Welcome, Guest
Username: Password: Remember me

TOPIC: Speed of the mesher: regression ?

Speed of the mesher: regression ? 1 week 5 days ago #45987

  • nicogodet
  • nicogodet's Avatar
  • OFFLINE
  • Expert Boarder
  • Posts: 166
  • Thank you received: 40
Seems similar

Title
Name New Mesh
Type 2D T3 Scalar Mesh
Directory
Filename
NoDataValue -999
DoublePrecision False
Option_ResampleOutline True
Option_GrowThreshold 1.200000
Option_CheckGrowThreshold True
Option_DefaultEdgeLength 35.000000
Option_AutoSmoothMesh True
Option_AutoSmoothLoopCount 400
Option_ResampleLinesOnly False
Option_OverwriteMesh True
Option_NewResampler_Experimental False
Performance_ResampleShorelines 0.000 sec
Performance_ResampleSoftlines 0.000 sec
Query_MaxNodeCount 125250
Performance_SpaceEstimation 10 sec
Performance_AddFixedGeometry 2 sec
Performance_AddOutlineNodes 0.000 sec
Performance_PrepareOutline 0.000 sec
Performance_AddDensityPolygonSeedNodes 0.000 sec
Performance_ComputeInteriorNodes 1.8 min
Performance_GenerateNodes 6 sec
Performance_Triangulation 2 sec
Performance_DoCullSmallElements 0.000 sec
Performance_DoCullElementsOutside 5 sec
Performance_ValidateOuterEdgeElements 3.3 min
Performance_RemoveUnusedNodes 5 sec
Performance_MergeSubMeshes 1.9 min
Performance_FinalizeSubmeshAndIslandEdgeConnectivity 6 sec
Performance_InitNewMeshAttributesAndActivateMesh 0.000 sec
Performance_AutoSmoothMesh 3 sec
Performance_DoMakeAllElementsCounterClockWise 0.000 sec
Performance_Cleanup 0.000 sec
Performance_DoGenerateMesh 7.6 min
The administrator has disabled public write access.

Speed of the mesher: regression ? 1 week 4 days ago #45988

  • a.barton
  • a.barton's Avatar
  • OFFLINE
  • Moderator
  • Posts: 772
  • Thank you received: 188
If I understand correctly... upgrading from BK 3.12.24-beta to BK 3.12.25-beta has improved the speed from 8.0 min. down to 7.6 min. - minus 10s (waiting) = 7.4 min. for the same test case.

And in BK 3.12.25, 44.6% (3.3 min) of the total time (7.4 min.) is taken up by "Performance_ValidateOuterEdgeElements", while 24.3% (1.8 min.) of the total time (7.4 min.) is taken up by Performance_ComputeInteriorNodes (previously 2.5 min.).

Ok. That jives with what I was messing with when changing from 3.12.24 to 3.12.25

Besides 0s for a total time... :woohoo:... for you, what is a reasonable total time for meshing your test case?

Kindly... Alan
The administrator has disabled public write access.

Speed of the mesher: regression ? 1 week 4 days ago #45989

  • nicogodet
  • nicogodet's Avatar
  • OFFLINE
  • Expert Boarder
  • Posts: 166
  • Thank you received: 40
This is very acceptable!

I will do the comparison with BK 3.3.4 and report back
The administrator has disabled public write access.

Speed of the mesher: regression ? 2 days 7 hours ago #46006

  • a.barton
  • a.barton's Avatar
  • OFFLINE
  • Moderator
  • Posts: 772
  • Thank you received: 188
> This is very acceptable!

Nice! Happy to hear that... I'll move on to other things then.

Kindly... Alan

PS Happy New Year 2025!
The administrator has disabled public write access.
Moderators: Serrer, a.barton

The open TELEMAC-MASCARET template for Joomla!2.5, the HTML 4 version.