Please ignore the previous email. It aims to be sent to another mailing list.
Kent ________________________________ From: Hdf-forum <[email protected]> on behalf of Kent Yang <[email protected]> Sent: Monday, January 06, 2014 1:06 PM To: [email protected] Subject: Re: [Hdf-forum] Error building HDF5 1.8.X with Intel compilers version 14 Hi Ted, I have a quick review about your ESDIS metadata archive slides. Slide 4 still needs to be improved. HDF-EOS5 doesn't add any projection parameters and a few others. HDF-EOS2 also has those information. Perhaps I didn't explain well in my previous email. What I mean is for information in the Grid*, <the Projection: CharacterString> really only tells the projection name such as <Projection: GCTP_GEO>. More information on how to retrieve the latitude and longitude of this projection(projection parameters etc.) is also stored in the StructMetadata. I hope that your footnote can add that piece of information. You may just say "Projection parameters and other relevant info. are also preserved" or something like that. Kent ________________________________ From: Hdf-forum <[email protected]> on behalf of Scot Breitenfeld <[email protected]> Sent: Monday, January 06, 2014 12:16 PM To: [email protected] Subject: Re: [Hdf-forum] Error building HDF5 1.8.X with Intel compilers version 14 We have confirmed there is a bug introduced in the Intel compiler, version 14, which causes the tests to fail with the option --enable-fortran2003. You can see the discussion at: http://software.intel.com/en-us/forums/topic/498698 No time frame was given for this bug being fixed by Intel, hopefully it will be soon... We also fixed an issue (in the trunk) with h5sselect_hyperslab_f when compiling with Intel version 14 and --enable-fortran. All the tests pass (at least on the limited platforms we have Intel 14.0 installed on). This fix will be in the next release. Thanks, Scot ________________________________ From: Hdf-forum <[email protected]> on behalf of Eduardo Jauch <[email protected]> Sent: Wednesday, November 06, 2013 4:33 PM To: [email protected] Subject: Re: [Hdf-forum] Error building HDF5 1.8.X with Intel compilers version 14 Hello Kyle, I'm having the same problem, not only in Scientific Linux 6.4, but also on Fedora 19. First I thought that it was related to the fact that I was using both distributions on virtual machine (Virtual Box). I got to compile everything fine ONCE, after re-installing Scientific Linux, but since then, the same error happens all the time and I was unable to reproduce the "success". I figured out what is the problem, but not the cause of it. First, with some changes in the makefile to create a trace, I get this. ============================ Fortran API: fortranlib_test Test Log ============================ ========================== FORTRAN tests ========================== FORTRANLIB_TEST is linked with HDF5 Library version 1.8 release 11 Mounting test PASSED Reopen test PASSED File open/close test PASSED File free space test PASSED Dataset test PASSED Extendible dataset test PASSED Basic dataspace test PASSED Reference to object test PASSED Reference to dataset region test PASSED forrtl: severe (174): SIGSEGV, segmentation fault occurred Image PC Routine Line Source lt-fortranlib_tes 000000000049BC99 Unknown Unknown Unknown lt-fortranlib_tes 000000000049A610 Unknown Unknown Unknown lt-fortranlib_tes 00000000004703C2 Unknown Unknown Unknown lt-fortranlib_tes 00000000004562D3 Unknown Unknown Unknown lt-fortranlib_tes 000000000042C95B Unknown Unknown Unknown libpthread.so.0 0000003B1C40EFA0 Unknown Unknown Unknown libhdf5_fortran.s 00007FB43829BAAC h5sselect_hypersl 993 H5Sf.c libhdf5_fortran.s 00007FB43828F547 h5s_mp_h5sselect_ 1332 H5Sff.f90 lt-fortranlib_tes 000000000041EC68 test_basic_select 905 tH5Sselect.f90 lt-fortranlib_tes 000000000040D567 MAIN__ 129 fortranlib_test.f90 lt-fortranlib_tes 000000000040D1B6 Unknown Unknown Unknown libc.so.6 0000003B1BC21B75 Unknown Unknown Unknown lt-fortranlib_tes 000000000040D0A9 Unknown Unknown Unknown Than, with some more tests, I found that the soubroutine nh5sselect_hyperslab_c, during the test that causes the SIGSEGV, is being called twice, but shouldn't (at least I think). The second time it is executed (a thread problem?) , the "stride" array is not defined, causing the error. I tried to use the --disable-production and a new error happened, this time when testing the C library: ============================ Testing testhdf5 ============================ testhdf5 Test Log ============================ For help use: /home/user/Work/hdf5-1.8.11/test/.libs/lt-testhdf5 -help Linked with hdf5 version 1.8 release 11 Testing -- Configure definitions (config) Testing -- Encoding/decoding metadata (metadata) Testing -- Checksum algorithm (checksum) Testing -- Ternary Search Trees (tst) Testing -- Memory Heaps (heap) Testing -- Skip Lists (skiplist) Testing -- Reference Counted Strings (refstr) Testing -- Low-Level File I/O (file) Testing -- Generic Object Functions (objects) Testing -- Dataspaces (h5s) Testing -- Dataspace coordinates (coords) Testing -- Shared Object Header Messages (sohm) Testing -- Attributes (attr) Testing -- Selections (select) Testing -- Time Datatypes (time) Testing -- References (reference) Testing -- Variable-Length Datatypes (vltypes) Testing -- Variable-Length Strings (vlstrings) Testing -- Group & Attribute Iteration (iterate) Testing -- Array Datatypes (array) Testing -- Generic Properties (genprop) Testing -- UTF-8 Encoding (unicode) Testing -- User-Created Identifiers (id) Testing -- Miscellaneous (misc) *** UNEXPECTED VALUE from H5Gopen2 should be -1, but is 33554432 at line 4935 in tmisc.c !!! 1 Error(s) were detected !!! So... Seems that something is not working here... I'm assuming that probably is something with the intel fortran 14... For example, --enable-fortran2003 do not works for sure with intel fortran 14... Em 06-11-2013 20:16, kyle kanos escreveu: On my University's cluster, I've built HDF5 1.8.7 using the Intel 11.1 and Intel 12.0 suites without any issue. On my personal computer, running Scientific Linux 6.4, I have Intel 14.0 and cannot built either 1.8.7 or 1.8.11. On both versions, I get the same error under the Fortran tests. I compile it with zlib, and properly exporting the Intel compilers (export CC=icc && export FC=ifort && export FCC=icpc), using $> ./configure --enable-fortran --with-zlib=/usr/local/zlib-1.8.7 I have also tried different permutations involving --enable-cxx, --disable-shared, and removing the --with-zlib flags, but nothing has changed the SIGSEGV error I get: ========================== FORTRAN tests ========================== FORTRANLIB_TEST is linked with HDF5 Library version 1.8 release 11 Mounting test PASSED Reopen test PASSED File open/close test PASSED File free space test PASSED Dataset test PASSED Extendible dataset test PASSED Basic dataspace test PASSED Reference to object test PASSED Reference to dataset region test PASSED forrtl: severe (174): SIGSEGV, segmentation fault occurred Image PC Routine Line Source fortranlib_test 0000000000713379 Unknown Unknown Unknown fortranlib_test 0000000000711CF0 Unknown Unknown Unknown fortranlib_test 00000000006E7AA2 Unknown Unknown Unknown fortranlib_test 00000000006CD9B3 Unknown Unknown Unknown fortranlib_test 00000000006A403B Unknown Unknown Unknown libpthread.so.0 0000003CBDC0F500 Unknown Unknown Unknown fortranlib_test 000000000043676C Unknown Unknown Unknown fortranlib_test 0000000000429D3D Unknown Unknown Unknown fortranlib_test 0000000000417B0B Unknown Unknown Unknown fortranlib_test 0000000000404CEC Unknown Unknown Unknown fortranlib_test 00000000004048D6 Unknown Unknown Unknown libc.so.6 0000003CBD41ECDD Unknown Unknown Unknown fortranlib_test 00000000004047C9 Unknown Unknown Unknown I have zero issues if I install this using the gcc compilers, but my code needs the Intel compilers. Is this an error related to the newer version of the Intel compiler, or is this something that I've done wrong? _______________________________________________ Hdf-forum is for HDF software users discussion. [email protected]<mailto:[email protected]> http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org -- [X]<www.ist.utl.pt> [X] <www.mohid.com> TECNICO Lisboa - MARETEC Sec??o de Ambiente e Energia Av. Rovisco Pais 1049-001 Lisboa Tel.: +351 21 8419424 Fax: +351 21 8417365 URL: www.mohid.com<http://www.mohid.com>
_______________________________________________ Hdf-forum is for HDF software users discussion. [email protected] http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org
