Thank you for checking! Elena ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Elena Pourmal The HDF Group http://hdfgroup.org 1800 So. Oak St., Suite 203, Champaign IL 61820 217.531.6112 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
On May 19, 2014, at 10:59 AM, Sean Byland <[email protected]> wrote: > Elena, > FWIW, I changed the example you linked to: > dset = H5Dcreate (file, DATASET, H5T_NATIVE_LDOUBLE, space, H5P_DEFAULT, > H5P_DEFAULT, H5P_DEFAULT); > status = H5Dwrite (dset, H5T_NATIVE_INT, H5S_ALL, H5S_ALL, H5P_DEFAULT, > wdata[0]); > > To check the following test failure: > Testing hard int -> long double conversions > *FAILED* > > And visual inspection showed: > [ 0 -1 -2 -3 -4 -5 -6] > [ 0 0 0 0 0 0 0] > [ 0 1 2 3 4 5 6] > [ 0 2 4 6 8 10 12] > > To be the same output as before, so I suspect the library might me fine. > > Sean B. > > From: Elena Pourmal <[email protected]> > Reply-To: HDF Users Discussion List <[email protected]> > Date: Friday, May 16, 2014 9:30 PM > To: HDF Users Discussion List <[email protected]> > Cc: Kimberly Browning <[email protected]> > Subject: Re: [Hdf-forum] make check failures with gcc/4.9.0 and HDF5 1.8.13 > (long double conversions) > > Sean, > > On May 16, 2014, at 10:43 AM, Sean Byland <[email protected]> wrote: > >> Thanks Elena, >> Unfortunately it still fails regardless of the optimization level. Do you >> have any other suggestions ? > > Unfortunately, no. We will need to investigate. Hopefully, this is a problem > with the tests and not with the library. > > One thing you may try is to store some data of each of those types > signed/unsigned char, short, int, long as long double in HDF5 (see > http://www.hdfgroup.org/ftp/HDF5/examples/examples-by-api/hdf5-examples/1_8/C/H5T/h5ex_t_int.c > ) and see if the problem is still there. > > Elena > >> I tried it with g++/gfortran pointing to gcc/4.9.0 and gcc pointing it to >> 4.8.2 (because C++ name mangling and fortran modules are often the >> incompatible elements) and all of the test pass but I'd prefer to use 4.9.0 >> for all of the compilers. >> >> Sean >> >> From: Elena Pourmal <[email protected]> >> Reply-To: HDF Users Discussion List <[email protected]> >> Date: Thursday, May 15, 2014 5:32 PM >> To: HDF Users Discussion List <[email protected]> >> Cc: Kimberly Browning <[email protected]> >> Subject: Re: [Hdf-forum] make check failures with gcc/4.9.0 and HDF5 1.8.13 >> (long double conversions) >> >> Hi Sean, >> >> Thank you for trying with the newly released gcc! I'll add the issue to >> JIRA; we will try to support gcc 4.9.* in the next release. >> >> As a workaround please try to use different level of optimization (edit the >> config/gnu-flags file; search for "Production" string and then edit gcc-4.* >> line). >> >> Elena >> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >> Elena Pourmal The HDF Group http://hdfgroup.org >> 1800 So. Oak St., Suite 203, Champaign IL 61820 >> 217.531.6112 >> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >> >> >> >> >> On May 15, 2014, at 2:45 PM, Sean Byland <[email protected]> wrote: >> >>> When building with hdf5/1.8.13 with the following configure line: >>> ./configure CC=gcc CXX=g++ FC=gfortran --disable-silent-rules >>> --enable-fortran --enable-fortran2003 --enable-static --with-pic >>> --disable-sharedlib-rpath --enable-shared --enable-cxx >>> >>> With the gcc, g++ and gfortran resolving to gcc/4.9 I see the following >>> failures: >>> grep "FAILED" gcc49.check.out >>> Testing hard signed char -> long double conversions >>> *FAILED* >>> Testing hard unsigned char -> long double conversions >>> *FAILED* >>> Testing hard short -> long double conversions >>> *FAILED* >>> Testing hard unsigned short -> long double conversions >>> *FAILED* >>> Testing hard int -> long double conversions >>> *FAILED* >>> Testing hard unsigned int -> long double conversions >>> *FAILED* >>> Testing hard long -> long double conversions >>> *FAILED* >>> >>> Has anyone else observed this ? Is there a workaround. Thanks. >>> >>> Sean B. >>> >>> From: Elena Pourmal <[email protected]> >>> Reply-To: HDF Users Discussion List <[email protected]> >>> Date: Thursday, May 15, 2014 2:35 PM >>> To: HDF Users Discussion List <[email protected]> >>> Subject: Re: [Hdf-forum] make check-install fails with HDF5 1.8.13 >>> >>> Ekin, >>> >>> On May 15, 2014, at 1:42 PM, Ekin Akoglu <[email protected]> wrote: >>> >>>> Dear Elena, >>>> >>>> Changing to the directory hdf5-1.8.13/fortran/examples and then issuing >>>> >>>> $ h5fc -o h5_crtdat h5_crtdat.f90 >>>> >>>> works. However, the check-install complains about not being able to open >>>> h5_crtdat.o to write, which already exists. I checked the permissions of >>>> that file and they are correct: 644, so write is allowed. >>>> >>>> Do you have an idea? >>>> >>> No idea at this point. We couldn't reproduce the problem here on one of >>> our machines with Intel Fortran compiler. >>> >>> Could you please try to run run-fortran-ex.sh in that directory? >>> >>> ./run-fortran-ex.sh >>> >>> Thank you! >>> >>> Elena >>>> Thank you, >>>> >>>> Ekin >>>> >>>> >>>> On 15 May 2014 19:47, Elena Pourmal <[email protected]> wrote: >>>>> Ekin, >>>>> >>>>> It looks like that for some reason Fortran compiler cannot write *.o >>>>> file. Could it be that you ran out of space? Is there quota on your >>>>> system? Was /tmp partition full? >>>>> >>>>> I would try to run make check-install the second time, or go to >>>>> /usr/local/hdf5-1.8.13/share/hdf5_examples/fortran and try to run the >>>>> following command >>>>> >>>>> ../../../bin/h5fc -o h5_crtdat h5_crtdat.f90 >>>>> >>>>> and see if it works. >>>>> >>>>> Elena >>>>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >>>>> Elena Pourmal The HDF Group http://hdfgroup.org >>>>> 1800 So. Oak St., Suite 203, Champaign IL 61820 >>>>> 217.531.6112 >>>>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >>>>> >>>>> >>>>> >>>>> >>>>> On May 15, 2014, at 10:30 AM, Ekin Akoglu <[email protected]> wrote: >>>>> >>>>>> Dear Elena, >>>>>> >>>>>> The OS is 64-bit Debian Wheezy 7.5. The compiler is Intel Fortran >>>>>> Compiler 2013 SP1 Update 3, which is exactly 14.0.3 20140422. My GCC >>>>>> version is 4.7.2. I built and installed HDF5 as follows: >>>>>> >>>>>> $ ./configure --prefix=/usr/local/hdf5-1.8.13 --enable-fortran >>>>>> --enable-fortran2003 >>>>>> $ make >>>>>> $ make check >>>>>> # make install >>>>>> $ make check-install >>>>>> >>>>>> Everything except the "make check-install" step worked flawlessly. >>>>>> >>>>>> Thank you. >>>>>> >>>>>> Kind regards, >>>>>> >>>>>> Ekin >>>>>> >>>>>> On 15 May 2014 17:11, Elena Pourmal <[email protected]> wrote: >>>>>>> Ekin, >>>>>>> >>>>>>> Could you please provide OS and compiler information? It would also >>>>>>> help if you post the outputs of the commands used to build and install >>>>>>> the HDF5 libraries. >>>>>>> >>>>>>> Thank you! >>>>>>> >>>>>>> Elena >>>>>>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >>>>>>> Elena Pourmal The HDF Group http://hdfgroup.org >>>>>>> 1800 So. Oak St., Suite 203, Champaign IL 61820 >>>>>>> 217.531.6112 >>>>>>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> On May 15, 2014, at 9:10 AM, Ekin Akoglu <[email protected]> wrote: >>>>>>> >>>>>>>> Dear all, >>>>>>>> >>>>>>>> I have just installed the HDF5 version 1.8.13 and the check-install >>>>>>>> process failed with: >>>>>>>> >>>>>>>> ============================ >>>>>>>> Testing run-fortran-ex.sh >>>>>>>> ============================ >>>>>>>> >>>>>>>> ################# h5_crtdat ################# >>>>>>>> error: can't open file h5_crtdat.o for write >>>>>>>> compilation aborted for h5_crtdat.f90 (code 1) >>>>>>>> messed up compiling h5_crtdat.f90 >>>>>>>> rm: cannot remove `*.o': No such file or directory >>>>>>>> rm: cannot remove `*.h5': No such file or directory >>>>>>>> >>>>>>>> make[3]: *** [installcheck-local] Error 1 >>>>>>>> make[3]: Leaving directory >>>>>>>> `/home/ekin/Tools/hdf5-1.8.13/fortran/examples' >>>>>>>> make[2]: *** [installcheck-local] Error 1 >>>>>>>> make[2]: Leaving directory `/home/ekin/Tools/hdf5-1.8.13/fortran' >>>>>>>> make[1]: *** [installcheck-recursive] Error 1 >>>>>>>> make[1]: Leaving directory `/home/ekin/Tools/hdf5-1.8.13/fortran' >>>>>>>> make: *** [installcheck-recursive] Error 1 >>>>>>>> >>>>>>>> Regards, >>>>>>>> >>>>>>>> Ekin >>>>>>>> >>>>>>>> On 15 May 2014 15:20, <[email protected]> wrote: >>>>>>>>> >>>>>>>>> >>>>>>>>> To view from a browser see: >>>>>>>>> http://www.hdfgroup.org/newsletters/newsletter140.html >>>>>>>>> >>>>>>>>> -------------------------------------------------------------------------- >>>>>>>>> The HDF Group Home Page: http://www.hdfgroup.org/ >>>>>>>>> Support Services: >>>>>>>>> http://www.hdfgroup.org/services/ >>>>>>>>> -------------------------------------------------------------------------- >>>>>>>>> >>>>>>>>> Newsletter #140 >>>>>>>>> May 15, 2014 >>>>>>>>> >>>>>>>>> CONTENTS >>>>>>>>> . Release of HDF5-1.8.13 >>>>>>>>> >>>>>>>>> - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - >>>>>>>>> - - - >>>>>>>>> >>>>>>>>> Release of HDF5 1.8.13 >>>>>>>>> ====================== >>>>>>>>> >>>>>>>>> The HDF5-1.8.13 release is now available, and can be downloaded from >>>>>>>>> the >>>>>>>>> HDF Group Downloads page: >>>>>>>>> >>>>>>>>> http://www.hdfgroup.org/downloads/ >>>>>>>>> >>>>>>>>> It can also be obtained directly from: >>>>>>>>> >>>>>>>>> http://www.hdfgroup.org/HDF5/release/obtain5.html >>>>>>>>> >>>>>>>>> HDF5-1.8.13 is a minor release, with a few new features and changes. >>>>>>>>> Important changes include: >>>>>>>>> >>>>>>>>> o The H5free_memory (C) api was added to free memory that was >>>>>>>>> allocated >>>>>>>>> by HDF5 API calls. This was added specifically for Windows, but >>>>>>>>> may be >>>>>>>>> useful anywhere that the library uses a different memory manager >>>>>>>>> than >>>>>>>>> the application. >>>>>>>>> >>>>>>>>> o New wrappers were added for C++ (many) and Fortran >>>>>>>>> (h5pset_file_image_f and h5pget_file_image_f). >>>>>>>>> >>>>>>>>> o Mac OS X 10.9 Mavericks is now supported. >>>>>>>>> >>>>>>>>> o Many improvements were added in support of CMake, including the >>>>>>>>> addition of support for CMake 2.8.11. >>>>>>>>> >>>>>>>>> o Improvements were added to the Core VFD (or Memory VFD). >>>>>>>>> >>>>>>>>> o The following changes were made to Parallel HDF5: >>>>>>>>> >>>>>>>>> - The MPI-POSIX VFD was removed. Application developers should >>>>>>>>> always use the MPI-IO VFD. >>>>>>>>> >>>>>>>>> - Parallel HDF5 now supports collective I/O on point selections. >>>>>>>>> >>>>>>>>> o A few changes were added to the tools (h5diff, h5dump, and >>>>>>>>> h5repack). >>>>>>>>> >>>>>>>>> This release contains many other changes that are not listed here. >>>>>>>>> Please be sure to read the Release Notes for a comprehensive list of >>>>>>>>> new features and changes: >>>>>>>>> >>>>>>>>> >>>>>>>>> http://www.hdfgroup.org/ftp/HDF5/current/src/hdf5-1.8.13-RELEASE.txt >>>>>>>>> >>>>>>>>> Changes that affect maintainers of HDF5-dependent applications are >>>>>>>>> listed in "HDF5 Software Changes from Release to Release". See: >>>>>>>>> >>>>>>>>> http://www.hdfgroup.org/HDF5/doc/ADGuide/Changes.html >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> Future Changes to Supported Platforms >>>>>>>>> ------------------------------------- >>>>>>>>> >>>>>>>>> Please be aware that after this release, we will be moving to >>>>>>>>> CMake 3.0, and will add testing for VS 2013. VS 2012 will remain >>>>>>>>> the standard. >>>>>>>>> >>>>>>>>> In addition, we will only support Cygwin built with CMake, and will >>>>>>>>> be dropping VS 2010. >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> ------------------------------------------------------------- >>>>>>>>> For questions regarding these or other HDF issues, contact: >>>>>>>>> >>>>>>>>> [email protected] >>>>>>>>> >>>>>>>>> >>>>>>>>> _______________________________________________ >>>>>>>>> Hdf-forum is for HDF software users discussion. >>>>>>>>> [email protected] >>>>>>>>> http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org >>>>>>>> >>>>>>>> _______________________________________________ >>>>>>>> Hdf-forum is for HDF software users discussion. >>>>>>>> [email protected] >>>>>>>> http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org >>>>>>> >>>>>>> >>>>>>> _______________________________________________ >>>>>>> Hdf-forum is for HDF software users discussion. >>>>>>> [email protected] >>>>>>> http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org >>>>>>> >>>>>> >>>>>> _______________________________________________ >>>>>> Hdf-forum is for HDF software users discussion. >>>>>> [email protected] >>>>>> http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org >>>>> >>>>> >>>>> _______________________________________________ >>>>> Hdf-forum is for HDF software users discussion. >>>>> [email protected] >>>>> http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org >>>>> >>>> >>>> _______________________________________________ >>>> Hdf-forum is for HDF software users discussion. >>>> [email protected] >>>> http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org >>> >>> _______________________________________________ >>> Hdf-forum is for HDF software users discussion. >>> [email protected] >>> http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org >> >> _______________________________________________ >> Hdf-forum is for HDF software users discussion. >> [email protected] >> http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org > > _______________________________________________ > Hdf-forum is for HDF software users discussion. > [email protected] > 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. [email protected] http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org Twitter: https://twitter.com/hdf5
