Subruns: Difference between revisions

From MidasWiki
Jump to navigation Jump to search
No edit summary
No edit summary
Line 2: Line 2:
to split a big data file into several individual data files. This feature might be helpful if a data file from a single run gets too large (>4 GB for example) and if shorter runs are not wanted for efficiency reasons. The scheme works as follows:
to split a big data file into several individual data files. This feature might be helpful if a data file from a single run gets too large (>4 GB for example) and if shorter runs are not wanted for efficiency reasons. The scheme works as follows:
   
   
In the  ODB [[Explanation of Keys in the ODB /Logger/Channels tree #/Logger/Channels/<logging-channel-number>/Settings subtree | /Logger/Channels/x/Settings/ subtree]] except where noted
In the  ODB [[Keys in the ODB /Logger/Channels tree #/Logger/Channels/<logging-channel-number>/Settings subtree | /Logger/Channels/x/Settings/ subtree]] except where noted


* If a new subrun is to be started  
* If a new subrun is to be started  
** after a certain '''time''' :
** after a certain '''time''' :
*** Set  [[Explanation of Keys in the ODB /Logger tree #/Logger/Subrun duration | /Logger/Subrun duration]] key to the required time in seconds.
*** Set  [[Keys in the ODB /Logger tree #/Logger/Subrun duration | /Logger/Subrun duration]] key to the required time in seconds.
** or after a certain '''number of bytes''' :
** or after a certain '''number of bytes''' :
*** Set  [[Explanation of Keys in the ODB /Logger/Channels tree #/Logger/Channels/0/Settings/Subrun Byte limit | Subrun Byte limit key]] to the required number of bytes for the subrun
*** Set  [[Keys in the ODB /Logger/Channels tree #/Logger/Channels/0/Settings/Subrun Byte limit | Subrun Byte limit key]] to the required number of bytes for the subrun


*    Set [[Explanation of Keys in the ODB /Logger/Channels tree #/Logger/Channels/0/Settings/Filename | Filename key]] to something like run%05d_%02d.mid. The first %05d will be replaced by the run number, while the second one will be replaced by the subrun number. This will result in files such as
*    Set [[Keys in the ODB /Logger/Channels tree #/Logger/Channels/0/Settings/Filename | Filename key]] to something like run%05d_%02d.mid. The first %05d will be replaced by the run number, while the second one will be replaced by the subrun number. This will result in files such as


     run00001_00.mid    run #1
     run00001_00.mid    run #1
Line 22: Line 22:




Each subrun will contain an ODB dump if this is turned on via [[Explanation of Keys in the ODB /Logger/Channels tree #/Logger/Channels/0/Settings/ODB Dump | ODB Dump key]].
Each subrun will contain an ODB dump if this is turned on via [[Keys in the ODB /Logger/Channels tree #/Logger/Channels/0/Settings/ODB Dump | ODB Dump key]].
The stopping of the "main" run (after four subruns in the above example) can be done in the usual way (event limit in the front-end, manually through [[odbedit]], etc.).
The stopping of the "main" run (after four subruns in the above example) can be done in the usual way (event limit in the front-end, manually through [[odbedit]], etc.).


'''Note:''' the user may manually request a new subrun to the started by setting the  [[Explanation of Keys in the ODB /Logger tree #/Logger/Next subrun | Next subrun key]] to "y".
'''Note:''' the user may manually request a new subrun to the started by setting the  [[Keys in the ODB /Logger tree #/Logger/Next subrun | Next subrun key]] to "y".

Revision as of 11:34, 12 August 2013

A subrun scheme is available in the MIDAS event logger utility (mlogger) to split a big data file into several individual data files. This feature might be helpful if a data file from a single run gets too large (>4 GB for example) and if shorter runs are not wanted for efficiency reasons. The scheme works as follows:

In the ODB /Logger/Channels/x/Settings/ subtree except where noted

  • If a new subrun is to be started
  • Set Filename key to something like run%05d_%02d.mid. The first %05d will be replaced by the run number, while the second one will be replaced by the subrun number. This will result in files such as
   run00001_00.mid    run #1
   run00001_01.mid      "
   run00001_02.mid      "
   run00001_03.mid      "
   run00002_00.mid    run #2
   run00002_01.mid      "
   run00002_02.mid      "
   run00002_03.mid      "


Each subrun will contain an ODB dump if this is turned on via ODB Dump key. The stopping of the "main" run (after four subruns in the above example) can be done in the usual way (event limit in the front-end, manually through odbedit, etc.).

Note: the user may manually request a new subrun to the started by setting the Next subrun key to "y".