v25-rc1 testing and feedback

[rweardley]

Hi, I've made a start on updating some cases to v25-rc1 to test out the differences - in general it was pretty smooth and the v25 syntax was easy to pick up, but I did notice a few things that didn't work as expected or that could be clarified. The examples I was working with are here https://github.com/kyleadm/nekrs_cardinal_workshop_material/tree/main/nekrs for v23, or https://github.com/kyleadm/nekrs_cardinal_workshop_material/tree/v25_nekrs/nekrs for the v25 versions.

• nekrs --help par doesn't show the change in block names e.g. [PRESSURE] -> [FLUID PRESSURE] and [VELOCITY] -> [FLUID VELOCITY]
• checkpointInterval = 0 doesn't output a file at the end (used to, at least in v23).
• RANS k-tau only outputs the S01 field (k), not S02 (tau), when writing files (in v23 the S02 would also be written to file).
• Mesh interpolation when restarting from a case with a different .re2 mesh based on the same geometry isn't working (worked in v23). If +int isn't specified, it reads the solution from the different mesh and interpolates within elements anyway, and when +int is specified to set pointwise interpolation this doesn't work (see below). This case worked in v23, and nekrs --help par still indicates +int should work:

# in .par:
startFrom = "coarsePipe0.f00000+int"            # uncomment to restart from N=1 solution with different .re2 mesh
[MESH]
file = pipe2.re2

# in output log:
setting IC ... 
restart options: int  
reading checkpoint ...
 fileName: coarsePipe0.f00000
 user variables: time mesh pressure velocity 
Aborting in /home/rupert/NekRS/v25-pre/source/src/core/iofld/iofldNek.cpp:L147
read:
read attribute interpolate not supported!

--------------------------------------------------------------------------
MPI_ABORT was invoked on rank 0 in communicator MPI_COMM_WORLD
with errorcode 1.

[stgeke]

Thank you for your feedback. I’ve addressed the reported issues, and the fixes will be included in the upcoming release.