20040930: netCDF build under Linux using Intel icc & ifort

Riwal

> To: address@hidden
> From: Riwal Plougonven <address@hidden>
> Subject: problem in installing netCDF 
> Organization: University of St Andrews
> Keywords: 200409292201.i8TM1BUE020738 netCDF Linux Intel compilers

The above message contained the following:

> I am trying to install the netCDF libraries on a linux machine, using
> intel compilers (ifort and icc). I encounter an error in the 'make
> test' step.
> 
> I am not sure of the values I give to the flags; I have tried different
> values suggested on the website, but always obtain the same failures in
> the test.
> 
> Hence I have attached and included below the elements that you
> recommend we send for reporting problems.
> 
> Thanking you in advance for your help, 
> 
> Riwal Plougonven
> University of St Andrews
> Scotland
>   
> 
> -------------------------
> 
> uname -a
> 
> Linux storm1 2.4.22-1.2199.nptl #1 Wed Aug 4 12:21:48 EDT 2004 i686 i686 
> i386 GNU/Linux
> 
> VERSION:
> 3.5.1
> 
> which ifort
> /usr/bin/ifort
> 
> which icc
> /usr/bin/icc
> 
> 
> 
> Note on the choice of the compilers: 
> 
> The aim is to be able to use WRF, and it is advised to use the same 
> compilers for both netCDF and WRF; and I will use the intel compilers for
> WRF. 
...

The file "macros.make" contained the following:

    # Preprocessing:
    M4              = m4
    M4FLAGS         = -B10000
    CPP             = icc -E
    CPPFLAGS        = $(INCLUDES) $(DEFINES) -DNDEBUG -DpgiFortran
    FPP             = 
    FPPFLAGS        = 
    CXXCPPFLAGS     = $(CPPFLAGS)


    # Compilation:
    CC              = icc
    CXX             = icc
    FC              = ifort
    F90             = ifort
    CFLAGS          = -O
    CXXFLAGS        = 
    FFLAGS          = -O -mp -static-libcxa
    F90FLAGS        = -O -mp -static-libcxa

The file "test.log" contained the following:

    Making `test' in directory /user/riwal/CDF/netcdf-3.5.1-Inst7/src/nc_test
    ...
    *** Testing nc_put_var_float ... 
            FAILURE at line 2831 of test_put.c: Numeric conversion not 
representable
            FAILURE at line 2878 of test_put.c: Numeric conversion not 
representable
            FAILURE at line 697 of test_put.c: Var value read not that expected
            FAILURE at line 697 of test_put.c: Var value read not that expected
            FAILURE at line 697 of test_put.c: Var value read not that expected
            FAILURE at line 697 of test_put.c: Var value read not that expected
     1194 good comparisons. 
            ### 6 FAILURES TESTING nc_put_var_float! ###

The problem is that the C compiler is not generating code that correctly 
handles floating-point arithmetic.

This problem is solved by having the "-mp" option in the CFLAGS
environment variable prior to executing the configure script.  More 
information is available at

    
http://my.unidata.ucar.edu/cgi-bin/getfile?file=/content/support/help/MailArchives/netcdf/msg02387.html

Please try the following:

    1.  Perform steps 3 through 5 at

            
http://my.unidata.ucar.edu/content/software/netcdf/INSTALL.html#ReportingProblems

    2.  Ensure that the environment variable CFLAGS contains the option
        "-mp" (without the quotes).

    3.  Perform steps 6 through 9 at

            
http://my.unidata.ucar.edu/content/software/netcdf/INSTALL.html#ReportingProblems

Please let me know if this helps.

Regards,
Steve Emmerson

> NOTE: All email exchanges with Unidata User Support are recorded in the
> Unidata inquiry tracking system and then made publicly available
> through the web.  If you do not want to have your interactions made
> available in this way, you must let us know in each email you send to us.