[Fwd: Re: [cam-users] compilation error]

[Maisa Rojas Corradi]

a couple of months ago I was tying (unsuccesfully) to compile and run CAM
on our linux cluster. I deceided now to try again and I believe I am a few
steps closer!!!

The model compiles with no complain, but it is nor running properly. Here
are the messages I get:

]% mpimon bld/cam < namelist -- node-11 2
t_setoption: option disabled: Usr Sys
t_setoption: option disabled: Usr Sys
            2 pes participating in computation
 -----------------------------------
 NODE#  NAME
  0  node-11  1  node-11
------------------------------------------------------------
 NCAR Community Atmospheric Model (CAM)
 $Name: cam3_0_brnchT_p1 $
 $Date: 2004/05/20 18:36:01 $
 ------------------------------------------------------------
 (Online documentation is available on the CAM
  home page: http://www.ccsm.ucar.edu/models/atm-cam/
  License information is available as a link from above or from:
  home page: http://www.ccsm.ucar.edu/models/atm-cam/license.html)
 ------------------------------------------------------------
 DATE 10/01/04 TIME 16:38:03
 ------------------------------------------------------------
 DYCORE is EUL
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

and it has been there for the last hour!

First, question: I guess message: t_setoption: option disabled: Usr Sys is
no good. So I did what you suggested a while ago and did:

bld]% nm f_wrappers.o
         U strncpy
         U t_initialize
00000010 T t_initializef_
         U t_pr
00000020 T t_prf_
         U t_reset
00000040 T t_resetf_
         U t_setoption
00000050 T t_setoptionf_
         U t_stamp
00000080 T t_stampf_
         U t_start
000000b0 T t_startf_
         U t_stop
00000100 T t_stopf_

You said that if there was no underscore then the Makefile was doing
thinks not properly. Do you have any suggestion on what to change in my
Makefile??


many thanks,

Maisa






> On Mon, 19 Jul 2004, Maisa Rojas Corradi wrote:
>
>> Hello,
>> I am trying to compile CAM on our linux cluster with portland group
fortran. and I get the following error when trying to link all the **.o
files:
>>
>> atm_lndMod.o(.text+0x4de): undefined reference to `t_startf_'
>
> As Phil Rasch says, this external reference comes from the file
> f_wrappers.o.
> In the directory where the model is built, I'd first ensure that this
file exists.  If it does not exist then it either failed to compile or
"configure"
> failed to build in the need for it.  Make sure Filepath (built by
configure)
> contains the directory ending with utils/timing.
>
> If f_wrappers.o does exist, please do "nm f_wrappers.o".  If this spits
out names
> with zero or two underscores appended (e.g. t_startf or t_startf__),
then it
> is likely that the Makefile constructed by configure set a #define
incorrectly.  If "nm f_wrappers.o" spits out names with a single
underscore
> appended (e.g. t_startf_) this is correct and the next question is
whether the load step is handling the file properly.
>
> Jim Rosinski
>


-- 
Maisa Rojas Corradi
Departamento de Geofisica
Universidad de Chile
Blanco Encalada 2002
Tel: 678 43 00



-- 
Maisa Rojas Corradi
Departamento de Geofisica
Universidad de Chile
Blanco Encalada 2002
Tel: 678 43 00