Hello,
I wanted to ask the reason behind the current version of NETGEN in salome, since salome 9.9 the version used inside it is 5.3.1 with a modification in 9.12 (5.3.1_with_CA S_7.8) which from this version onward is still the same.
I am just curious about the reason behind staying at this version, where current netgen is quite ahead of it, i could not find from when the 5.3.1 vr dates but 6.2 first commit is from 24. I imagine that there is a performance gain between the newer versions and the current one used. The only information about this i have found is in this post where it was mentioned the going back to 9.9, and that that year was going to be continued the porting of newer versions. It also mentionned the major difference of multithreading of netgen by itself, instead of the current implementation of salome where we find ourself into partitioning the solid into different sections with internal faces that will be ‘fixed’ by the partitioning.
is there any particular reason for this? I can imagine:
- missing of budget to dedicate to the updating of the code to newer version on this specific regard?
- uncompatibility of new version with salome due to the OCCT dependency? (i saw that lastest version they are compatible with 8.0 release)
- bugs/undesired performance drops in specific points between new vs old version?
- GMSH prioritization development over NETGEN?