current Time: Sun Dec 15 00:50:59 GMT 2019

Powered by FUDForum. Page generated in 0.01942 seconds
.:: Contact :: Home ::.

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

Back to the top