[MITgcm-support] mitgcm with cheapaml and mnc packages

[Ed Hill]

Hi Nicolas,

On Wed, 2005-09-14 at 11:34 -0400, Nicolas Wienders wrote:

> I am trying to get rid of some probleme we have while trying to
> compile
> the model in a configuration including the mnc and cheapaml packages.
> 
> 
> 1- just for information, what does mean
>    === End of Compilation 1 ===
>    === End of Compilation 2 ===
>    === End of Compilation 3 ===
> What are these steps corresponding to?

In some of our Fortran files there are multiple subroutines and/or
functions defined.  Its likely that the compiler is reporting the above
after compiling each subroutine, function, etc.


> 2- Going back to the checkpoint57o_post we re were able to compile
> and run mitgcm with the cheapaml package turned on.
> Problems are coming when we add the mnc package:
> 
> 
>     a- checkpoint.for will not compile and the error message is the
> following:
> 
> 
> ocnpc84:~/MITgcm/verification/climode/build mitgcm$ make -j2
> /opt/ibmcmp/xlf/8.1/bin/f77 -qfixed=132 -qsuffix=f=for -O3 -c
> checkpoint.for
> cat mnc_var.F | ../../../tools/set64bitConst.sh |  /usr/bin/cpp
> -traditional -P -xassembler-with-cpp -DWORDLENGTH=4
> -DNML_TERMINATOR=" /" -DHAVE_CLOC -DHAVE_STAT -DIGNORE_TIME
> -I/usr/local/netcdf/include > mnc_var.for
> ** set_write_global_pickup   === End of Compilation 1 ===
> cat MNC_CW_READWRITE_I.F | ../../../tools/set64bitConst.sh
> |  /usr/bin/cpp -traditional -P -xassembler-with-cpp -DWORDLENGTH=4
> -DNML_TERMINATOR=" /" -DHAVE_CLOC -DHAVE_STAT -DIGNORE_TIME
> -I/usr/local/netcdf/include > MNC_CW_READWRITE_I.for
> ** read_checkpoint   === End of Compilation 2 ===
> ** write_checkpoint   === End of Compilation 3 ===
> cat MNC_CW_READWRITE_RL.F | ../../../tools/set64bitConst.sh
> |  /usr/bin/cpp -traditional -P -xassembler-with-cpp -DWORDLENGTH=4
> -DNML_TERMINATOR=" /" -DHAVE_CLOC -DHAVE_STAT -DIGNORE_TIME
> -I/usr/local/netcdf/include > MNC_CW_READWRITE_RL.for
> "checkpoint.for", 1516-036 (S) Entity tair has undefined type.
> ** write_checkpoint_now   === End of Compilation 4 ===
> 1501-511  Compilation failed for file checkpoint.for.
> make: *** [checkpoint.o] Error 1
> make: *** Waiting for unfinished jobs....
> ocnpc84:~/MITgcm/verification/climode/build mitgcm$ ls MNC*
> MNC_CW_READWRITE_I.F    MNC_CW_READWRITE_I.for  MNC_CW_READWRITE_RL.F
> MNC_CW_READWRITE_RL.for MNC_CW_READWRITE_RS.F   MNC_OPTIONS.h
> MNC_PARAMS.h            MNC_SIZE.h
> ocnpc84:~/MITgcm/verification/climode/build mitgcm$

The above is somewhat difficult to follow since you specified the "-j2"
option and the errors from compilation of one file are being interleaved
between output from compilation of another file (remember, the builds
happening in parallel when you use "-j2").

But the error seems to be an undefined variable with the name "tair" in
the file "checkpoint.for".  As far as I can tell, this is a custom
modification that you folks made because there is no "tair" variable in
the standard version of our "checkpoint.F" file.


>     b - the different calls to the mnc_cw subroutines are present but
> not the subroutines
>          themselves in the ".for" files.

Sorry, I don't understand the above.


>     c- trying to compile mnc_var.for manually i get: 

> "mnc_var.for", line 422.17: 1514-007 (S) Array tr which has inherited
> or object time dimensions is not permitted to be an element of a
> COMMON block.

[*snip*]

> ** mnc_var_write_any   === End of Compilation 17 ===
> 1501-511  Compilation failed for file mnc_var.for.

I'm afraid I don't understand whats happening in the above errors.


> Any idea of what might be going on?  Thanks a lot. nico

I don't know whats going on.  I suspect one or more of the following:

  - a "make" problem where the *.template files are not being 
     properly expanded
  - C-preprocessor problem(s)
  - you've made some local changes to your code that have errors
  - you've made some local changes to your code that don't have 
    errors but they are now resulting in conflicts/problems when 
    you "merging" your changes with our recent updates that we've 
    made to the main model

but its hard to say.

Since I know nothing about the "cheapaml" package (this is the first
time I've even heard of it!), I have no idea how to reproduce your
errors.  Could you perhaps put a copy (say, as a tar file) of your code
on a web server somewhere so that we could download it and look at it?
I don't know how else to debug these problems.

Ed

-- 
Edward H. Hill III, PhD
office:  MIT Dept. of EAPS;  Rm 54-1424;  77 Massachusetts Ave.
             Cambridge, MA 02139-4307
emails:  eh3 at mit.edu                ed at eh3.com
URLs:    http://web.mit.edu/eh3/    http://eh3.com/
phone:   617-253-0098
fax:     617-253-4464