Re: 20041118: netCDF Fortran (77 + 90) - array size limitation

>To: address@hidden
>From: "Paquita Zuidema" <address@hidden>
>Subject: 20041118: netCDF Fortran (77 + 90) - array size limitation
>Organization: CU/CIRES
>Keywords: 200411182221.iAIMLur5029116 netCDF MacOS-X

Paquita,

> Thanks for your detailed reply.
> 
> I'm using an Absoft fortran 90 interface. I get the
> error from calling a netcdf function, it occurs at 
> line 35 in the attached file "read_sonia.f95" - so
> I have a valid file and variable identifier, but, I
> get the message "segmentation violation" within the
> absoft fx debugger, and a "segmentation fault"
> if I run it outside the debugger on the command line.
> 
> I'm also attaching two files, one which read_sonia.f95
> can read successfully, named sonia2_homo.opt, and
> the other that read_sonia.f95 can't - sonia1_homo.opt.
> 
> One obvious thought was that I lacked enough static
> storage - I increased it on the mac, and also compile
> read_sonia with a static flag, to no avail. I've attached
> the makefile ("sonia_Makefile","make.common")

I doubt if that is the problem, but can't be sure, because I don't
have access yet to Absoft Fortran on a Mac OS X platform.  I've
applied for a trial copy, but it may be 48 hours before I get the
necessary key to download and install it.

In the meantime, I think you should be checking the return status of
your netCDF calls, just to make sure nothing is going wrong before the
segmentation violation.  If the returned status from any call is not
equal to nf90_error, you should print out the associated error
message.  For example:

        status=nf90_Inquire_Dimension(ncid,dimid,len=nXedges)
        if (status /= nf90_noerr) print *, trim(nf90_strerror(status))

> I wonder if perhaps I have installed netcdf correctly?
> I have attached the output from "make test" as "test.log".
> I did not see a binary that stated it would work w/
> absoft fortran, hence installed it myself. Note the
> last line of "test.log" - 2 files don't quite compare.
> Is this telling me/us something ?

Yes, the output you sent from "make test" indicates a problem with the
C library, on which the Fortran 90 library depends.  Where you see:

> Making `test' in directory /Users/pz/netcdf-3.5.1/src/libsrc
> 
> ar cru libnetcdf.a attr.o dim.o error.o libvers.o nc.o ncio.o ncx.o putget.o 
> string.o v1hpg.o v2i.o var.o
> ranlib libnetcdf.a
> acc -o t_nc -O  t_nc.o -L. -lnetcdf 
> ./t_nc
> dimrename: IXX
> nc_close ret = 0
> 
> reopen id = 6 for filename test.nc
> NC done
> GATTR VAR VATTR
> VATTR
> VATTR
> VATTR
> VATTR
> VATTR
> fill_seq indices       0 1 0  8.000000 != 
> 9969209968386868600000000000000000000.000000
> indices        0 1 1  9.000000 != 9969209968386868600000000000000000000.000000
> indices        0 1 2  10.000000 != 
> 9969209968386868600000000000000000000.000000
 ...
> indices        7 6 5  445.000000 != 
> 9969209968386868600000000000000000000.000000
> indices        7 6 6  446.000000 != 
> 9969209968386868600000000000000000000.000000
> indices        7 6 7  447.000000 != 
> 9969209968386868600000000000000000000.000000
> Done
> got val = 3.250000
> got val = 0.000000
> got val = 2.718282
> got val = 82555
> got val = 97
> got NC_CHAR val = A (0x41) 
> got NC_CHAR val = B (0x42) 
> got NC_CHAR val = "The red death had long devastated the country."
> got val = A (0x41) 
> got val = B (0x42) 
> got val = "The red death had long devastated the country."
> got vals = 0.000000 ... 9969209968386868600000000000000000000.000000
> re nc_close ret = 0
> cmp test.nc test_nc.sav
> test.nc test_nc.sav differ: char 2113, line 5

you should instead see something like this:

  Making `test' in directory /home/russ/netcdf/build/mort-36b4/src/libsrc

  ar cru libnetcdf.a attr.o dim.o error.o libvers.o nc.o ncio.o ncx.o putget.o 
string.o v1hpg.o v2i.o var.o
  ranlib libnetcdf.a
  cc -o t_nc -O  t_nc.o -L. -lnetcdf 
  ./t_nc
  dimrename: IXX
  nc_close ret = 0

  reopen id = 7 for filename test.nc
  NC done
  GATTR VAR VATTR
  VATTR
  VATTR
  VATTR
  VATTR
  VATTR
  fill_seq indices       1 2 3  75.000000 != 2.718282
  Done
  got val = 3.250000
  got val = 0.000000
  got val = 2.718282
  got val = 82555
  got val = 97
  got NC_CHAR val = A (0x41) 
  got NC_CHAR val = B (0x42) 
  got NC_CHAR val = "The red death had long devastated the country."
  got val = A (0x41) 
  got val = B (0x42) 
  got val = "The red death had long devastated the country."
  got vals = 0.000000 ... 447.000000
  re nc_close ret = 0
  cmp test.nc test_nc.sav
  *** Success ***

> If you tell me you can read "sonia1_homo.opt" just fine
> w/ read_sonia.f95, that is a good clue too that I didn't
> install netcdf quite right.

On a Sun/Solaris 5.8 system, I was able to read "sonia1_homo.opt" just
fine w/ read_sonia.f95 using the Sun f95 compiler and version 3.5.1 of
netCDF.

> Thanks for your time, I really appreciate it.
> 
> If I did the installation incorrectly, may I run what I did
> by you ?

Sure.

--Russ

_____________________________________________________________________

Russ Rew                                         UCAR Unidata Program
address@hidden          http://www.unidata.ucar.edu/staff/russ