[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[GEMPAK #WTP-700036]: GDDIAG segmentation fault



Evan,

I believe I have a solution for you, hope you don't mind playing guinea pig, 
there are a few steps to follow, please let me know if you have any questions 
or need further explanation.

Attached is a file gditoh.f

As user gempak"

* cd $GEMPAK/source/gridlib/gd/

* mv the existing file gditoh.f to gditoh_orig.f

* cp the attached gditoh.f to $GEMPAK/source/gridlib/gd/ - you should now have 
the new gditoh.f and the original backup in this directory.

in this same directory type:

"make all"
"ar -r $OS_LIB/gemlib.a gditoh.o"

then

* cd to $GEMPAK/source/programs/gd/gddiag/

type:

"make all"
"make install"

At this point you should attempt to run the GDDIAG script as you would before.  
I have tried this with the script and data files you sent and the segmentation 
fault no longer occurs.  Would like to see if this is the case on your end as 
well.

Michael James
Unidata User Support




> 
> Hi Michael,
> 
> I tried to recompile with g77 with no luck, the segmentation fault still 
> occurred.  If you know of any other programs similar to gdb, I could give 
> those a shot to dig down into what part of the Fortran source code is causing 
> the Segmentation Fault.  Otherwise, I'm ok on this end since the script's 
> working smoothly on a 32 bit version of GEMPAK.
> 
> Regards,
> Evan
> 
> --- On Mon, 5/4/09, Unidata GEMPAK Support <address@hidden> wrote:
> 
> > From: Unidata GEMPAK Support <address@hidden>
> > Subject: [GEMPAK #WTP-700036]: GDDIAG segmentation fault
> > To: address@hidden
> > Cc: address@hidden, address@hidden
> > Date: Monday, May 4, 2009, 3:21 PM
> > Evan,
> >
> > I didn't run into any problems with GDDIAG using the grid
> > file you supplied on my Fedora 9 64bit system (GEMPAK 5.11.4
> > compiled with gfortran).
> >
> > However, I have read about GDDIAG problems before with RH5
> > and gfortran, specifically seg faulting during computations.
> > Not exactly sure why such problems exist but in the past
> > users have found that recompiling with g77 rather than
> > gfortran solved the issue.  Is this possible for you to
> > try?
> >
> > Michael
> >
> >
> >
> > >
> > > Hi Michael,
> > >
> > > Thanks for continuing to follow up on this support
> > request.  I apologize for taking over a week to
> > respond.
> > >
> > > I uploaded the following files to the Unidata RAMADDA
> > Data Repository
> > >
> > > crt_sunshine_fcst_1deg.grd
> > > crt_sunshine_fcst4.csh
> > >
> > > The first GDDIAG GEMPAK command in the csh script
> > results in a Segmentation fault:
> > >
> > > source /data/gempak/Gemenviron
> > >
> > > gddiag << GDDIAG4_EOF
> > > GDFILE=
> > > GDOUTF=
> > > GFUNC=
> > > GDATTIM=
> > > GLEVEL=0
> > > GVCORD=
> > > GRDNAM=
> > > GRDTYP=S
> > > GPACK=none
> > > GRDHDR=
> > > PROJ=
> > > GRDAREA=
> > > KXKY=
> > > MAXGRD=
> > > CPYFIL=
> > > ANLYSS=
> > >
> > GDFILE=/data/weather/input/daily/crt_sunshine_fcst/crt_sunshine_fcst_1deg.grd
> > >
> > GDOUTF=/data/weather/input/daily/crt_sunshine_fcst/crt_sunshine_fcst_1deg.grd
> > > GDATTIM=090503/0000F009
> > > GVCORD=lcly
> > > GFUNC=QUO ( SUB ( cld01, cld01^090503/0000F003 ), 06
> > )
> > > GRDNAM=dtcld^090503/0000F009
> > > r
> > >
> > > I tried to run a gdb stack trace but that didn't
> > work.  Here's the STDOUT from my attempt:
> > >
> > > GEMPAK-GDDIAG>GDOUT=
> > > GEMPAK-GDDIAG>GDOUTF=
> > > GEMPAK-GDDIAG>GFUNC=
> > > GEMPAK-GDDIAG>GDATTIM=
> > > GEMPAK-GDDIAG>GLEVEL=0
> > > GEMPAK-GDDIAG>GVCORD=
> > > GEMPAK-GDDIAG>GRDNAM=
> > > GEMPAK-GDDIAG>GRDTYP=
> > > GEMPAK-GDDIAG>GPACK=none
> > > GEMPAK-GDDIAG>GRDHDR=
> > > GEMPAK-GDDIAG>PROJ=
> > > GEMPAK-GDDIAG>GRDAREA=
> > > GEMPAK-GDDIAG>KXKY=
> > > GEMPAK-GDDIAG>MAXGRD=
> > > GEMPAK-GDDIAG>CPYFIL=
> > > GEMPAK-GDDIAG>ANLYSS=
> > >
> > GEMPAK-GDDIAG>GDFILE=/data/weather/input/daily/crt_sunshine_fcst/crt_sunshine_fcst_1deg.grd
> > >
> > GEMPAK-GDDIAG>GDOUTF=/data/weather/input/daily/crt_sunshine_fcst/crt_sunshine_fcst_1deg.grd
> > > GEMPAK-GDDIAG>GDATTIM=090503/0000F009
> > > GEMPAK-GDDIAG>GVCORD=lcly
> > > GEMPAK-GDDIAG>GFUNC=QUO ( SUB ( cld01,
> > cld01^090503/0000F003 ), 06 )
> > > GEMPAK-GDDIAG>GRDNAM=dtcld^090503/0000F009
> > > GEMPAK-GDDIAG>r
> > > Segmentation fault
> > > [evan@rs1 crt_sunshine_fcst]$ gdb gddiag core GNU gdb
> > Fedora (6.8-27.el5) Copyright (C) 2008 Free Software
> > Foundation, Inc.
> > > License GPLv3+: GNU GPL version 3 or later 
> > > <http://gnu.org/licenses/gpl.html>
> > > This is free software: you are free to change and
> > redistribute it.
> > > There is NO WARRANTY, to the extent permitted by
> > law.  Type "show copying"
> > > and "show warranty" for details.
> > > This GDB was configured as
> > "x86_64-redhat-linux-gnu"...
> > > (no debugging symbols found)
> > > /data/weather/input/daily/crt_sunshine_fcst/core: No
> > such file or directory.
> > > (gdb) q
> > >
> > > Sorry I'm not very familiar with this program. 
> > This same script works fine on my 32-bit server.
> > >
> > > Regards,
> > > Evan Lowery
> > >
> >
> > Ticket Details
> > ===================
> > Ticket ID: WTP-700036
> > Department: Support GEMPAK
> > Priority: Normal
> > Status: Open
> >
> >
> 
> 
> 
> 
> 


Ticket Details
===================
Ticket ID: WTP-700036
Department: Support GEMPAK
Priority: Critical
Status: Open

Attachment: gditoh.f
Description: Binary data