[Mitgcm-support] Re: bug in genmake?

mitgcm-support at dev.mitgcm.org mitgcm-support at dev.mitgcm.org
Wed Jul 9 15:54:52 EDT 2003


Hi,

as you can see from my original email, Patrick's fix was already my 
suspicion. I will check this in sometime soon together with our new 
case. For this I have few questions. The SunFire mpi-compiler is called 
mpf77 and there are also specific paths, e.g. for the mpi-includes. 
These will go in the case SunOS+sunmpi (or maybe 
SunFire+mpi,SunOS+CT+mpi?) Or should I directly edit the SunOS+mpi, 
since the paths are standard for any Sun MPI system (according to 
Stephan, I hope I didn't misunderstand anything). Or is this all to 
specific and not interesting for the "general public" and you would 
rather not have me do this at all?

Where do I include the compile (mpf77) so that I don't have specify it 
at the command line (-fc=mpf77)? When I do it within the case SunOS+mpi, 
Alistair's $likelysuspects systems further up in the code still tells me 
that it uses f77 (which it does not in the end, but it's confusing)

So far I have this:
   case SunOS+mpi:
     set FC         = ( 'mpf77' )
     set LINK       = ( 'mpf77' )
     set LN         = ( '/usr/bin/ln -s' )
     set CPP        = ( '/usr/ccs/lib/cpp -P' )
     set DEFINES    = ( ${DEFINES} '-DTARGET_SUN -DWORDLENGTH=4' )
     set INCLUDES   = ( '-I/opt/SUNWhpc/include' )
     set LIBS       = ( '-lmpi -lthread -lsocket -lnsl' )
     set FFLAGS     = ( '-stackvar -explicitpar -vpara -e -u -noautopar 
-xtypemap=real:64,double:64,integer:32 -fsimple=0' )
     set FOPTIM     = ( '-dalign -O3' )
     set NOOPTFLAGS = ( '-dalign -O0' )
     set NOOPTFILES = ( 'barrier.F different_multiple.F 
external_fields_load.F ini_vertical_grid.F ini_spherical_polar_grid.F 
ini_cori.F mon_printstats_rl.F mon_printstats_rs.F')
     breaksw

Martin

Chris Hill wrote:
> Hi martin
> 
>  patricks fix is right. However, as you note, socket and nsl may not be
> needed on your sunfire. If stephan has an example set of options for
> compile and link of any mpi programs, the option and pathnames that he
> has may help.
> 
> chris
> 
> 
> 
> 
>>-----Original Message-----
>>From: Patrick Heimbach [mailto:heimbach at MIT.EDU] 
>>Sent: Wednesday, February 26, 2003 6:48 AM
>>To: Martin Losch
>>Cc: support at mitgcm.org
>>Subject: Re: bug in genmake?
>>
>>
>>Look at the line
>>    set LIBS   = ...    \
>>It's incomplete and should have the remaining piece further down
>>    '-lsocket -lnsl' )
>>Might be that the latter are not even needed.
>>Maybe try withouth?
>>
>>-p.
>>
>>Martin Losch wrote:
>>
>>>Hi there,
>>>
>>>I tried to compile the code for the -platform=SunOS+mpi. 
>>
>>The genmake 
>>
>>>complains with "too many )'s". This is the piece of script 
>>
>>in question 
>>
>>>(sorry for the autowrapping by my mailtool, as usual I have 
>>
>>no control):
>>
>>>case SunOS+mpi:
>>>  set LN         = ( '/usr/bin/ln -s' )
>>>  set CPP        = ( '/usr/ccs/lib/cpp -P' )
>>>  set DEFINES    = ( ${DEFINES} '-DTARGET_SUN 
>>
>>-DWORDLENGTH=4 -D_d=E' )
>>
>>>  set INCLUDES   = ( '-I/usr/local/mpi/include' )
>>>  set LIBS       = ( '-L/usr/local/mpi/lib/solaris/ch_shmem -lmpi 
>>>-lthread' \
>>>  set FFLAGS     = ( '-stackvar -explicitpar -vpara -e -u 
>>
>>-noautopar 
>>
>>>-xtypemap=real:64,double:64,integer:32 -fsimple=0' )
>>>  set FOPTIM     = ( '-dalign -O3 -xarch=v9' )
>>>  set NOOPTFLAGS = ( '-dalign -O0 -xarch=v9' )
>>>                       '-lsocket -lnsl' )
>>>  set NOOPTFILES = ( 'barrier.F different_multiple.F 
>>>external_fields_load.F ini_vertical_grid.F 
>>
>>ini_spherical_polar_grid.F 
>>
>>>ini_cori.F mon_printstats_rl.F mon_printstats_rs.F')
>>>    breaksw
>>>
>>>I assume that the '-lsocket -lnsl' ) line is out of place 
>>
>>and (maybe) 
>>
>>>belongs to further up after -lthread' \
>>>but I dare not change it myself and check it in before 
>>
>>someone with the 
>>
>>>proper compentence (Chris?) has given his OK. I also plan 
>>
>>to develop a 
>>
>>>case which is specific to our SunFire architecture here at AWI, if 
>>>that's of interest for you.
>>>
>>>Martin
>>>
>>
>>
>>-- 
>>_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
>>Patrick Heimbach     Massachusetts Institute of Technology
>>FON: +1/617/253-5259                    EAPS, Room 54-1518
>>FAX: +1/617/253-4464               77 Massachusetts Avenue
>>mailto:heimbach at mit.edu                 Cambridge MA 02139
>>http://www.mit.edu/~heimbach/                       U.S.A.
>>
> 
> 


-- 
Martin Losch
Alfred Wegener Institute for Polar and Marine Research
Postfach 120161, 27515 Bremerhaven, Germany;
Tel./Fax: ++49(0471)4831-1872/1797





More information about the MITgcm-support mailing list