[visit-developers] Exception going to lala land

Miller, Mark C. miller86 at llnl.gov
Thu Jul 30 13:54:02 EDT 2015


Hi Allen,

Yes, that sounds plausible.

We (obviously) cannot have exceptions get thrown *across* collective MPi calls. Depending on circumstances, MPI may abort (a good thing in this case) or hang (worst possible outcome).

We've seen similar situations in other parts of VisIt. My recollection is that Kathleen audited for this issue in another part of VisIt (generic database maybe).

Sounds like you may need a CATCHALL down the PICs code and then some logic to decide if/how to continue executing an approaching collective

Mark


From: Allen Sanderson <allen at sci.utah.edu<mailto:allen at sci.utah.edu>>
Reply-To: "visit-developers at elist.ornl.gov<mailto:visit-developers at elist.ornl.gov>" <visit-developers at elist.ornl.gov<mailto:visit-developers at elist.ornl.gov>>
Date: Thursday, July 30, 2015 10:45 AM
To: "visit-developers at elist.ornl.gov<mailto:visit-developers at elist.ornl.gov>" <visit-developers at elist.ornl.gov<mailto:visit-developers at elist.ornl.gov>>
Subject: [visit-developers] Exception going to lala land

Hello,

In the PICS code I have discovered a valid exception firing off. It gets caught but it never returns. This exception happens on the root processor. And I am suspecting that the other processors get hung up waiting for the root to provide data but it is long gone. So they sit and wait in lala land. Does this explanation seem plausible??

Cheers,

Allen



Exception: (VisItException) /Users/allen/Projects/VisIt/trunk/src/avt/IVP/avtParICAlgorithm.C, line 959: engine_par: avtParICAlgorithm::DoSendICs() Sending ICs to yourself. Error.
catch(VisItException) /Users/allen/Projects/VisIt/trunk/src/avt/Pipeline/AbstractFilters/avtFilter.C:313
Exception occurred in avtLCSFilter
Filter avtMissingDataRemovalFilter releasing its data
Done Updating avtLCSFilter
Executing avtStaggeringFilter
Done executing avtStaggeringFilter
Filter avtLCSFilter releasing its data
Filter avtLCSFilter releasing its data
Done Updating avtStaggeringFilter
Executing avtGhostZoneAndFacelistFilter
Using facelist filter only.
Entered update for avtFacelistFilter
This source should not load balance the data.
Executing avtFacelistFilter
Done executing avtFacelistFilter
Done Updating avtFacelistFilter
Done executing avtGhostZoneAndFacelistFilter
Done Updating avtGhostZoneAndFacelistFilter
Executing avtCondenseDatasetFilter
Done executing avtCondenseDatasetFilter
Filter avtGhostZoneAndFacelistFilter releasing its data
Filter avtGhostZoneFilter releasing its data
A data object sink's current input was fed back into itself, ignoring...
Filter avtFacelistFilter releasing its data
Done Updating avtCondenseDatasetFilter
Executing avtVertexNormalsFilter
Done executing avtVertexNormalsFilter
Filter avtCondenseDatasetFilter releasing its data
Done Updating avtVertexNormalsFilter
Executing avtCompactTreeFilter
Done executing avtCompactTreeFilter
Filter avtVertexNormalsFilter releasing its data
Done Updating avtCompactTreeFilter
Executing avtActualExtentsFilter
Unable to determine data extents -- dataset needs an update
Unable to determine spatial extents -- dataset needs an update
Done executing avtActualExtentsFilter
Filter avtCompactTreeFilter releasing its data
Done Updating avtActualExtentsFilter
Memory after first execution was: size = 39606080, rss = 178434048
Done with iterating Updates on pipeline 2
Engine::GatherData:
  writer->MustMergeParallelStreams()=true
  useCompression=false
  respondWithNull=false
  scalableThreshold=2000000
  currentTotalGlobalCellCount=1
  cellCountMultiplier=1



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://elist.ornl.gov/pipermail/visit-developers/attachments/20150730/b730d46f/attachment.html>


More information about the visit-developers mailing list