#893 closed Defect (fixed)
Is it possible to run ORCA05 with KPP closure scheme and MPI?
Reported by: | jharlass | Owned by: | nemo |
---|---|---|---|
Priority: | normal | Milestone: | Unscheduled |
Component: | OCE | Version: | v3.1 |
Severity: | Keywords: | KPP Miscellaneous OPA ORCA05 TKE v3.1 vertical_diffusivity | |
Cc: | jharlass@… |
Description
I want to run ORCA05 (L46) with the KPP closure scheme (instead of TKE), set via the CPP-keys 'key_zdfkpp' and 'key_kpplktb'. The User guide does not help, since in all versions the chapter is incomplete.
There is an explicit query in zdfini.F90 (line 130), if the key 'zdfkpp' is set and if 'lk_mpp' is true (running in MPI). If both conditions are fulfilled, the model produces an error message ("The KPP scheme is not ready to run in MPI"), but continues computing until the zonal velocity gets larger than 20m/s.
Is this only an unimportant old error message or is the KPP scheme really not capable of running in MPI?
Best regards!
Commit History (0)
(No commits)
Change History (8)
comment:1 Changed 12 years ago by rblod
comment:2 Changed 11 years ago by gm
- Resolution set to fixed
- Status changed from new to closed
What Rachid said is exact. The stop put there means that KPP scheme has never been tested in mpp.A priori, it may work properly, but who knows... If someone want to use it on multi processors, he must verify that mono/multi proc runs give the same results (and this will possibly require changes in the code).
comment:3 Changed 8 years ago by nicolasmartin
- Keywords vertical_diffusivity added; vertical diffusivity removed
comment:4 Changed 6 years ago by nemo
- Keywords changed from ORCA05, vertical_diffusivity, KPP, TKE to KPP ORCA05 TKE vertical_diffusivity
comment:5 Changed 6 years ago by nemo
- Keywords Misc. added
comment:6 Changed 6 years ago by nemo
- Keywords Misc. removed
comment:7 Changed 2 years ago by nemo
- Keywords Miscellaneous OPA r3.1 added
comment:8 Changed 2 years ago by nemo
- Keywords v3.1 added; r3.1 removed
Hi Jan
When looking at KPP routine, I don't see particular reason to avoid running it on mpp, so the stop could just be an historical remaining. This being said, I would be very careful in using KPP in NEMO:
Rachid