Page History

...

Function:	runInLoopParallelProcessorCount, set the amount of cores available to Delft-FEWS when running ensemble workflowsworkflow portions in parallel in one forecasting shell instance

Module Name:	runInLoopParallelProcessorCount
Where to Use?	global properties file
Why to Use?	to speed-up ensemble runs on multi core machines
Description:	The runInLoopParallelProcessorCount en try entry in the global properties files indicated the number of cores Delft-FEWS may use when running ensemble members in a loop
Preconditions:	2009-02 release, multi core cpu or multi cpu computer
Outcome(s):	speed-up of the computations
Scheendump(s):	link to attached screendump(s) for displays only
Remark(s):	The speedup that may be obtained is highly dependent on the type of module you are running
Available since:	DelftFEWS200902

Overview

Delft-FEWS can, within one Forecasting Shell instance, split ensemble workflows (that have the runInLoop element set to true) over multiple cores. Based on the available amount of cores a number of queues is made, one for each core. When running the activity the different ensemble members are added to the different queues. Since FEWS 2017.01 it is optionally possible to run parallel activities on multiple forecasting shells. This can also apply for ensemble workflows as is explained in section 'parallel' on page 06 Configuring WorkFlows.

...

By default the General Adapter runs are performed in the %TEMP_DIR% directory.

Configuration

By default Delft-FEWS will only use one core and all tasks are run one after another. To enable the parallel running of ensemble members the runInLoopParallelProcessorCount entry must be set in the global properties file. Here you either specify the number of cores to use or specify 100 to use all available cores.

...

For all internal Delft-FEWS modules that have been tested no changes are needed to the configuration. For external modules that are run using the General adapter some changes may be needed to the configuration.

Tested modules

Module	Remarks
Transformation (old)	Test ok
Interpolation	Test ok. Interpolation via DLL not tested
TransformationModule (new)	Test ok. Interpolation via DLL not tested
pcrTransformation	Test ok
General Adapter	test ok

Sample input and output (with explanation of all options)

An example for parallel running an import and then processing of the import modules is:

...

The colon characters ":" will be replaced by an underscore "_". Use the internal variables with % characters (like %TEMP_DIR%) and the global.properties variables with $ characters (like e.g. $DUMP_DIR$).

Combination of ensembles and normal time series

You may have a combination of for example a rainfall ensemble with other non-ensemble timeseries (like structure operation or boundary levels). In case you run an ensemble with the runInLoop option at workflow level by default all timeseries are used from the defined ensembleId and member, also for a non-ensemble timeseries. That means that you should configure transformations to create also an ensemble of your non-ensemble series, which is of course not convenient. To enable non-ensemble timeseries FEWS has an option that overrules the ensemble member forcing at workflow level. Therefore you define for the non-ensemble series the fixed ensembleId "main". This is a "virtual" ensemble.

...

Code Block

	xml
	xml

<errorModelSet>
  <inputVariable variableId="observation">
    <timeSeriesSet>
      <moduleInstanceId>Import</moduleInstanceId>
      <valueType>scalar</valueType>
      <parameterId>Q.obs</parameterId>
      <locationId>NA_Mastenbroek</locationId>
      <timeSeriesType>external historical</timeSeriesType>
      <timeStep unit="hour"/>
      <relativeViewPeriod unit="hour" start="-96" end="0" startOverrulable="true" endOverrulable="false"/>
      <readWriteMode>read only</readWriteMode>
      <ensembleId>main</ensembleId>
    </timeSeriesSet>
  </inputVariable>
  <inputVariable variableId="update_run">
    <timeSeriesSet>
      <moduleInstanceId>Sobek_Update</moduleInstanceId>
      <valueType>scalar</valueType>
      <parameterId>Q.sim.hist</parameterId>
      <locationId>NA_Mastenbroek</locationId>
      <timeSeriesType>simulated historical</timeSeriesType>
      <timeStep unit="hour"/>
      <relativeViewPeriod unit="hour" start="-96" end="0" startOverrulable="true" endOverrulable="false"/>
      <readWriteMode>read only</readWriteMode>
      <ensembleId>main</ensembleId>
    </timeSeriesSet>
  </inputVariable>
  <inputVariable variableId="forecast_run">
    <timeSeriesSet>
      <moduleInstanceId>Sobek_Forecast</moduleInstanceId>
      <valueType>scalar</valueType>
      <parameterId>Q.sim.for</parameterId>
      <locationId>NA_Mastenbroek</locationId>
      <timeSeriesType>simulated forecasting</timeSeriesType>
      <timeStep unit="hour"/>
      <relativeViewPeriod unit="hour" start="-96" end="120" startOverrulable="true" endOverrulable="true"/>
      <readWriteMode>read only</readWriteMode>
      <ensembleId>EPS</ensembleId>
    </timeSeriesSet>
  </inputVariable>
  <autoOrderMethod>
    <orderSelection>true</orderSelection>
    <order_ar>3</order_ar>
    <order_ma>1</order_ma>
    <subtractMean>true</subtractMean>
    <boxcoxTransformation>false</boxcoxTransformation>
    <lambda>0</lambda>
    <observedTimeSeriesId>meting</observedTimeSeriesId>
    <simulatedTimeSeriesId>forecast_run</simulatedTimeSeriesId>
    <simulatedTimeSeriesId>update_run</simulatedTimeSeriesId>
    <outputTimeSeriesId>corrected</outputTimeSeriesId>
  </autoOrderMethod>
  <interpolationOptions>
    <interpolationType>linear</interpolationType>
    <gapLength>6</gapLength>
  </interpolationOptions>
  <minResult>0</minResult>
  <ignoreDoubtful>true</ignoreDoubtful>
  <outputVariable variableId="corrected">
    <timeSeriesSet>
      <moduleInstanceId>Sobek_Forecast</moduleInstanceId>
      <valueType>scalar</valueType>
      <parameterId>Q.updated.for</parameterId>
      <locationId>NA_Mastenbroek</locationId>
      <timeSeriesType>simulated forecasting</timeSeriesType>
      <timeStep unit="hour"/>
      <relativeViewPeriod unit="hour" start="-96" end="120" startOverrulable="true" endOverrulable="true"/>
      <readWriteMode>add originals</readWriteMode>
      <ensembleId>EPS</ensembleId>
    </timeSeriesSet>
  </outputVariable>
</errorModelSet>

Error and warning messages

Description of errors and warnings that may be generated

Error:	Error message
Action:	Action to fix

Known issues

Running modules in parallel means you will use more memory

...

Other factors that influence this are the amount of data being retrieved and store in the FEWS database in relation to the total execution time and (in the case of an external module) the amount of data written to and read from the file system.

Technical reference

Entry in moduleDescriptors:	none
Link to schema:	none

Page tree

Versions Compared

Old Version 23

New Version 24

Key

Contents