Custom Query (2547 matches)
Results (37 - 39 of 2547)
Ticket | Resolution | Summary | Owner | Reporter | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
#2689 | fixed | latest Fmake_WORK.sh changes assume ln understands the -r option | acc | acc | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Description |
ContextRecent improvements to correctly enable the specification and use of alternative MY_SRC directories may break on some systems. AnalysisThe cause is the use of ln -r to create links relative to the current directory. This option is not understood by all versions of ln. RecommendationIt will be more robust to be set the links explicitly. This works:
and will be submitted later if no one has a better idea. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
#2688 | fixed | restartability issue with ISF at NEMO 4.0.x | davestorkey | davestorkey | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Description |
ContextMy eORCA025 configuration at NEMO 4.0.4 fails restartability and I have traced this to the following lines in sbc_isf_init: !! compute first level of the top boundary layer DO ji = 1, jpi DO jj = 1, jpj ik = 2 !!gm potential bug: use gdepw_0 not _n DO WHILE ( ik <= mbkt(ji,jj) .AND. gdepw_n(ji,jj,ik) < rzisf_tbl(ji,jj) ) ; ik = ik + 1 ; END DO misfkt(ji,jj) = ik-1 END DO END DO (This applies to the parametrised option with closed cavities). For one particular ice shelf the initialisation of misfkt is changed after a restart, thus changing the results. If I follow Gurvan's suggestion in the comment and use gdepw_0 instead of gdepw_n, it fixes the problem. Analysis... RecommendationI see that the equivalent code in the trunk uses gdepw_0, so I propose to make this change in 4.0_HEAD as well. (But I would like to understand the initialisation of gdepw_n a bit better to see exactly what is going wrong - isn't this potentially a problem for other bits of the code??). |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
#2687 | fixed | nesting tools - interpolation of 2D-3D variables | systeam | clem | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Description |
The nesting tools for version 4.0-HEAD is not able to interpolate data with dimensions: (x,y) (x,y,z) It needs a time_counter dimension. It is constraining. RecommendationEventhough this version is outdated, it is still used for 4.0 nemo versions, so I think we should release this constrain on time_counter. I coded it, so ready to be committed. If time is a dimension in the input file, it needs to be the unlimited dimension though. |