[Wrf-users] WRF Parallel Benchmark on WRF v2.2

[John Michalakes]

Those error messages are harmless -- they have to do with the fact that the
benchmark datasets are from 2.1 and there's a slight incompatibility but nothing
that would affect results or the timings.  To suppress that message, comment out
the call to the routine wrf_message at around line 364 in the file
external/esmf_time_f90/Meat.F90 .

John

> -----Original Message-----
> From: Jan Ploski [mailto:Jan.Ploski at offis.de]
> Sent: Thursday, July 12, 2007 2:56 AM
> To: wrfhelp
> Cc: wrf-users at ucar.edu
> Subject: Re: [Wrf-users] WRF Parallel Benchmark on WRF v2.2
>
>
> wrf-users-bounces at ucar.edu schrieb am 05/30/2007 03:49:21 AM:
>
> > Please add these two namelist record in the downloaded namelist.input
> > file:
> >
> >   &fdda
> >   /
> >
> >   &grib2
> >   /
> >
> > If it doesn't work, send me the namelist.input file.
>
> Hello,
>
> Please forgive my late reply. I got around to running the benchmark again
> now. Your suggestion solved the problem I originally described (see my
> quoted message below). However, there is still yet another issue. Although
> the model appears to run, some error messages are output:
>
> Timing for main: time 2001-10-25_00:33:36 on domain   1:    8.57300
> elapsed seconds.
>  ERROR timeaddmonths():  MM out of range
> Timing for main: time 2001-10-25_00:34:48 on domain   1:    7.56600
> elapsed seconds.
>  ERROR timeaddmonths():  MM out of range
> Timing for main: time 2001-10-25_00:36:00 on domain   1:    7.49700
> elapsed seconds.
>  ERROR timeaddmonths():  MM out of range
> Timing for main: time 2001-10-25_00:37:12 on domain   1:    7.58600
> elapsed seconds.
>  ERROR timeaddmonths():  MM out of range
> ...
>
> I am attaching my namelist.input file per your request. Thank you!
>
> Best regards,
> Jan Ploski
>
>
>
> >
> > wrfhelp
> > > Message: 1
> > > Date: Mon, 28 May 2007 19:11:51 +0200
> > > From: Jan Ploski <Jan.Ploski at offis.de>
> > > Subject: [Wrf-users] WRF Parallel Benchmark on WRF v2.2
> > > To: wrf-users at ucar.edu
> > > Message-ID: <465B0D57.80600 at offis.de>
> > > Content-Type: text/plain; charset=us-ascii; format=flowed
> > >
> > > Hello,
> > >
> > > I tried running the conus12km_2001 benchmark from
> > > http://www.mmm.ucar.edu/wrf/WG2/bench/ using WRF v2.2 (not using WRF
> > > v2.1.1 as recommeded on the page). I got the following error from
> > > wrf.exe:
> > >
> > >   -------------- FATAL CALLED ---------------
> > >   FATAL CALLED FROM FILE:  module_configure.b  LINE:     136
> > >   module_configure: initial_config: error reading namelist
> > >   -------------------------------------------
> > >
> > > or, with strace:
> > >
> > > ...
> > > lseek(3, 4096, SEEK_SET)                = 4096
> > > read(3, ",    0.1,\n non_hydrostatic      "..., 4096) = 1363
> > > read(3, "", 4096)                       = 0
> > > write(2, " ", 1 )                        = 1
> > > write(2, "-------------- FATAL CALLED ----"..., 43-------------- FATAL
> > > CALLED ---------------) = 43
> > > write(2, "\n", 1
> > > )                       = 1
> > > write(1, " FATAL CALLED FROM FILE:  module"..., 60 FATAL CALLED FROM
> > > FILE:  module_configure.b  LINE:     136
> > > ) = 60
> > > write(2, " ", 1 )                        = 1
> > > write(2, "module_configure: initial_config"..., 56module_configure:
> > > initial_config: error reading namelist) = 56
> > > write(2, "\n", 1
> > > )                       = 1
> > > write(2, " ", 1 )                        = 1
> > > write(2, "--------------------------------"...,
> > > 43-------------------------------------------) = 43
> > > write(2, "\n", 1
> > > )                       = 1
> > > write(2, "FORTRAN STOP\n", 13FORTRAN STOP
> > > )          = 13
> > > close(3)                                = 0
> > >
> > > (Are there better ways of getting useful error messages?)
> > >
> > > Is there a workaround for this problem to make the benchmark run? If
> > > not, are there any "official" benchmarks for WRF v2.2?
>