domain size specification

27 views
Skip to first unread message

Tiago Silva (Cefas)

unread,
May 28, 2012, 10:35:24 AM5/28/12
to getm-...@googlegroups.com

Hi,

 

I am setting up a parallel run with GETM 2.2 and it is failing here:

 

    init_time

        Time step:         1.50000000000000       seconds

        Time format:               2

        Start:          2007-12-01 00:00:00

        Stop:           2007-12-01 01:00:00

         ==>            0  day(s) and         3600  seconds ==>         2400

  micro time steps

    init_domain test

FATAL ERROR: read_par_setup: Global grid sizes do not match

FATAL ERROR:   Expected           -1  by           -1

FATAL ERROR:   Read             1225  by          538

 

Looking at read_par_setup.F90:

 

   if (iextr_read /= iextr .OR. jextr_read /= jextr) then

      FATAL 'read_par_setup: Global grid sizes do not match'

      FATAL '  Expected ',iextr,' by ',jextr

      FATAL '  Read     ',iextr_read,' by ',jextr_read

      stop

   end if

 

It seems that iextr and jextr  are still -1 as initialised and haven’t been update to the domain size. Can you tell me where this should be read from?

The values in par_setup.dat (iextr_read, jextr_read) are the correct ones.

 

This is the &domain section of the getm.inp:

 

&domain

   vert_cord = 1,

   maxdepth = 55.0,

   bathymetry = 'topo.nc',

   longitude = -2.8,

   latitude = 52.0,

   f_plane = .true.,

   openbdy = .true.,

   bdyinfofile = 'bdyinfo.dat',

   crit_depth = 0.5,

   min_depth = 0.2,

   kdum = 25,

   ddu = -1.0,

   ddl = -1.0,

   d_gamma = 40.0,

   gamma_surf = .true.,

   il = -1,

   ih = -1,

  jl = -1,

   jh = -1,

   z0_method = 0,

   z0_const = 0.005,

/

 

Thanks and I apologise for the corporate disclaimer at the end of this email,

Tiago

 

 

 





This email and any attachments are intended for the named recipient only. Its unauthorised use, distribution, disclosure, storage or copying is not permitted. If you have received it in error, please destroy all copies and notify the sender. In messages of a non-business nature, the views and opinions expressed are the author's own and do not necessarily reflect those of Cefas. Communications on Cefas’ computer systems may be monitored and/or recorded to secure the effective operation of the system and for other lawful purposes.


Karsten Bolding

unread,
May 29, 2012, 1:07:18 AM5/29/12
to getm-...@googlegroups.com
Morning

What is the content of the .dim file?

Karsten
--
http://www.getm.eu

Tiago Silva (Cefas)

unread,
May 29, 2012, 4:42:16 AM5/29/12
to getm-...@googlegroups.com

I forgot to say this is a STATIC compilation

STATIC += -DSPHERICAL -DITERATE_VERT_ADV -DUV_TVD -DNONNEGSALT


The dim file sits in the directory from where I run make.


integer, parameter :: iextr=1225,jextr=538
integer, parameter :: imin=1,imax=96,jmin=1,jmax=48
integer, parameter :: kmax=15


Thanks,
tiago

Karsten Bolding

unread,
May 29, 2012, 4:49:02 AM5/29/12
to getm-...@googlegroups.com
do you have -DSTATIC anywhere?
--
http://www.getm.eu

Knut

unread,
May 29, 2012, 5:19:52 AM5/29/12
to getm-...@googlegroups.com
Hi Tiago,

> do you have -DSTATIC anywhere?
>

if you are using the official setups framework of 2.2 probably not.
In the past -DSTATIC was added by default to STATIC (Rules.make?), but
now you have to specify it manually in your testcase Makefile. Your
error encounts because dynamic memory allocation is not yet possible for
parallel runs.

Knut

Tiago Silva (Cefas)

unread,
May 29, 2012, 6:28:00 AM5/29/12
to getm-...@googlegroups.com
Thanks Karsten and Knut,

I have added
ADDARGS= -DSTATIC

to:
~/getm-setups/git/Rules.make

And it is now setting iextr and jextr and build time.

Thanks,
Tiago

Knut

unread,
May 29, 2012, 6:38:55 AM5/29/12
to getm-...@googlegroups.com
Dear Tiago,

>
> I have added
> ADDARGS= -DSTATIC
>
> to:
> ~/getm-setups/git/Rules.make

although this solves your problem, modifications of "global" settings
are definitely not a nice solution...

BTW:
ADDARGS was thought to be used in a cmdline way as "make ADDARGS=...".

Knut

Tiago Silva (Cefas)

unread,
May 29, 2012, 6:53:38 AM5/29/12
to getm-...@googlegroups.com
OK, so would you put it in Myrun/Makefile?

STATIC += -DSTATIC -DSPHERICAL -DITERATE_VERT_ADV -DUV_TVD -DNONNEGSALT




-----Original Message-----
From: getm-...@googlegroups.com [mailto:getm-...@googlegroups.com] On Behalf Of Knut
Sent: 29 May 2012 11:39
To: getm-...@googlegroups.com

Knut

unread,
May 29, 2012, 6:59:13 AM5/29/12
to getm-...@googlegroups.com
Tiago Silva (Cefas) wrote:
> OK, so would you put it in Myrun/Makefile?
>
> STATIC += -DSTATIC -DSPHERICAL -DITERATE_VERT_ADV -DUV_TVD -DNONNEGSALT

I would do so :-)

Karsten Bolding

unread,
May 29, 2012, 8:04:09 AM5/29/12
to getm-...@googlegroups.com
On Tue, May 29, 2012 at 12:59:13PM +0200, Knut wrote:
> Tiago Silva (Cefas) wrote:
> >OK, so would you put it in Myrun/Makefile?
> >
> >STATIC += -DSTATIC -DSPHERICAL -DITERATE_VERT_ADV -DUV_TVD -DNONNEGSALT

why was -DSTATIC removed from the Rules.make belonging to the test
cases? Did I do by accident?

>
> I would do so :-)
>
>

Karsten

--
http://www.getm.eu

Knut

unread,
May 29, 2012, 8:06:49 AM5/29/12
to getm-...@googlegroups.com
Dear Karsten,
> why was -DSTATIC removed from the Rules.make belonging to the test
> cases? Did I do by accident?
>

commit 03fb0d62b079c3279b6659528ef9abca8163662e
Author: Karsten Bolding <kar...@bolding-burchard.com>
Date: Tue Apr 12 13:49:51 2011 +0200

-DSTATIC moved from Rules.make to individual Makefiles


I think it is ok.

Knut

Karsten Bolding

unread,
May 29, 2012, 8:12:03 AM5/29/12
to getm-...@googlegroups.com
I must confess I completely forgot the good arguments for doing that!!!
>
>
> I think it is ok.
>
> Knut
>

Karsten

--
http://www.getm.eu
Reply all
Reply to author
Forward
0 new messages