Re: [Hdf-forum] H5IMPORT test fail
Hi Dave, Please try these: 1. ./configure ... 2. make # build all binary 3. make check# run testings The reasons of the failures you saw is because h5dump is arranged to be built AFTER h5import. When you did make check directly, the h5dump that was needed by the H5IMPORT test was not built yet, resulting in the no such file failure. It is our mistake to arrange h5dump built later than h5import. We will fix it for the next release. Thank you for reporting this failure. -Albert Cheng The HDF Group On 5/27/13 1:28 AM, Dave wrote: Dear hdf5 users I'm trying to build HDF5 library in ubuntu 12.04 64bit LTS using gcc 4.7.3 $ ./configure --prefix=/opt/hdf5 --with-szlib=/opt/hdf5 --with-zlib=/opt/hdf5 \ $ make check ('szlib' and 'zlib' are already installed in /opt/hdf5) And I got 8 fails during running 'make check' as below. I've installed hdf5 tools from ubuntu software center and checked 'h5dump' worked fine but test didn't make it. It seems that there is no posts to refer for the problem in this forum. Please help this newbie. Thank you in advance. == H5IMPORT tests started == Testing ASCII I32 rank 3 - Output BE h5importtestutil.sh: line 165: ../h5dump: Is a directory h5importtestutil.sh: line 168: ../h5dump: No such file or directory PASSED Testing ASCII I16 rank 3 - Output LE - CHUNKED - extended h5importtestutil.sh: line 165: ../h5dump: Is a directory h5importtestutil.sh: line 168: ../h5dump: No such file or directory PASSED Testing ASCII I8 - rank 3 - Output I8 LE-Chunked+Extended+Compressed h5importtestutil.sh: line 165: ../h5dump: Is a directory h5importtestutil.sh: line 168: ../h5dump: No such file or directory PASSED Testing ASCII UI16 - rank 2 - Output LE+Chunked+Compressed h5importtestutil.sh: line 165: ../h5dump: Is a directory h5importtestutil.sh: line 168: ../h5dump: No such file or directory PASSED Testing ASCII UI32 - rank 3 - Output BE h5importtestutil.sh: line 165: ../h5dump: Is a directory h5importtestutil.sh: line 168: ../h5dump: No such file or directory PASSED Testing ASCII F32 - rank 3 - Output LE h5importtestutil.sh: line 165: ../h5dump: Is a directory h5importtestutil.sh: line 168: ../h5dump: No such file or directory PASSED Testing ASCII F64 - rank 3 - Output BE + CHUNKED+Extended+Compressed h5importtestutil.sh: line 165: ../h5dump: Is a directory h5importtestutil.sh: line 168: ../h5dump: No such file or directory PASSED Testing BINARY F64 - rank 3 - Output LE+CHUNKED+Extended+Compressed h5importtestutil.sh: line 165: ../h5dump: Is a directory h5importtestutil.sh: line 168: ../h5dump: No such file or directory PASSED Testing H5DUMP-BINARY F64 - rank 3 - Output LE+CHUNKED+Extended+Compreh5importtestutil.sh: line 186: ../h5dump: Is a directory One or more of the required fields (RANK, DIMENSION-SIZES) missing. Configuration parameters are invalid in dbinfp64.h5.dmp. Error in processing the configuration file: dbinfp64.h5.dmp. Program aborted. HDF5-DIAG: Error detected in HDF5 (1.8.11) thread 0: #000: H5L.c line 824 in H5Lexists(): unable to get link info major: Symbol table minor: Object not found #001: H5L.c line 2765 in H5L_exists(): path doesn't exist major: Symbol table minor: Object already exists #002: H5Gtraverse.c line 861 in H5G_traverse(): internal path traversal failed major: Symbol table minor: Object not found #003: H5Gtraverse.c line 755 in H5G_traverse_real(): component not found major: Symbol table minor: Object not found *FAILED* Testing BINARY I8 - rank 3 - Output I16LE + Chunked+Extended+Compresseh5importtestutil.sh: line 165: ../h5dump: Is a directory h5importtestutil.sh: line 168: ../h5dump: No such file or directory PASSED Testing H5DUMP-BINARY I8 - rank 3 - Output I16LE + Chunked+Extended+Coh5importtestutil.sh: line 186: ../h5dump: Is a directory One or more of the required fields (RANK, DIMENSION-SIZES) missing. Configuration parameters are invalid in dbinin8.h5.dmp. Error in processing the configuration file: dbinin8.h5.dmp. Program aborted. HDF5-DIAG: Error detected in HDF5 (1.8.11) thread 0: #000: H5L.c line 824 in H5Lexists(): unable to get link info major: Symbol table minor: Object not found #001: H5L.c line 2765 in H5L_exists(): path doesn't exist major: Symbol table minor: Object already exists #002: H5Gtraverse.c line 861 in H5G_traverse(): internal path traversal failed major: Symbol table minor: Object not found #003: H5Gtraverse.c line 755 in H5G_traverse_real(): component not found major: Symbol table minor: Object not found *FAILED* Testing BINARY I16 - rank 3 - Output order LE + CHUNKED + extended h5importtestutil.sh: line 165: ../h5dump: Is a directory h5importtestutil.sh: line 168: ../h5dump: No such file or directory PASSED Testing H5DUMP-BINARY I16 - rank 3 - Output order LE + CHUNKED + extenh5importtestutil.sh: line 186: ../h5dump
Re: [Hdf-forum] building hdf5-1.8.10-patch1: H5_CFLAGS variable problems for Cray compiler
Hi Neil, Since the configure process is executed in the front end which is a Linux system, configure naturally thinks it is working in a Linux system, therefore it goes to the Linux configure settings. I think the right solution is to tell configure to build for a cross-compiling Cray system. This will take some effort and won't be available in the pending release v1.8.11. We are planning to provide it in the next release. -Albert Cheng The HDF Group On 5/2/13 3:49 PM, Neil wrote: Hi Albert, We are using a Cray XK6 running Cray's Compute Linux Environment (CLE) with the Cray CCE compiler. The front end nodes run SLES Linux and the compute nodes run CLE. So when building HDF5 (and any other application/library) it's a cross compilation. I can get the library to build successfully, both serial and parallel versions by hacking the config/linux-gnulibcl to remove the inclusion of the intel-flags and intel-fflags files (as stated previously). It would be handy if the configure script could correctly identify the Cray CCE compiler as some users may not be aware that this is the problem. We use pretty much the same method as described in the release_docs/INSTALL_parallel. As the doc says, the key is to set the RUNPARALLEL and RUNSERIAL environment variables. With the Cray CCE compiler it doesn't generate Fortran module files by default, so you have to make sure that -em is set in FFLAGS. For CFLAGS and CXXFLAGS we also have to set -h tolerant -h system_alloc. I'll try to post our configure options for both serial and parallel builds tomorrow when I next have access to the system in case this is of help. Cheers, Neil. -- View this message in context: http://hdf-forum.184993.n3.nabble.com/building-hdf5-1-8-10-patch1-H5-CFLAGS-variable-problems-for-Cray-compiler-tp4026036p4026154.html Sent from the hdf-forum mailing list archive at Nabble.com. ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org
Re: [Hdf-forum] HDF5 on Redhat
HDF5 configure detects you are using a parallel MPI compiler (.../mpicc) and turns on --enable-parallel by default which set RUNPARALLEL to .../mpirun. Your version of mpirun probably needs mpd to present, thus the failure. (Note: this is an educated guess, not necessary what happened in your machine.) If you want to use parallel HDF5, you need to setup your environment to run MPI programs. You should check with your system administrator how to do that in your machine. If you don't mean to build parallel HDF5, then don't use mpicc to build HDF5. But if you want to use mpicc but not to build parallel HDF5, then you should do, CC=.../mpicc ./configure --disable-parallel ... Hope this helps. -Albert Cheng The HDF Group On 5/1/13 6:03 PM, brown wrap wrote: I have just compiled HDF5 1.8.10 on RHEL Server 5.9. If I compile it with gcc and run the test, make check, everything is fine. If I compile it using: CC=/share/apps/opt/intel/impi/4.1.0.024/intel64/bin/mpicc ./configure --prefix=/share/apps/hdf5-impi-intel-64 I get the following errors: cannot connect to local mpd (/tmp/mpd2.console_ramos); possible causes: 1. no mpd is running on this host 2. an mpd is running but was started without a console (-n option) Command exited with non-zero status 255 0.14user 0.02system 0:00.23elapsed 73%CPU (0avgtext+0avgdata 37520maxresident)k 0inputs+0outputs (0major+5476minor)pagefaults 0swaps make[4]: *** [t_mpi.chkexe_] Error 1 make[4]: Leaving directory `/users/ramos/hdf5-1.8.10/testpar' make[3]: *** [build-check-p] Error 1 make[3]: Leaving directory `/users/ramos/hdf5-1.8.10/testpar' make[2]: *** [test] Error 2 make[2]: Leaving directory `/users/ramos/hdf5-1.8.10/testpar' make[1]: *** [check-am] Error 2 make[1]: Leaving directory `/users/ramos/hdf5-1.8.10/testpar' make: *** [check-recursive] Error 1 ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org
Re: [Hdf-forum] building hdf5-1.8.10-patch1: H5_CFLAGS variable problems for Cray compiler
Hi Neil, What kind of Cray and OS version do you use? Does it have a frontend for compiling and a backend for compute? If so, what kind of machine and OS version is the frontend? There are problems with the configure process for the Cray platform that we will not be able to fix it in time for this pending release. We plan to fix it in the next release. Meanwhile, can you follow the instruction in release_docs/INSTALL_parallel, under section 2.4. Hopper (Cray XE6) (for v1.8 and later) to see if you can build HDF5 in your Cray? By the way, are you building the serial HDF5 library or the Parallel HDF5 which requires MPI support? -Albert Chen The HDF Group On 4/30/13 4:57 AM, Neil wrote: Hi again, As promised, here is the version output information for the Cray compiler (CCE). Note: this version information is obtained using the command: cc -V. |cc -v| and |cc --version| are not supported. |cc -V /opt/cray/xt-asyncpe/5.17/bin/cc: INFO: Compiling with CRAYPE_COMPILE_TARGET=native. Cray C : Version 8.1.7 Tue Apr 30, 2013 09:48:17| I have also tried out HDF5 1.8.11-pre2, but it also incorrectly identifies the Cray compiler as the Intel compiler. Hope this helps, Neil. View this message in context: Re: building hdf5-1.8.10-patch1: H5_CFLAGS variable problems for Cray compiler http://hdf-forum.184993.n3.nabble.com/building-hdf5-1-8-10-patch1-H5-CFLAGS-variable-problems-for-Cray-compiler-tp4026036p4026141.html Sent from the hdf-forum mailing list archive http://hdf-forum.184993.n3.nabble.com/ at Nabble.com. ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org
[Hdf-forum] HDF5 1.8.11 release candidate is available for testing
Dear HDF users, A pre-release candidate version of HDF5 1.8.11 is available for testing and can be downloaded at the following link: http://www.hdfgroup.uiuc.edu/ftp/pub/outgoing/hdf5/hdf5-1.8.11/hdf5-1.8.11-pre2.tar.gz For Windows users, there is an equivalent http://www.hdfgroup.uiuc.edu/ftp/pub/outgoing/hdf5/hdf5-1.8.11/hdf5-1.8.11-pre2.zip http://www.hdfgroup.uiuc.edu/ftp/pub/outgoing/hdf5/hdf5-1.8.11/hdf5-1.8.11-pre2.tar.gz The tar ball contains fixes for a few CMake problems reported on FORUM for the previous release candidate. Specifically, the requested enhancement for building parallel HDF5 with CMake has been addressed. Last time we failed to announce that HDF5 1.8.11 release will contain two new features that may be of interest to anyone who works with the compressed data in HDF5. - Writing a compressed chunk bypassing HDF5 library internals for a better speed. We added a high-level function H5DOwrite_chunk documented here http://www.hdfgroup.org/HDF5/doc_test/H51.8docs/Advanced/DirectChunkWrite/UsingDirectChunkWrite.pdf -Dynamically loaded filters in HDF5 will allow to load a compression library at run time. With this feature the users will be able to use the HDF5 library and tools distributed by The HDF Group with the thrid-party HDF5 compression methods (and other filters). The document that describes the feature and the BZI2 plugin example is available from http://www.hdfgroup.org/HDF5/doc_test/H51.8docs/Advanced/DynamicallyLoadedFilters/HDF5DynamicallyLoadedFilters.pdf If you have some time to test this pre-release, we would greatly appreciate it. Newly reported issues will be documented and will be addressed in the next release. Thank you! -Albert Cheng The HDF Group ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org
Re: [Hdf-forum] building hdf5-1.8.10-patch1: H5_CFLAGS variable problems for Cray compiler
Hi Neil, I worked in a Cray about 3 to 4 years ago. Back then the cc and ftn were actually cross compilers provided by PGI. Maybe they are changed and supported by Intel. One of our staff has more recent experience with Cray XE6, Cray Linux Environment (CLE) PrgEnv-pgi/4.0.46 hopper.nersc.gov pgcc 12.5-0 64-bit target on x86-64 Linux -tp shanghai pgf90 12.5-0 64-bit target on x86-64 Linux -tp shanghai pgCC 12.5-0 64-bit target on x86-64 Linux -tp shanghai We are in a pre-release testing for v1.8.11 and a pre-release source tarball is available at: http://www.hdfgroup.uiuc.edu/ftp/pub/outgoing/hdf5/hdf5-1.8.11/hdf5-1.8.11-pre1.tar.gz After your have downloaded the tar file, untar it into hdf5-1.8.11-pre1/, cd into it and tried to follow the instruction in release_docs/INSTALL_parallel, under 2.4. Red Storm (Cray XT3) (for v1.8 and later) or 2.5. Hopper (Cray XE6) (for v1.8 and later). The section 2.4 is more detailed but it is old information for Red Storm. The section 2.5 is current. Please see which one is more relevant to your Cray. (And let me know too.) If it still fails for you in any way, please send me all these information: 1. Machine information (Cray so and so) 2. All commands used and their output. 3. compiler versions (e.g. output of cc --version) Thanks. -Albert Cheng The HDF Group On 4/24/13 9:57 AM, Neil wrote: Hi, I too have had the same problem with both HDF5 1.8.5 and 1.8.8 with the Cray compiler. The -std=c99 flag gets added because the HDF5 configure script incorrectly identifies the Cray compiler has the Intel compiler and consequently it automatically sets a load of compiler flags that are intended for the Intel compilers. To workarround this edit config/linux-gnulibcl and comment out the two lines that source files named intel-flags and intel-fflags. It would be handy if future versions of HDF5 could detect the Cray compiler correctly. Hope this helps, Neil. -- View this message in context: http://hdf-forum.184993.n3.nabble.com/building-hdf5-1-8-10-patch1-H5-CFLAGS-variable-problems-for-Cray-compiler-tp4026036p4026136.html Sent from the hdf-forum mailing list archive at Nabble.com. ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org
Re: [Hdf-forum] HDF5 1.8.11 release candidate is available for testing -- HighLevel library missing
Hi Pedro, Thank you for your feedback. I think you are on a Windows system but could you give more details like OS versions, platform specifics (32, 64 bits,... ), compilers, ... I will check with our Cmake person of your suggestions. A question about your suggestion 2. Wouldn't it be better to use folder names to specify different versions? E.g., ...\hdf5-1.8.11-pre1\HDF5.sln ...\hdf5-1.8.10\HDF5.sln On 4/18/13 9:47 PM, Pedro Vicente wrote: Hi Albert I tried the CMake build with T:\hdf5-1.8.11-pre1\buildcmake .. and unlike the previous hdf5-1.8.10 reported errors, no errors this time, However, the High-level library is not included in the build. The netCDF build requires the Dimension Scales library, so this is a must for netCDF users. 1) Would it be possible to have a CMake build for the High-level library ? 2) I have a suggestion, if you could name the CMake generated Visual Studio solution with the HDF5 release version name, like hdf5-1.8.11-pre1.sln instead of HDF5.sln This is helpful for anyone that uses several versions Thanks Pedro -- Pedro Vicente, Earth System Science University of California, Irvine http://www.ess.uci.edu/ - Original Message - From: Albert Cheng ach...@hdfgroup.org To: HDF Users Discussion List hdf-forum@hdfgroup.org Sent: Friday, April 12, 2013 1:59 PM Subject: [Hdf-forum] HDF5 1.8.11 release candidate is available for testing Hello everyone, A pre-release candidate version of HDF5 1.8.11 is available for testing and can be downloaded at the following link: http://www.hdfgroup.uiuc.edu/ftp/pub/outgoing/hdf5/hdf5-1.8.11/hdf5-1.8.11-pre1.tar.gz If you have some time to test this pre-release, we would greatly appreciate it. We try to test on a wide variety of platforms and environments but are unable to test everywhere so feedback from the user community is always welcome. Please note that while the release notes contained in the pre-release are reflective of the changes and additions present in this release, the 'platforms tested' and 'tested configurations' sections have yet to be updated for this version of HDF5. We plan to release HDF5 1.8.11 in mid-May barring the discovery of any critical issues. Thank you! The HDF Group ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org
[Hdf-forum] HDF5 1.8.11 release candidate is available for testing
Hello everyone, A pre-release candidate version of HDF5 1.8.11 is available for testing and can be downloaded at the following link: http://www.hdfgroup.uiuc.edu/ftp/pub/outgoing/hdf5/hdf5-1.8.11/hdf5-1.8.11-pre1.tar.gz If you have some time to test this pre-release, we would greatly appreciate it. We try to test on a wide variety of platforms and environments but are unable to test everywhere so feedback from the user community is always welcome. Please note that while the release notes contained in the pre-release are reflective of the changes and additions present in this release, the 'platforms tested' and 'tested configurations' sections have yet to be updated for this version of HDF5. We plan to release HDF5 1.8.11 in mid-May barring the discovery of any critical issues. Thank you! The HDF Group ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org
Re: [Hdf-forum] [Paraview] Fortran wrapper of HDF5
Years ago, I heard one explanation for the seeming contradiction between the C order and the Fortran order that the C order is computer science based while the Fortran order is mathematics based. Here is what it means. The C order, aka row major order, is last index changes first. So, a 10 by 10 two dimensional array is stored in memory like this: (0,0), (0,1), ..., (0,9), (1,0), (1,1),..., (9,9) which is as natural as 00, 01, ..., 09, 10, 11, ..., 99. The Fortran order, aka column major order, is first index changes first. So, a 10 by 10 two dimensional array is stored in memory like this: (1,1), (2,1), ..., (10,1), (1, 2), (2,2), ..., (10, 10) which seems unnatural. But if you recall when you learned the Cartesian coordination in high school mathematics. The x-axis is horizontal and y-axis is vertical. Points on the Cartesian plain are (x, y): (1,1), (2,1), ..., (10, 1), ... Since Fortran, was first known as FORTRAN, was derived from The IBM Mathematical Formula Translating System, the Fortran order is natural too. One more comment, the two ordering, though seem like a simple rotation at two dimensions, are not really the same in physics for higher dimensions. The C order is in left-hand rule, while the Fortran order is in right-hand rule. If mixed up, you may attempt to operate the human heart in the wrong side. :-) -Albert On 3/14/13 11:57 PM, Elena Pourmal wrote: John, I think you explained it well. Looks like Jason is not along. Let me try to explain one more time To address Jason's I've read the section of the hdf5 page that Pradeep linked to (C vs Fortran ordering) several times, and for the life of me, I can't figure out whether the hdf5 format ordering is transparent to the application or not Yes, it is transparent. Think about N-dim array in any language as 1-dim flatten array. Transparency means that one always gets the same 1-dim flatten array of data in any programming language used. Dimensions of the user's array (dataset dimensions) are stored in HDF5 file as 1-dim array following a convention that the last element of that array has the size of the fastest changing dimension of the user's array. File format is INDEPENDENT from which language an array is written (and yes, it is turned to be a C convention :-) When an array is written from Fortran, the value of its first dimension is stored in the last element of the 1-dim array in the HDF5 file. When Fortran library queries dimensions for an array (dataset dimensions), it knows as does the C library, that the last element of the 1-dim array in the file is the size of the fastest changing dimension of the user's array. It will be the first dimension for the Fortran array and the last dimension for the C array. I.e., Fortran wrappers flip dimensions, but NEVER touch user's data. If one reads data written by a Fortran application using a C application (and vice versa), data will appear to be transposed, but it will be same data if you think about it as 1-dim flatten array! Is it even more confusing now? :-) Elena ~~ Elena Pourmal The HDF Group http://hdfgroup.org 1800 So. Oak St., Suite 203, Champaign IL 61820 217.531.6112 ~~ On Mar 13, 2013, at 2:28 PM, Biddiscombe, John A. wrote: I'm forwarding this to the hdf mailing list just in case someone can explain it better, or correct me You write a Fortran array with dimensions say arr[10,20,30] {x=10,y=20,z=30}, and this goes onto disk as an array with dimension sizes 0 = 10 1 = 20 2 = 30 The convention is simply that the dimensions are listed in row major order, so if you read the data in c into an array arr[z][y][x] then everything will look fine. nine times out of ten, I've found that fortran arrays are declared as arr[x,y,z] and c arrays as arr[z][y][x] so the programmer has already flipped the orders of the dimensions and everything works out. All hdf does is say that the dimensions are listed in row major order, you can interpret the data how you like. Hdf doesn't say 'this is X, this is Y, this is Z' - it only says, dim 0 is size 10, dim 1 size 20, dim 2 size 30 - you may add metadata yourself to tell the user which 'should' be X/Y/Z if you wish. If the OP has data physically stored as fortran array[z,y,x] then the data will be written out transposed relative to what we expect when reading into paraview, then transposing will be necessary (we won't go into coordinate transforms to achieve the same flipping at the graphics end). Did I get that right? JB -Original Message- From: paraview-boun...@paraview.org mailto:paraview-boun...@paraview.org [mailto:paraview-boun...@paraview.org] On Behalf Of Jason Fleming Sent: 13 March 2013 19:51 To: parav...@paraview.org mailto:parav...@paraview.org Subject: Re: [Paraview] Fortran wrapper of HDF5 I've read the section of the hdf5 page that Pradeep linked
Re: [Hdf-forum] make check FAILED testlibinfo.sh on Mac OS X 10.7.5 with GNU and Intel compilers
(George, thanks for the good information.) Pat, I copied a copy of libhdf5.7.dylib I generated in a Darwin 11.4.2 (Lion) to a Darwin 10.8.0 (Snow Leopard) and did $ strings libhdf5.7.dylib /dev/null strings: object: libhdf5.dylib malformed object (unknown load command 8) Puzzled, I read man strings and found this: Description Strings looks for ASCII strings in a binary file or standard input. Strings is useful for identifying random object files and many other things. A string is any sequence of 4 (the default) or more printing characters ending with a newline or a null. Unless the - flag is given, strings looks in _all sections of the object files_ except the (__TEXT,__text) section. If no files are specified standard input is read. So, I guess strings detects libhdf5.7.dylib is an object and tries to find the sections. But since the object file is generated in a newer OS (Lion), the older strings in Snow Leopard does not understand the newer format. I then tried this in the Snow Leopard system: $ strings libhdf5.dylib | grep SUMMARY SUMMARY OF THE HDF5 CONFIGURATION My suggestions to you: 1. Ask your system support staff to update the strings command in your system; Or 2. Edit the test/testlibinfo.sh file by changing this line: if strings $1 | grep SUMMARY OF THE HDF5 CONFIGURATION /dev/null; then to if strings $1 | grep SUMMARY OF THE HDF5 CONFIGURATION /dev/null; then ^ Meanwhile, I will update the future release of HDF5 to use this solution. It will take a micro seconds to run it but it avoids this system versions issue. Hope this helps. -Albert On 2/3/13 12:37 PM, George N. White III wrote: It looks like you have an old version of the strings program. On Snow Leopard with Xcode 3.2 I have: $ /usr/bin/strings /usr/bin/strings | grep cctools cctools-750 $ /usr/bin/strings /opt/local/bin/strings | grep cctools cctools-836 $ port provides /opt/local/bin/strings /opt/local/bin/strings is provided by: cctools $ port info cctools cctools @836 (devel) Variants: llvm29, llvm30, [+]llvm31, llvm32, universal Description: A set of essential tools to support development on Mac OS X and Darwin. Conceptually similar similar to binutils on other platforms. Homepage: http://opensource.apple.com/source/cctools/ Build Dependencies: ld64 Library Dependencies: cctools-headers Runtime Dependencies: llvm-3.1 Platforms:darwin License: APSL-2 GPL-2+ Maintainers: jerem...@macports.org mailto:jerem...@macports.org, openmaintai...@macports.org mailto:openmaintai...@macports.org cctools-836 works, but cctools-750 fails. $ /opt/local/bin/strings /opt/local/lib/libhdf5.7.dylib | grep SUMMARY SUMMARY OF THE HDF5 CONFIGURATION $ /usr/bin/strings /opt/local/lib/libhdf5.7.dylib | grep SUMMARY /usr/bin/strings: object: /opt/local/lib/libhdf5.7.dylib malformed object (unknown load command 8) I have also run into situations where Apple's ld64 fails but macports build of the Apple open source tool worked (e.g.,LDFLAGS=-B/opt/local/libexec/ld64). On Fri, Feb 1, 2013 at 11:53 AM, pkabl...@atmos.umd.edu mailto:pkabl...@atmos.umd.edu wrote: Thanks for the advice, Albert. I get the same error when I execute the if statement from the command line. However, I get no error when I execute $ strings /etc/group /dev/null So, any idea why libhdf5.7.dylib would be a 'malformed object'? I have included the output from make below with the hope that it may help. -Pat Making all in src make[1]: Entering directory `/Users/pkablick/Downloads/tarballs/hdf5-1.8.10-patch1-intel-serial/src' make all-am make[2]: Entering directory `/Users/pkablick/Downloads/tarballs/hdf5-1.8.10-patch1-intel-serial/src' ...[skipped] On 1/30/13 4:36 PM, acheng at hdfgroup.org http://hdfgroup.org wrote: The testlibinfo.sh does something simple in the first test. It just does this: if strings ../src/.libs/libhdf5.7.dylib | grep SUMMARY OF THE HDF5 CONFIGURATION /dev/null; then echo PASSED else echo FAILED nerrors=`expr $nerrors + 1` fi From the message you showed, I think the error was from the command strings. Can you cd test and hand run the above commands? (They are Bourn sh style commands. So, if you use cshell, do a sh first.) If you see the same error message, try this: $ strings /etc/group /dev/null If you see the same error message again, something is wrong with the command strings executable. In that case, check with your system support staff. Let me know how it turns out. -Albert Cheng On 1/28/13 12:02 PM, pkablick at atmos.umd.edu http://atmos.umd.edu wrote: Hello, I'm trying to build HDF5 on Mac OS
Re: [Hdf-forum] make check FAILED testlibinfo.sh on Mac OS X 10.7.5 with GNU and Intel compilers
The testlibinfo.sh does something simple in the first test. It just does this: if strings ../src/.libs/libhdf5.7.dylib | grep SUMMARY OF THE HDF5 CONFIGURATION /dev/null; then echo PASSED else echo FAILED nerrors=`expr $nerrors + 1` fi From the message you showed, I think the error was from the command strings. Can you cd test and hand run the above commands? (They are Bourn sh style commands. So, if you use cshell, do a sh first.) If you see the same error message, try this: $ strings /etc/group /dev/null If you see the same error message again, something is wrong with the command strings executable. In that case, check with your system support staff. Let me know how it turns out. -Albert Cheng On 1/28/13 12:02 PM, pkabl...@atmos.umd.edu wrote: Hello, Noob here. I'm trying to build HDF5 on Mac OS X 10.7.5, but I've run into a problem that I haven't been able to fix. Any help or suggestions are appreciated. I've tried to build with both GNU and Intel compilers, and I get the same failure with both compiler suites on make check: Testing testlibinfo.sh *** Error ignored Finished testing testlibinfo.sh testlibinfo.sh Test Log Check file ../src/.libs/libhdf5.7.dylib strings: object: ../src/.libs/libhdf5.7.dylib malformed object (unknown load command 15) FAILED Check file ../src/.libs/libhdf5.a PASSED Check file testhdf5-SKIP- ***1 errors encountered*** 0.31 real 0.03 user 0.05 sys Finished testing testlibinfo.sh Based on suggestions from other fora, here is some information about the dynamic library using file, otool and lipo: $ file src/.libs/libhdf5.7.dylib src/.libs/libhdf5.7.dylib: Mach-O 64-bit dynamically linked shared library x86_64 $ otool -L src/.libs/libhdf5.7.dylib src/.libs/libhdf5.7.dylib: /usr/local/lib/libhdf5.7.dylib (compatibility version 8.0.0, current version 8.4.0) /usr/lib/libz.1.dylib (compatibility version 1.0.0, current version 1.2.5) /usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 159.1.0) mac64/libcilkrts.5.dylib (compatibility version 0.0.0, current version 0.0.0) /usr/lib/libstdc++.6.dylib (compatibility version 7.0.0, current version 52.0.0) /usr/lib/libgcc_s.1.dylib (compatibility version 1.0.0, current version 1094.0.0) $ lipo -info src/.libs/libhdf5.7.dylib Non-fat file: src/.libs/libhdf5.7.dylib is architecture: x86_64 I run the following bash script to perform the build: #!/bin/bash # version of HDF5 export HDF5=hdf5-1.8.10-patch1 # location of installation export PREFIX=/usr/local # location of compiler export COMPREFIX=/usr # c preprocessor export CPP='icc -E' # c compiler export CC=$COMPREFIX/bin/icc # archiver and linker export AR=$COMPREFIX/bin/xiar export LD=$COMPREFIX/bin/xild # c flags export CFLAGS='-O0' # linker flags export LDFLAGS=-L$COMPREFIX/lib # compression libraries used by HDF5 export ZLIB=$COMPREFIX rm -rf $HDF5-intel-serial tar xzf $HDF5.tar.gz mv $HDF5 $HDF5-intel-serial cd $HDF5-intel-serial ./configure --prefix=$PREFIX --with-zlib=$ZLIB # tell make to continue even if errors are encountered export HDF5_Make_Ignore=yes # use shared libraries export HDF5_USE_SHLIB=yes make all make check Also, here is some system information that may be helpful: $uname -a Darwin kestrel.local 11.4.2 Darwin Kernel Version 11.4.2: Thu Aug 23 16:25:48 PDT 2012; root:xnu-1699.32.7~1/RELEASE_X86_64 x86_64 $ which icc which ifort which icpc /usr/bin/icc /usr/bin/ifort /usr/bin/icpc $ icc --version ifort --version icpc --version icc (ICC) 13.0.1 20121010 Copyright (C) 1985-2012 Intel Corporation. All rights reserved. ifort (IFORT) 13.0.1 20121010 Copyright (C) 1985-2012 Intel Corporation. All rights reserved. icpc (ICC) 13.0.1 20121010 Copyright (C) 1985-2012 Intel Corporation. All rights reserved. $ which gcc which gfortran which g++ /usr/local/bin/gcc /usr/local/bin/gfortran /usr/local/bin/g++ $ gcc --version gfortran --version g++ --version gcc (GCC) 4.8.0 20120603 (experimental) Copyright (C) 2012 Free Software Foundation, Inc. GNU Fortran (GCC) 4.8.0 20120603 (experimental) Copyright (C) 2012 Free Software Foundation, Inc. g++ (GCC) 4.8.0 20120603 (experimental) Copyright (C) 2012 Free Software Foundation, Inc. ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org ___ Hdf-forum is for HDF software users discussion. Hdf-forum
Re: [Hdf-forum] 1.8.10 pre-release problem
Hi Scott, We are not using cmake to build HDF5 library in our Mac systems, yet. I have a built of h5dump in my Mountain Lion system and used it to dump the file you provided. It worked fine (see output h5dump.out in the ftp space below.) Here is my solution, for now: 1. I have deposited my h5dump executable and your hdf5 data file in my FTP outgoing directories as below. Please downloaded it and try it out. ftp://ftp.hdfgroup.uiuc.edu/pub/outgoing/acheng/h5dump/ mdsum output to verify you got the right files. da2f13df506878b406146524eff7d3da diblock2s_History.h5 8fe01cd79b43ba6f528ee7be314da2b1 h5dump 7d1047e6339ad415f3a2cecc8c3aa4ab h5dump.out 2. If my executable works in your system, try use the following steps to build your own and see if it works. ./configure; make; make check; make install; Let me know if this solution works for you, for now. -Albert Cheng On 10/17/12 5:37 PM, Scott Sides wrote: thanks a lot. Let me know if you find a solution and if you need any more info. Scott On 10/17/12 4:31 PM, Peter Cao wrote: Hi Scott, Thank you for the information and the sample file. We will build 1.8.10 with the same configuration and try to reproduce the problem. Thanks --pc On 10/17/2012 3:27 PM, Scott Sides wrote: Here's the info from our build/configure system for hdf5-1.8.10. I didnt do the 'make check' test. h5dump seg faults when trying to access file. Scott #!/bin/bash # Clear cmake cache to ensure a clean configure. rm -rf CMakeFiles CMakeCache.txt '/Users/swsides/contrib/cmake/bin/cmake' \ -DCMAKE_INSTALL_PREFIX:PATH=/Users/swsides/contrib/hdf5-1.8.10-pre1-ser \ -DCMAKE_BUILD_TYPE:STRING=Release \ -DCMAKE_COLOR_MAKEFILE:BOOL=FALSE \ -DCMAKE_VERBOSE_MAKEFILE:BOOL=TRUE \ -DCMAKE_C_COMPILER:FILEPATH='/usr/bin/gcc' \ -DCMAKE_CXX_COMPILER:FILEPATH='/usr/bin/g++' \ -DCMAKE_Fortran_COMPILER:FILEPATH='/usr/local/bin/gfortran' \ -DCMAKE_C_FLAGS:STRING='-pipe -fPIC' \ -DCMAKE_CXX_FLAGS:STRING='-pipe -fPIC' \ -DCMAKE_Fortran_FLAGS:STRING='-fPIC' \ -DHDF5_BUILD_TOOLS:BOOL=ON \ -DHDF5_BUILD_HL_LIB:BOOL=ON \ -DHDF5_BUILD_FORTRAN:BOOL=ON \ /Users/swsides/chompstall/builds/hdf5-1.8.10-pre1 this is the command I used for looking at the file thats attached. /Users/swsides/contrib/hdf5-1.8.10-pre1-ser/bin/h5dump diblock2s_History.h5 Thanks! Scott On 10/17/12 2:19 PM, Albert Cheng wrote: Hi Scot, I would like to ask for some more information about h5dump seg fault in your Mountain Lion system. Did you also use gcc compilers to build hdf5 v1.8.10 pre-release in your mountain lion system? What configure option(s) did you use to build the library and tools? How does h5dump seg fault on you? Did it seg fault during make check? Or did it seg fault when you used it to access your own HDF5 data files? If the former (during make check), can you please send me whatever output it shows when it seg faulted. If the latter, any chance I may get a copy of your HDF5 files and the command you used? Thanks. -Albert On 10/17/12 11:28 AM, Scott Sides wrote: Dave, thanks for pointing that out. I had two windows open. Here are the results for my Mac Lion. polymer:~ swsides$ gfortran --version ;gcc --version ; uname -a GNU Fortran (GCC) 4.6.2 20111019 (prerelease) Copyright (C) 2011 Free Software Foundation, Inc. GNU Fortran comes with NO WARRANTY, to the extent permitted by law. You may redistribute copies of GNU Fortran under the terms of the GNU General Public License. For more information about these matters, see the file named COPYING i686-apple-darwin11-llvm-gcc-4.2 (GCC) 4.2.1 (Based on Apple Inc. build 5658) (LLVM build 2336.1.00) Copyright (C) 2007 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. Darwin polymer.txcorp.com 11.4.2 Darwin Kernel Version 11.4.2: Thu Aug 23 16:25:48 PDT 2012; root:xnu-1699.32.7~1/RELEASE_X86_64 x86_64 On 10/17/12 10:23 AM, Scott Sides wrote: [swsides@volt lists]$ gfortran --version ; gcc --version ; uname -a GNU Fortran (GCC) 4.4.4 20100726 (Red Hat 4.4.4-13) Copyright (C) 2010 Free Software Foundation, Inc. GNU Fortran comes with NO WARRANTY, to the extent permitted by law. You may redistribute copies of GNU Fortran under the terms of the GNU General Public License. For more information about these matters, see the file named COPYING gcc (GCC) 4.4.4 20100726 (Red Hat 4.4.4-13) Copyright (C) 2010 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. Linux volt.txcorp.com 2.6.32-71.29.1.el6.x86_64 #1 SMP Mon Jun 27 19:49:27 BST 2011 x86_64 x86_64 x86_64 GNU/Linux On 10/17/12 10:14 AM, Raymond Lu wrote: Scot, What compiler do you use? What version is it? Thanks. Ray On Oct 17, 2012, at 10:03 AM, Scott Sides wrote: I just tried
Re: [Hdf-forum] HDF5 1.8.9 release candidate is available for testing
On 10/16/12 12:17 PM, Sean McBride wrote: On Tue, 16 Oct 2012 11:38:15 -0500, Albert Cheng said: Please note that while the release notes contained in the pre-release are reflective of the changes and additions present in this release, the 'platforms tested' and 'tested configurations' sections have yet to be updated for this version of HDF5. For the Macs listed, might I suggest adding the Xcode version number? You can get it from Xcode's 'About Xcode' menu. I'd put something like Xcode 4.5.1 (4G1004). This entry is suspicious: Mac OS X Mountain Lion 10.8.1 GCC 4.2.1 gcc (owl)GNU Fortran (GCC) 4.6.1 gfortran GCC 4.2.1. g++ Apple clang version 4.0 (cc) Apple clang version 4.0 (c++) Are you sure you're using gcc 4.2.1 there? The current Xcodes on Mountain Lion only come with clang, and not gcc. Cheers, Hi Sean, I maintained the Mountain entry. I started using Mountain Lion just over a month. Things are still a bit uncertain for me, especially in the compiler areas as Apple seems to have made some changes in Mountain Lion. For example, in snow leopard, both cc and gcc are links to gcc-4.2. In Lion, cc and gcc both links to llvm-gcc-4.2. Then things changed in Mountain Lion. Here is what I have in my Mac Mountain Lion. % sw_vers ProductName:Mac OS X ProductVersion:10.8.2 BuildVersion:12C60 % uname -a Darwin owl 12.2.0 Darwin Kernel Version 12.2.0: Sat Aug 25 00:48:52 PDT 2012; root:xnu-2050.18.24~1/RELEASE_X86_64 x86_64 % ll /usr/bin/gcc lrwxr-xr-x 1 root wheel 12 Aug 24 13:53 /usr/bin/gcc - llvm-gcc-4.2 % gcc --version i686-apple-darwin11-llvm-gcc-4.2 (GCC) 4.2.1 (Based on Apple Inc. build 5658) (LLVM build 2336.11.00) Copyright (C) 2007 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. % ll /usr/bin/cc lrwxr-xr-x 1 root wheel 5 Aug 24 13:53 /usr/bin/cc - clang % cc --version Apple clang version 4.0 (tags/Apple/clang-421.0.60) (based on LLVM 3.1svn) Target: x86_64-apple-darwin12.2.0 Thread model: posix About adding the Xcode version number. Here is what shows in my Mountain Lion. % xcodebuild -version Xcode 4.5.1 Build version 4G1004 Would Mac users prefer us reporting: 1. Only Xcode version and NO more cc or gcc version? 2. Show both Xcode and cc or gcc verison? ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org
Re: [Hdf-forum] 1.8.10 pre-release problem
Hi Scot, I would like to ask for some more information about h5dump seg fault in your Mountain Lion system. Did you also use gcc compilers to build hdf5 v1.8.10 pre-release in your mountain lion system? What configure option(s) did you use to build the library and tools? How does h5dump seg fault on you? Did it seg fault during make check? Or did it seg fault when you used it to access your own HDF5 data files? If the former (during make check), can you please send me whatever output it shows when it seg faulted. If the latter, any chance I may get a copy of your HDF5 files and the command you used? Thanks. -Albert On 10/17/12 11:28 AM, Scott Sides wrote: Dave, thanks for pointing that out. I had two windows open. Here are the results for my Mac Lion. polymer:~ swsides$ gfortran --version ;gcc --version ; uname -a GNU Fortran (GCC) 4.6.2 20111019 (prerelease) Copyright (C) 2011 Free Software Foundation, Inc. GNU Fortran comes with NO WARRANTY, to the extent permitted by law. You may redistribute copies of GNU Fortran under the terms of the GNU General Public License. For more information about these matters, see the file named COPYING i686-apple-darwin11-llvm-gcc-4.2 (GCC) 4.2.1 (Based on Apple Inc. build 5658) (LLVM build 2336.1.00) Copyright (C) 2007 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. Darwin polymer.txcorp.com 11.4.2 Darwin Kernel Version 11.4.2: Thu Aug 23 16:25:48 PDT 2012; root:xnu-1699.32.7~1/RELEASE_X86_64 x86_64 On 10/17/12 10:23 AM, Scott Sides wrote: [swsides@volt lists]$ gfortran --version ; gcc --version ; uname -a GNU Fortran (GCC) 4.4.4 20100726 (Red Hat 4.4.4-13) Copyright (C) 2010 Free Software Foundation, Inc. GNU Fortran comes with NO WARRANTY, to the extent permitted by law. You may redistribute copies of GNU Fortran under the terms of the GNU General Public License. For more information about these matters, see the file named COPYING gcc (GCC) 4.4.4 20100726 (Red Hat 4.4.4-13) Copyright (C) 2010 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. Linux volt.txcorp.com 2.6.32-71.29.1.el6.x86_64 #1 SMP Mon Jun 27 19:49:27 BST 2011 x86_64 x86_64 x86_64 GNU/Linux On 10/17/12 10:14 AM, Raymond Lu wrote: Scot, What compiler do you use? What version is it? Thanks. Ray On Oct 17, 2012, at 10:03 AM, Scott Sides wrote: I just tried the new 1.8.10 pre-release on Mac Lion and the binaries do not work. In particular the h5dump binary seg faults. Actualy, this problem seem to be persistent across lion, mountain lion and snow leopard since 1.8.7. We've had to revert to 1.8.7 on mac for some time. Is this problem known and is a fix being considered? Thanks, -- * Dr. Scott W. Sides swsi...@txcorp.com Tech-X Corporation 5621 Arapahoe Avenue, Suite A Boulder, CO 80303 office: (720) 974-1849 fax : (303) 448-7756 * ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org
[Hdf-forum] HDF5 1.8.9 release candidate is available for testing
Hello everyone, A pre-release candidate version of HDF5 1.8.10 is available for testing and can be downloaded at the following link: http://www.hdfgroup.uiuc.edu/ftp/pub/outgoing/hdf5/hdf5-1.8.10-pre1/ If you have some time to test this pre-release, we would greatly appreciate it. We try to test on a wide variety of platforms and environments but are unable to test everywhere so feedback from the user community is always welcome. Please note that while the release notes contained in the pre-release are reflective of the changes and additions present in this release, the 'platforms tested' and 'tested configurations' sections have yet to be updated for this version of HDF5. We plan to release HDF5 1.8.10 in mid-November barring the discovery of any critical issues. Thank you! The HDF Group ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org
Re: [Hdf-forum] h5edit test/Makefile.am shows 'check_PROGRAMS' recursively defined
On 9/13/2011 11:02 AM, Rhys Ulerich wrote: Hi Albert, h5edit.1.0.0.tar.gz (md5sum 065013585cc483fa7cd945020e3d967f) shows the fatal error test/Makefile.am:22: variable `check_PROGRAMS' recursively defined after configure when 'make' is invoked. Modifying test/Makefile.am to read check_PROGRAMS= $(EXTRA_TEST) fixes that issue. 'make check' seems to check nothing. Could you tell me more about the machine you test h5edit on? For example, uname -a output, compilers used. Thanks. Sure. Thanks for looking into it. $ uname -a Linux setun 2.6.32-31-generic #60-Ubuntu SMP Thu Mar 17 22:15:39 UTC 2011 x86_64 GNU/Linux $ gcc --version gcc (GCC) 4.5.1 $ automake --version automake (GNU automake) 1.11.1 $ autoconf --version autoconf (GNU Autoconf) 2.68 Also, I used CC=h5pcc (from a parallel build of HDF5). FWIW, this problem doesn't seem build-system version-dependent to me. I suspect the problem is a mixture of stale development artifacts lingering in the source tree. - Rhys Hi Rhys, I discovered that I have forgot to include an autoconfigure related file in the release tarball that will cause some problems during the make stage. But the error should not be fatal. Are you able to do the following build and installation stage? 1) % ./configure # assuming h5cc is in your $PATH 2) % gmake # some non-fatal error messages 3) % gmake check # to run tests 4) % gmake prefix=install-path# install h5edit in install-path/bin If it does not work for you, could you please email me (ach...@hdfgroup.org) the output of the above commands (or whatever commands you used)? Thanks. -Albert h5edit developer ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org
Re: [Hdf-forum] redhat and hdf5
Hi John, Redhat (equivalent to CentOS, sort of) is one of the HDF5 development platforms. You should be quite okay running HDF5 in Redhat on x86 hardware. Parallel I/O is a different issue. It depends what parallel file systems you use. Two commercial systems, GPFS (by IBM) and Lustre (now owned by Oracle) are possible options. Their performance varies and are highly dependent on the characteristics of the I/O pattern of your applications. -Albert Cheng On 7/26/2011 12:28 PM, John Knutson wrote: I've recently been informed that we're going to be changing operating systems and hardware architecture to redhat on an x86 blade system (or something of the sort). I'm starting to investigate what all the implications might be and am digging through the archives for items of interest, but in the mean-time I was wondering if anyone out there was using HDF5 on RedHat (no, I don't know which flavor, either, we haven't been given much information), and what complications or advantages they ran across in the process. One thing of particular interest to me is the parallel I/O - I'd be interested in hearing comments about that WRT RH. I'm not entirely sure it's relevant to our system but it's worth researching anyway. ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org
Re: [Hdf-forum] unresolved external symbol _H5PGET_DXPL_MPIO_C
Hi Jerome, You are right. Thank you for reporting it. It will be fixed in the next release. -Albert Cheng On 6/29/2011 6:58 PM, Soumagne Jerome wrote: Hi, As I needed hdf5 parallel/fortran support on windows (I needed it for our parallel driver), I got a strange error at compile time: Error 53 error LNK2019: unresolved external symbol _H5PGET_DXPL_MPIO_C referenced in function _H5FDMPIO_mp_H5PGET_DXPL_MPIO_F H5FDmpioff.obj Error 54 fatal error LNK1120: 1 unresolved externals C:\Projects\hdf5-vfd\bin\RelWithDebInfo\hdf5_fortrandll.dll It appears that somehow a typo was inserted in the file fortran/src/H5f90proto.h line 974 H5_FCDLL int_f nh5pget_dxpl_mpio_rc(hid_t_f *prp_id, int_f* data_xfer_mode); ^ which should be H5_FCDLL int_f nh5pget_dxpl_mpio_c(hid_t_f *prp_id, int_f* data_xfer_mode); This apparently fixed my build. Just had a quick look, it's still present in the svn trunk. Jerome -- Jérôme Soumagne Scientific Computing Research Group CSCS, Swiss National Supercomputing Centre Galleria 2, Via Cantonale | Tel: +41 (0)91 610 8258 CH-6928 Manno, Switzerland | Fax: +41 (0)91 610 8282 ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org
Re: [Hdf-forum] HDF with MPI and CPP
Hi Phil, Depends on mean with MPI. PHDF5 uses MPI-IO to support parallel I/O to HDF5 files and it has C and Fortran APIs but no C++ API. Therefore, if you want to do parallel access to HDF5 file via C++ calls, it won't work. If you use MPI for the compute and inter-processes communication but limit each HDF5 file access to a single process which in turn gathers and scatters HDF5 file data with other processes, that would still work. If you want to use parallel access to the HDF5 files, you would have to code your C++ application to call the PHDF5 C APIs. Hope this answers your question. -Albert At 04:46 AM 7/16/2010, Kraus Philipp wrote: Hi, I have had compiled the current HDF lib with --enabled-cxx and use the C++ interface. Now I would like to switch HDF with MPI. Can I use Cpp and MPI together? Thanks for the information Phil ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org ___ Hdf-forum is for HDF software users discussion. Hdf-forum@hdfgroup.org http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org