ENH: GAMGAgglomeration: optional updateInterval for moving mesh cases.

E.g. "pcorr.*" { solver GAMG; tolerance 1e-2; relTol 0; smoother GaussSeidel; cacheAgglomeration yes; updateInterval -1; }

assigned to @Mattijs

assigned to @andy and unassigned @Mattijs

added 267 commits

• 1f008851...cc580fc1 - 263 commits from branch develop
• 427e7e12 - ENH: GAMGAgglomeration: optional updateInterval for moving mesh cases.
• 43444b35 - ENH: GAMGAgglomeration: only re-do individual interfaces
• d1707135 - ENH: GAMGAgglomeration: don't update agglomeration table
• 87719fed - STYLE: whitespace

Compare with previous version

I'll rebase and retest. We left it unmerged since it hardcodes for cyclicA(C)MI. Not too much of a problem - there are plenty of other occurences.

Hey Mattijs, interesting. Shall I test the updateInterval thingy? We didn´t know that there is code for testing?.

Interesting indeed. We're currently looking at whether it can help with cyclicAMI caching...

So let me know if I shall test it on a real case.

• no support for processor agglomeration.
• only tested on mixerVesselAMI2D and only checked if it ran.

added 1 commit

• e36288c8 - ENH: GAMGAgglomeration: use area for weighting

Compare with previous version

added 1 commit

• fd7c3cfd - BUG: GAMGAgglomeration: construct from matrix

Compare with previous version

added 17 commits

• fd7c3cfd...96ccae97 - 11 commits from branch develop
• 6acce358 - ENH: GAMGAgglomeration: optional updateInterval for moving mesh cases.
• 2e332dc4 - ENH: GAMGAgglomeration: only re-do individual interfaces
• 9451d3a1 - ENH: GAMGAgglomeration: don't update agglomeration table
• fe1aa2e2 - STYLE: whitespace
• 62f1f841 - ENH: GAMGAgglomeration: use area for weighting
• 8f08149c - BUG: GAMGAgglomeration: construct from matrix

Compare with previous version

added 1 commit

• 606b8108 - ENH: GAMGAgglomeration: cache agglomeration table on moving mesh cases

Compare with previous version

mentioned in commit 3e00e05e

merged

mentioned in commit 0fa5062c