* commit '54957d37f5aa73912763dbb6e308555e863c43f4':
Commit copyright header change for src/H5PLpkg.c which was added after running script to make changes.
Add new files in release_docs to MANIFEST. Cimmit changes to Makefile.in(s) and H5PL.c that resulted from running autogen.sh.
Merge pull request #407 in HDFFV/hdf5 from ~LRKNOX/hdf5_lrk:hdf5_1_10_1 to hdf5_1_10_1
Change copyright headers to replace url referring to file to be removed and replace it with new url for COPYING file.
config/linux-gnulibc1 reapply filx for HDFFV-9439 which was partially removed, probably by merge from revise_chunks branch.
Tested with h5committest.new on kituo, ostrich, osx1010dev and platypus.
--disable-symbols is specified and to ensure that production
mode (where symbol stripping is usually a part of the FLAGS)
doesn't conflict with --enable-symbols. This will allow better
(though still limited) debugging of production/optimized code.
The PGI, XLC, Intel, and Solaris files were not modified and
do not generate the symbol removal flag at this time. They'll
be updated in future check-ins. In the meantime, production
mode on those platforms still usually strips symbols.
This check-in also updates the Cygwin-specific files to use
the new autotools FLAGS and option scheme when uncommon
(non-gnu, etc.) Fortran compilers are used.
Tested on: 64-bit Ubuntu 15.10 (Linux 4.2.0 x86_64) gcc 5.2.1
autotools serial (prod/debug w/ and w/o symbols)
symbol presence/absence checked with objdump
symbols, asserts, profiling, and optimization.
Also much refactoring, improved help, and --enable-<foo> options will
now emit errors on nonsense (e.g.: --enable-foo="asdfasdf").
The libhdf5.settings.in file was also reformatted.
Tested on: 64-bit Ubuntu 15.10 (Linux 4.2.0 x86_64) gcc 5.2.1
serial and parallel autotools w/ various options
For make installcheck, compile and run installed examples using the installed scripts. They were being compiled but not run.
Add Fortran2003 examples to the run-ex-fortran script when fortran2003 is enabled.
Set flag to -O3 in production mode for Intel compilers other than those with specified other settings.
Gentoo patches:
Remove unnecessary setting of LD_LIBRARY_PATH from configure.
Prevent potential array subscript out of bounds error in perform/pio_engine.c and sio_engine.c.
Tested with h5committest on jam, koala, ostrich and platypus, on emu and kite, and parallel tests on jam, koala and platypus with mpich built with gnu 4.8.2 and with intel compilers.
Purpose:
Remove H5_MPI_SPECIAL_COLLECTIVE_IO_WORKS and
H5_MPI_COMPLEX_DERIVED_DATATYPE_WORKS #defines from source.
Description:
Two advanced parallel functionalities, special collective IO and
complex derived datatypes, are not supported by older
implementations of mpi, and thus our code limits the use of these
features with #ifdefs and has checks in configure to set them (or
not). Unfortunately, configure can't actually run a parallel check
to see if these features are working (nor not) so it resorts to
looking in the config files where they are explicity enabled or
disabled based on versions of mpi, sytems being built on, or for
no documented reason at all (i.e. just set to on or off as some
'default'). Overriding these settings is easy if need be, provided
it is known that it needs to be done to get improved performance,
and oftentimes it is not.
Most new MPI implementations successfully handle the functionality
requested when these #defines are set, and many of the "turn these
features off" cases in the config files are for old (> 5 years)
versions of MPI and retired systems (such as NCSA's tungsten).
Therefore, the decision has been made to remove the support for
these old versions of MPI and systems that cannot handle these
behaviors. The #ifdefs and supporting setup in the config/ files
and configure script has been removed, and the code executed when
these options were not set removed from the source.
In passing, this commit also cleans up some whitespace issues in
both t_mpi.c and H5Dmpio.c. Furthermore, in t_mpi.c, the special
collective IO test was not getting regularly run due to it being
written to work only with four processes (we regularly test with
six, previously with three), and thus it failed when actually run
due to an out of bounds data buffer assignment. It has been
modified to run at any number of processes greater than four, and
the memory problem has been fixed so the test passes.
Tested:
jam, h5committest, ember
mpif90 is already the default parallel Fortran if mpicc is the C compiler.
Tested: Jam (serial and parallel). No need for committest since this change is
limited to Linux system.
gfortran is more likely installed with gcc now.
Tested:
jam and amani which covers both linux and linux64.
No h5committest because this is gnu compiler specific, usually
applicable to Linux.
rank 0 in job 77 honest1.ncsa.uiuc.edu_60945 caused collective abort of all ranks
exit status of rank 0: killed by signal 9
Up to this point, we decide to turn off the use of advanced MPI derived datatype features inside HDF5 so that complicated selection tests are ensured to be handled by HDF5. This check-in reflects this.
1. In H5Dwrite and H5Dread, let the data buffer point to a fake address if the application passes
in an empty buffer. This is mainly for MPIO programs that some processes may not have any
data to write or read but still participate the I/O. This solution solves some MPI's problem
like the ChaMPIon on tungsten which doesn't support empty buffer.
2. The ChaMPIon on tungsten doesn't support complex derived MPI data type correctly and collective
I/O when some processes don't have any data to write or read correctly. Detect the compiler
"cmpicc" in the system-specific config file and set the variables for these two cases to false.
The PHDF5 library already has set up a way to switch collective chunked I/O to independent
under these two cases.
3. A bug fix - During the work of the optimization for compound data I/O, the case for switching
collective chunked I/O to independent I/O was leftout. Fixed it by adding I/O caching to it in
H5D_multi_chunk_collective_io in H5Dmpio.c.
Tested on tungsten, cobalt, and kagiso for parallel; on linew and smirom for serial.
Daily tests failed for Absoft f95 compiler
Solution:
Failure was due to several typos in H5f90prpto.h file; fixed
I also removed redundant flags from the configration files.
Platforms tested: heping with f95, g95, pgf90; pommier with f95
H5_CFLAGS change
Description:
Converted a few platforms to use H5_CFLAGS to "test the waters"
Solution:
Modified a few config/* files to use H5_CFLAGS for non-essential
flags.
Platforms tested:
mir, heping
The previous check-in seems fixing the non-selection failure at heping.
Description:
release the special_collective_io_works macro at this platform.
Solution:
Platforms tested:
heping mpich 1.2.6
Misc. update:
Add collective chunk IO supports
Description:
Added a macro hdf5_mpi_special_collective_io_works to filter out
some mpi-io packages that don't support collective IO for no IO
contributions in some processes.
Solution:
1.
"hdf5_mpi_special_collective_io_works=${hdf5_mpi_special_collective_io_works=
'no'}"
has been added at the end of file ia64-linux-gnu and linux-gnulibc1
2. Some comments are added in ia64-linux-gnu and linux-gnulibc1.
3. Documents are added inside RELEASE.txt.
Platforms tested:
Parallel:
IBM AIX 5.2(copper)
Linux (heping) mpich-1.2.6
SDSC Teragrid mpich-1.2.5
Linux(Tungsten) mpich-1.2.6
Altix(NCSA cobalt)
Seq:
Linux(heping)
Misc. update:
Bug fix.
Description:
The previous way of detecting the MPICH version fails if the suspected
$CC does not give the expected MPICH version information. Though no
consequence, it generates error messages that alarm users.
Also, mpich2 gives the version information as if it were mpich v1.0.2
which was incorrectly included as an earlier version of the origianl
mpich.
Solution:
Implemented a more robust code to detect the version and narrow the
recognition down to v1.2.0-v1.2.5. Also, the code were duplicated in
two files (linux-gnulibc1 and ia64-linux-gnu). Combined them into a
new file called mpich.
Platforms tested:
Tested in heping using mpicc and mpich2/mpicc and also in tg-ncsa.
Misc. update:
Updated MANIFEST.
Description:
Add source code copyright notice; comment symbols vary by type of file.
Platforms tested:
No changes that should cause any problems. Nonetheless, the tree will
be checked out fresh and tests run immediately post-commit.
Code cleanup
Description:
Removed Fortran type sizes from config/* files.
Solution:
These lines weren't doing anything anymore; they've been replaced by automatic
type generation. This is just housekeeping to remove them from
the config/* files.
Platforms tested:
mir, sol, modi4
Bug fix
Description:
A whole chunk of this file somehow got deleted and the heavily reduced
file was accidentally checked in. This is the complete version of the file, with the
mpiicc fix.
Solution:
Platforms tested:
Misc. update:
Bug fix
Description:
Forgot to remove a debug statement before committing.
Solution:
The script printined out what it thought CC_BASENAME was for debugging.
Removed this print statement.
Platforms tested:
minor change, mcr
Misc. update:
Bug fix
Description:
Same bug as was just checked in for v16. Do the complicated
derived datatype check for compilers named mpiicc in addition to
mpicc.
Solution:
Platforms tested:
Misc. update:
Updating phase 2 work of collective IO
Description:
When the size is bigger than some value for MPI derived data type, mpich 1.2.5 and lower
version won't work any more, so in the configuration file, we will check the version number
and switch from collective IO to independent IO by turning off complex derived datatype.
Solution:
Platforms tested:
linux 2.4, AIX 5.1, Linux 2.4 IA64 and IRIX 6.5
Misc. update:
Automake version upgrade
Description:
Upgraded automake version from 1.6.2 to 1.9.5.
Changed bin/reconfigure script to use automake 1.9.5.
Changed configure.in and Makefiles to use new FCFLAGS and FC variables
instead of FFLAGS and F9X.
Automake and configure should now do the lion's share of the work
supporting Fortran 9X; macros in acsite.m4 are now mostly unused (will
be cleaned later).
Altered how configure handles pmake; now root-level Makefile.in is
processed by bin/reconfigure to have a .MAKEFLAGS target, since
automake no longer allows us to define unused variables.
Configure now always checks for C++ compiler even if it is not
used, since automake thinks this is the Right Thing To Do and
will break otherwise.
Platforms tested:
Sol, copper, heping, mir, sleipnir, eirene, pommier, kelgia, modi4.
Bug Fix/Code Cleanup/Doc Cleanup/Optimization/Branch Sync :-)
Description:
Generally speaking, this is the "signed->unsigned" change to selections.
However, in the process of merging code back, things got stickier and stickier
until I ended up doing a big "sync the two branches up" operation. So... I
brought back all the "infrastructure" fixes from the development branch to the
release branch (which I think were actually making some improvement in
performance) as well as fixed several bugs which had been fixed in one branch,
but not the other.
I've also tagged the repository before making this checkin with the label
"before_signed_unsigned_changes".
Platforms tested:
FreeBSD 4.10 (sleipnir) w/parallel & fphdf5
FreeBSD 4.10 (sleipnir) w/threadsafe
FreeBSD 4.10 (sleipnir) w/backward compatibility
Solaris 2.7 (arabica) w/"purify options"
Solaris 2.8 (sol) w/FORTRAN & C++
AIX 5.x (copper) w/parallel & FORTRAN
IRIX64 6.5 (modi4) w/FORTRAN
Linux 2.4 (heping) w/FORTRAN & C++
Misc. update:
Bug fix
Description:
"Teach" the configure scripts about Intel 7.1 compilers, which don't have
_quite_ the same compiler flags
Platforms tested:
Linux 2.4 (eirene)
Code cleanup
Description:
Default to using mpif90 if CC=mpicc and --enable-fortran flag is given,
but no F9X compiler is set.
Platforms tested:
Linux 2.4 (verbena)
Too minor to require h5committest
Description: Added support for Absoft Fortran compiler
Solution: Modified configuration file to check which Fortran compiler is used
and set appropriate flags.
Platforms tested: verbena with pgf90 and Absoft f95 compilers
Misc. update:
Bug Fix
Description:
The FFLAGS, CXXFLAGS, and CPPFLAGS variables weren't carrying
user-defined values into the Makefiles.
Solution:
Changed the flags from FOO="..." to FOO="$FOO ..."
Platforms tested:
Copper
Verbena
Arabica
Misc. update:
Update
Description:
Revamped the configuration system. The configurations for the Fortran
and C++ libraries are no longer separate from the "main"
configuration system. This involved removing the "configure*" and
"aclocal.m4" files from the fortran/ and c++/ subdirectories. Also
merging settings in the config/ subdirectories into the main config/
subdirectory.
Fortran header files had to be modified a little for Linux. It was
checking if it was a Linux machine by some #defines, however with the
-std=c99 switch, these defines weren't there. I added a check for
some other ones which should be there whether the -std=c99 switch is
used or not.
Platforms tested:
Verbena (Fortran & C++)
Sol (Fortran & C++)
Copper (Fortran & C++)
Modi4 (Parallel, Fortran, & C++)
Misc. update:
Fix
Description:
The "make *clean" options weren't recursing through the
subdirectories (if they exist).
Solution:
Force it to recurse through the subdirectories listed in SUBDIRS
Platforms tested:
Linux
----------------------
./config/depend.in
Fixed automatic dependencies. We were storing dependencies for
*.o files instead of *.lo files after shared libraries were
added.
./config/gnu-flags
./config/linux-gnulibc1
Moved `-march=pentiumpro -mcpu=pentiumpro -malign-double' from
the linux file to this file and caused it to depend on the CPU
name. This fixes one of Elena's bugs.
./src/H5B.c
./src/H5Bprivate.h
./src/H5D.c
./src/H5Dprivate.h
./src/H5Dpublic.h
./src/H5F.c
./src/H5Farray.c
./src/H5Fistore.c
./src/H5Fmpio.c
./src/H5Fprivate.h
./src/H5Fpublic.h
./src/H5Gnode.c
./src/H5P.c
./src/H5RA.c
./src/H5Sall.c
./src/H5Shyper.c
./src/H5Smpio.c
./src/H5Spoint.c
./src/H5Sprivate.h
./src/H5Tpublic.h
./test/istore.c
Added an H5Dget_storage_size() function that reports the
amount of storage allocated for raw data in a dataset.
Changed H5D_xfer_* to H5F_xfer_* because these properties are
more general than datasets. This also allows some of the
lower-level I/O functions to get this information easier.
./src/H5S.c
./src/H5Sall.c
Added two new functions H5S_all_read() and H5S_all_write()
which are optimizations that copy data directly between file
and memory without having to go through the scatter gather
step. This knocks quite a bit of time off the I/O and
reading/writing entire datasets is a fairly common operation.
./tools/h5ls.c
Reports the logical size of data, the allocated size of data,
and the percent utilization.
./MANIFEST
Removed old pablo files, added new files. Snapshots should now
start to work again.
./src/H5D.c
./src/H5Fmpio.c
Removed two warnings signed vs. unsigned comparisons and check
for overflow.
----------------------
./Makefile.in
./configure.in
./configure [REGENERATED]
./src/H5config.h.in [REGENERATED]
./config/commence.in
./config/conclude.in
./config/dec-osf4.x
./config/depend.in
./config/freebsd
./config/linux-gnu
./config/linux-gnulibc1
./config/linux-gnulibc2
./config/solaris2.x
./examples/Makefile.in
./src/Makefile.in
./test/Makefile.in
./testpar/Makefile.in
./tools/Makefile.in
./tools/testh5dump.sh
Changes that allow hdf5 to be configured and compiled in a
directory other than the source directory. This is especially
useful if you want to concurrently compile with different
compilers and/or compile flags, or if the hdf5 source is on
read-only media like a CDROM. The changes were surprisingly
easy ;-) Here's what you can do now...
$ mkdir /tmp/build1 # or something
$ cd /tmp/build1
$ /cdrom/hdf5/configure # where ever the source is
$ make
Paul, you'll have to change the testh5toh4 script similar to
the way I changed testh5dump.sh. I started working on it but
then gave up because of a number of problems: (1) I had to
comment out all the tests that let h5toh4 choose the output
file name because it always tried to put the output file in
the same directory as the input file, (2) if path names are
used during the h4 dump then they interfere with the
diff. The test works fine when run in the source directory,
but try this instead:
$ gunzip <hdf5-1.1.67.tar.gz |tar xf -
$ chmod -R ugo-w hdf5-1.1.67
$ mkdir build
$ cd build
$ ../hdf5-1.1.67/configure
$ make check
Dan, I didn't modify the pablo/Makefile.in because I have no
way to test it. I think all you need to do is add a couple
lines before the @COMMENCE@ line and add a couple search
directories for header files. Anyway, it seems like almost an
exact duplicate of the src/Makefile.in, so it shouldn't be a
problem...
./Makefile.in
Removes a few more temporary files during make clean and
distclean.
./configure.in
I fixed the creation of the time-stamp files so that the
initial make doesn't have to regenerated all the makefiles
(only GNU make users will see any change).
./src/H5R.c
Changed a return type from `intn' to `int' for an API
function.
./src/H5detect.c
Added the volatile qualifier to a couple more variables.
./tools/h5tools.c
Fixed a compiler warning about an unused local.
./tools/h5toh4.c
Initialized `optind' to 1 because it's zero on my system. It
should always be one on a unix system.
./tools/testh5toh4
Redirected "broken pipe" messages to /dev/null so outut is
formatted correctly.
