Eclipse Community Forums: Epsilon » etl in multithreaded standalone application

Help

Home

Home » Modeling » Epsilon » etl in multithreaded standalone application

Show: Today's Messages :: Show Polls :: Message Navigator

etl in multithreaded standalone application [message #1840509]

Fri, 16 April 2021 20:39

jo ber

Messages: 94
Registered: August 2018

Member

hi

I would like to know what are the guidelines to use etl modules in multithreaded application. In particular I would like to create a pool of etl modules which can be shared across threads. Each thread would take an etl module from the pool and assign a new source model/instances to be transformed. Do you see any error in this approach or situation that requires a particular attention?

Thank you king regards

Report message to a moderator

Re: etl in multithreaded standalone application [message #1840518 is a reply to message #1840509]

Sat, 17 April 2021 10:02

Dimitris Kolovos

Messages: 2163
Registered: July 2009
Location: York, UK

Senior Member

Hi,

In general, E*L modules have not been implemented with "recycling" in mind so my suggestion would be to re-parse your ETL transformation every time you need to use it. On my machine, when I parse the 2.5K-line transformation in [1] in a loop, after the first iteration it is parsed in about ~40ms so unless your transformations are huge, the performance impact should be negligible.

Best,
Dimitris

[1] https://bitbucket.org/bmyildiz/model-derivation-framework/src/e6adecebd44b354425bdedf15d0261c7a3c486bc/Model-Derivation%20Framework/Model%20Transformation%20Modules/jbcpp2uppaal_framework_version.etl?at=master&fileviewer=file-view-default

Report message to a moderator