compiling hdf5 on OS/X

Hi,

I am having troubles compiling hdf5-1.8.13 (with --enable-fortran) on OS X 10.9. The configure lines are:

    setenv CC gcc
    setenv F77 gfortran
    setenv CXX g++
    ./configure --with-szlib=/opt/local/szip/$F77
    --prefix=/opt/local/hdf5/$F77 --enable-fortran --enable-cxx

The error reported is:

    ./H5test_kind > H5fortran_detect.f90
       FC H5fortran_detect.o
    H5fortran_detect.f90:164.17:

         REAL(KIND= 16) :: a
                      1
    Error: Kind 16 not supported for type REAL at (1)
    H5fortran_detect.f90:167.21:

         a_size = SIZEOF(a)
                          1
    Error: Symbol 'a' at (1) has no IMPLICIT type
    H5fortran_detect.f90:176.27:

It gives the same error with either gfortran (4.3.6) OR ifort (13.0.2).

I found some references to this error when googling, but no resolution - any ideas?

--- Dave
Saint Mary's University

Dave,

You will need to set up the FC variable instead of F77 and use gfortran 4.8.2 or later.

Elena

···

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Elena Pourmal The HDF Group http://hdfgroup.org
1800 So. Oak St., Suite 203, Champaign IL 61820
217.531.6112
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

On Aug 28, 2014, at 10:56 AM, Dave Lane <dlane@ap.stmarys.ca<mailto:dlane@ap.stmarys.ca>> wrote:

Hi,

I am having troubles compiling hdf5-1.8.13 (with --enable-fortran) on OS X 10.9. The configure lines are:
setenv CC gcc
setenv F77 gfortran
setenv CXX g++
./configure --with-szlib=/opt/local/szip/$F77 --prefix=/opt/local/hdf5/$F77 --enable-fortran --enable-cxx
The error reported is:
./H5test_kind > H5fortran_detect.f90
  FC H5fortran_detect.o
H5fortran_detect.f90:164.17:

    REAL(KIND= 16) :: a
                 1
Error: Kind 16 not supported for type REAL at (1)
H5fortran_detect.f90:167.21:

    a_size = SIZEOF(a)
                     1
Error: Symbol 'a' at (1) has no IMPLICIT type
H5fortran_detect.f90:176.27:
It gives the same error with either gfortran (4.3.6) OR ifort (13.0.2).

I found some references to this error when googling, but no resolution - any ideas?

--- Dave
Saint Mary's University
_______________________________________________
Hdf-forum is for HDF software users discussion.
Hdf-forum@lists.hdfgroup.org<mailto:Hdf-forum@lists.hdfgroup.org>
http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org
Twitter: https://twitter.com/hdf5

My bug report to gcc.gnu deals with this issue:

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48404

In summary, there is a mismatch of the LD_LIBRARY_PATH and the gcc version used to compile the code on your OS.

The issue with the intel compiler is new, what is the output from the attached program using the intel compiler. Also use FC instead of F77 to set the fortran compiler.

Scot

testkind.f90 (497 Bytes)

ATT00001.htm (2.28 KB)

Scot,

I was able to build it in gfortran after installing gcc 4.9. Regarding ifort: below is the output from the program that you sent:

skolio:~ dlane$ ifort -v
ifort version 13.0.2
skolio:~ dlane$ ifort testkind.f90 -o testkind
skolio:~ dlane$ ./testkind
            4 1
            4 2
            4 3
            4 4
            4 5
            4 6
            8 7
            8 8
            8 9
            8 10
            8 11
            8 12
            8 13
            8 14
            8 15
           16 16
           16 17
           16 18
           16 19
           16 20
           16 21
           16 22
           16 23
           16 24
           16 25
           16 26
           16 27
           16 28
           16 29
           16 30
           16 31
           16 32
           16 33
           -1 34
            4 8 16 0 0 0
            0 0 0 0
skolio:~ dlane$

--- Dave

···

On 28/08/2014 1:49 PM, Scot Breitenfeld wrote:

My bug report to gcc.gnu deals with this issue:

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48404

In summary, there is a mismatch of the LD_LIBRARY_PATH and the gcc version used to compile the code on your OS.

The issue with the intel compiler is new, what is the output from the attached program using the intel compiler. Also use FC instead of F77 to set the fortran compiler.

Scot

_______________________________________________
Hdf-forum is for HDF software users discussion.
Hdf-forum@lists.hdfgroup.org
http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org
Twitter: https://twitter.com/hdf5

Hi Elena,

Thanks - I updated to 4.9.1 and was able to build it under gfortran. I also set the FC variable although it found the gfortran compiler with F77.

--- Dave

···

On 28/08/2014 1:42 PM, Elena Pourmal wrote:

Dave,

You will need to set up the FC variable instead of F77 and use gfortran 4.8.2 or later.

Elena
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Elena Pourmal The HDF Group http://hdfgroup.org
1800 So. Oak St., Suite 203, Champaign IL 61820
217.531.6112
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

On Aug 28, 2014, at 10:56 AM, Dave Lane <dlane@ap.stmarys.ca > <mailto:dlane@ap.stmarys.ca>> wrote:

Hi,

I am having troubles compiling hdf5-1.8.13 (with --enable-fortran) on OS X 10.9. The configure lines are:

    setenv CC gcc
    setenv F77 gfortran
    setenv CXX g++
    ./configure --with-szlib=/opt/local/szip/$F77
    --prefix=/opt/local/hdf5/$F77 --enable-fortran --enable-cxx

The error reported is:

    ./H5test_kind > H5fortran_detect.f90
      FC H5fortran_detect.o
    H5fortran_detect.f90:164.17:

        REAL(KIND= 16) :: a
                     1
    Error: Kind 16 not supported for type REAL at (1)
    H5fortran_detect.f90:167.21:

        a_size = SIZEOF(a)
                         1
    Error: Symbol 'a' at (1) has no IMPLICIT type
    H5fortran_detect.f90:176.27:

It gives the same error with either gfortran (4.3.6) OR ifort (13.0.2).

I found some references to this error when googling, but no resolution - any ideas?

--- Dave
Saint Mary's University
_______________________________________________
Hdf-forum is for HDF software users discussion.
Hdf-forum@lists.hdfgroup.org <mailto:Hdf-forum@lists.hdfgroup.org>
http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org
Twitter: https://twitter.com/hdf5

_______________________________________________
Hdf-forum is for HDF software users discussion.
Hdf-forum@lists.hdfgroup.org
http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org
Twitter: https://twitter.com/hdf5

I’m not sure then why it’s having problems with the intel compiler since the compiler does support KIND=16. You should have no problems compiling fortran/src/H5fortran_detect.f90, which you can check: ifort H5fortran_detect.f90

Scot

···

On Aug 28, 2014, at 2:01 PM, Dave Lane <dlane@ap.stmarys.ca<mailto:dlane@ap.stmarys.ca>> wrote:

Scot,

I was able to build it in gfortran after installing gcc 4.9. Regarding ifort: below is the output from the program that you sent:

skolio:~ dlane$ ifort -v
ifort version 13.0.2
skolio:~ dlane$ ifort testkind.f90 -o testkind
skolio:~ dlane$ ./testkind
           4 1
           4 2
           4 3
           4 4
           4 5
           4 6
           8 7
           8 8
           8 9
           8 10
           8 11
           8 12
           8 13
           8 14
           8 15
          16 16
          16 17
          16 18
          16 19
          16 20
          16 21
          16 22
          16 23
          16 24
          16 25
          16 26
          16 27
          16 28
          16 29
          16 30
          16 31
          16 32
          16 33
          -1 34
           4 8 16 0 0 0
           0 0 0 0
skolio:~ dlane$

--- Dave

On 28/08/2014 1:49 PM, Scot Breitenfeld wrote:
My bug report to gcc.gnu deals with this issue:

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48404

In summary, there is a mismatch of the LD_LIBRARY_PATH and the gcc version used to compile the code on your OS.

The issue with the intel compiler is new, what is the output from the attached program using the intel compiler. Also use FC instead of F77 to set the fortran compiler.

Scot

_______________________________________________
Hdf-forum is for HDF software users discussion.
Hdf-forum@lists.hdfgroup.org<mailto:Hdf-forum@lists.hdfgroup.org>
http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org
Twitter: https://twitter.com/hdf5

_______________________________________________
Hdf-forum is for HDF software users discussion.
Hdf-forum@lists.hdfgroup.org<mailto:Hdf-forum@lists.hdfgroup.org>
http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org
Twitter: https://twitter.com/hdf5

I'll give this a try and get back to you. It may be a few days as the professor who's mac I'm installing hdf5 on just unplugged it from the network and took it home!

---- Dave

···

On 28/08/2014 4:44 PM, Scot Breitenfeld wrote:

I'm not sure then why it's having problems with the intel compiler since the compiler does support KIND=16. You should have no problems compiling fortran/src/H5fortran_detect.f90, which you can check: ifort H5fortran_detect.f90

Scot

On Aug 28, 2014, at 2:01 PM, Dave Lane <dlane@ap.stmarys.ca > <mailto:dlane@ap.stmarys.ca>> wrote:

Scot,

I was able to build it in gfortran after installing gcc 4.9. Regarding ifort: below is the output from the program that you sent:

skolio:~ dlane$ ifort -v
ifort version 13.0.2
skolio:~ dlane$ ifort testkind.f90 -o testkind
skolio:~ dlane$ ./testkind
           4 1
           4 2
           4 3
           4 4
           4 5
           4 6
           8 7
           8 8
           8 9
           8 10
           8 11
           8 12
           8 13
           8 14
           8 15
          16 16
          16 17
          16 18
          16 19
          16 20
          16 21
          16 22
          16 23
          16 24
          16 25
          16 26
          16 27
          16 28
          16 29
          16 30
          16 31
          16 32
          16 33
          -1 34
           4 8 16 0 0 0
           0 0 0 0
skolio:~ dlane$

--- Dave

On 28/08/2014 1:49 PM, Scot Breitenfeld wrote:

My bug report to gcc.gnu deals with this issue:

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48404

In summary, there is a mismatch of the LD_LIBRARY_PATH and the gcc version used to compile the code on your OS.

The issue with the intel compiler is new, what is the output from the attached program using the intel compiler. Also use FC instead of F77 to set the fortran compiler.

Scot

_______________________________________________
Hdf-forum is for HDF software users discussion.
Hdf-forum@lists.hdfgroup.org
http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org
Twitter:https://twitter.com/hdf5

_______________________________________________
Hdf-forum is for HDF software users discussion.
Hdf-forum@lists.hdfgroup.org <mailto:Hdf-forum@lists.hdfgroup.org>
http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org
Twitter: https://twitter.com/hdf5

_______________________________________________
Hdf-forum is for HDF software users discussion.
Hdf-forum@lists.hdfgroup.org
http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org
Twitter: https://twitter.com/hdf5

It now compiles with ifort, provided the c compiler is the version 4.9 rather than the default one installed by mac ports. Thanks --- Dave

···

On 28/08/2014 4:44 PM, Scot Breitenfeld wrote:

I'm not sure then why it's having problems with the intel compiler since the compiler does support KIND=16. You should have no problems compiling fortran/src/H5fortran_detect.f90, which you can check: ifort H5fortran_detect.f90

Scot

On Aug 28, 2014, at 2:01 PM, Dave Lane <dlane@ap.stmarys.ca > <mailto:dlane@ap.stmarys.ca>> wrote:

Scot,

I was able to build it in gfortran after installing gcc 4.9. Regarding ifort: below is the output from the program that you sent:

skolio:~ dlane$ ifort -v
ifort version 13.0.2
skolio:~ dlane$ ifort testkind.f90 -o testkind
skolio:~ dlane$ ./testkind
           4 1
           4 2
           4 3
           4 4
           4 5
           4 6
           8 7
           8 8
           8 9
           8 10
           8 11
           8 12
           8 13
           8 14
           8 15
          16 16
          16 17
          16 18
          16 19
          16 20
          16 21
          16 22
          16 23
          16 24
          16 25
          16 26
          16 27
          16 28
          16 29
          16 30
          16 31
          16 32
          16 33
          -1 34
           4 8 16 0 0 0
           0 0 0 0
skolio:~ dlane$

--- Dave

On 28/08/2014 1:49 PM, Scot Breitenfeld wrote:

My bug report to gcc.gnu deals with this issue:

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48404

In summary, there is a mismatch of the LD_LIBRARY_PATH and the gcc version used to compile the code on your OS.

The issue with the intel compiler is new, what is the output from the attached program using the intel compiler. Also use FC instead of F77 to set the fortran compiler.

Scot

_______________________________________________
Hdf-forum is for HDF software users discussion.
Hdf-forum@lists.hdfgroup.org
http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org
Twitter:https://twitter.com/hdf5

_______________________________________________
Hdf-forum is for HDF software users discussion.
Hdf-forum@lists.hdfgroup.org <mailto:Hdf-forum@lists.hdfgroup.org>
http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org
Twitter: https://twitter.com/hdf5

_______________________________________________
Hdf-forum is for HDF software users discussion.
Hdf-forum@lists.hdfgroup.org
http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org
Twitter: https://twitter.com/hdf5