[Clas12_first_exp] Clara 4.3.0 update

[Vardan Gyurjyan]

Dear All,

We are releasing an update of the framework under the same version number. This update offers controlled horizontal scaling.
As you recall, for batch processing CLARA accepts input data-set in the form of a metadata file, containing the names of data files to be processed ( files.list). Before this update the files within the files.list was processed serially within the single DPE/node. In this case the only scaling factor was due to the CLARA vertical scaling, i.e. multi-threading through available cores of the node.
In this update we are presenting a new CLARA configuration parameter: farm.scaling
The default value is set to be 0 - no horizontal scaling, however in case user sets farm.scaling=n, that will tell CLARA to process groups of n files from the files.list data-set in parallel, on different farm nodes, thus speeding up input data-set processing n times. 

For more detailed documentation of this feature visit:
https://claraweb.jlab.org/clara//docs/clas/clara-scaling.html

Note: to activate this feature you need to update your local CLARA installations:
https://claraweb.jlab.org/clara//docs/clas/installation.html <https://claraweb.jlab.org/clara//docs/clas/installation.html>


Best regards,
-vardan

---------------------------------------------------
Vardan H. Gyurjyan, Ph.D.
Staff Scientist
Thomas Jefferson Accelerator Facility
Newport News, VA, 23606
E-mail: gurjyan at jlab.org
757-269-5879 (JLAB)



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/clas12_first_exp/attachments/20170823/03878479/attachment.html>