Added a new test to verify the MPIO can support independent
overlapping writes.
testphdf5.c:
Bracketed the MPIO write test with a #ifdef so that it is
tested only when the explicitely turned on.
Platform tested:
O2K.
----------------------
./src/H5.c
./src/H5Smpio.c
./src/H5T.c
./src/H5Tconv.c
./src/H5Tpublic.h
./src/H5detect.c
./src/H5private.h
./src/H5public.h
./test/big.c
The `long long' type isn't quite as prevalent as we had hoped;
there is at least one system that defines `__int64' but not
`long long'. Therefore, I've temporarily changed all
occurrences of `long long' to `long_long' and then #define
that in H5private.h based on the existence of `long long' or
`__int64'. This gets rid of some #ifdef's in other parts of the code.
The semantics of the hdf5 types are:
*int8_t Exactly one byte
*int16_t At least two bytes. We favor 32-bit integers over
16-bit integers if the 16-bit integer is 4-bytes wide
(Cray)
*int32_t At least four bytes.
*int64_t At least eight bytes
long_long The widest integral integer type
The H5Smpio.c contains debugging code which is non-portable.
./tools/h5ls.c
Changed the order native types are detected so we favor the name `int'
over `short' or `long' if two of them are the same.
./config/conclude.in
Added a rule to make test programs depend on the hdf5 library. This
fixes a minor bug where changing H5detect.c and then running `make
test' caused H5Tinit.c to not be recompiled and therefore the test
files are not relinked.
----------------------
./INSTALL
Added instructions about specifying a path for GNU zlib and
HDF4 headers and library.
Added comments for each of the tool names. Added h5toh4 as a
tool name.
./configure.in
./conigure [REGENERATED]
./src/H5config.h.in [REGENERATED]
The config/* scripts get invoked with $CC_BASENAME set the
base name of the compiler in order to make it easier to handle
setting compiler flags for different compilers in a big case
statement. For instance, if $CC has the value
/usr/local/mpi/bin/mpicc -ansi -64
then $CC_BASENAME will be `mpicc'. The $CC_BASENAME is not
set if $CC is not set.
Fixed alignment in `configure --help'.
An include and/or library path can be specified for GNU zlib
if configure can't find it in normal places. The "normal"
means wherever your compiler normally searches, including
search paths you've added through environment variables like
CPPFLAGS and LDFLAGS. The INSTALL file has instructions.
The `-ljpeg' library is detected.
If `ssize_t' is not found then a #define is added to
H5config.h similar to what we already do for `size_t'.
We detect the hdf5 header file `mfhdf.h' and libraries
`-lmfhdf' and `-ldf' and if found define the H5TOH4 and
TESTH5TOH4 Makefile variables. The user can specify an
include and/or library path. The INSTALL file has
instructions.
The `RUNTEST' variable has been split into `RUNSERIAL' and
`RUNPARALLEL' because these are different commands. The
makefile still uses `RUNTEST', which defaults to the RUNSERIAL
value. The new testpar/Makefile.in sets RUNTEST to the
RUNPARALLEL value. The default RUNSERIAL value is empty and
the default RUNPARALLEL value is `mpirun -np 2'. These can
both be overridden in the config/* files. To make the value
the empty string set it like `RUNPARALLEL=none' in the
config/* file.
The new testpar/Makefile is generated from testpar/Makefile.in
./config/commence.in
Added the RUNSERIAL and RUNPARALLEL makefile definitions.
./src/Makefile.in
The `H5detect' program is run with RUNSERIAL, which is empty
on all platforms except intel-osf1 (ASCI/Red)
./config/conclude.in
Removed the `.c.a:' implicit rule -- we don't use it any more.
./config/BlankForm
./config/alpha-dec
./config/alpha-dec-osf4.0
./config/freebsd2.2.7
./config/hpux10.20
./config/hpux9.03
./config/intel-osf1
./config/irix5.3
./config/irix6.2
./config/irix64
./config/linux
./config/powerpc-ibm-aix4.2.1.0
./config/rs6000-ibm-aix4.1.4.0
./config/solaris2.5
All of these files have been updated to hande multiple
compilers. Most of them assume `cc' if CC is not set. The
documentation in `BlankForm' has been updated and
documentation in the other files refer to `BlankForm'.
The intel-osf1 uses LDFLAGS instead of LIBS to specify the extra
library search paths.
./examples/Makefile.in
./test/Makefile.in
./tools/Makefile.in
Simplified by grouping some dependency information together
more succinctly.
The tools Makefile.in has been modified to work with the
`h5toh4' and `testh5toh4' programs. Until Paul checks these
in you may have problems compiling. If so, just create files
h5toh4.c and testh5toh4.c that contain:
#include <stdio.h>
int main(int argc, char *argv[]) {
fprintf(stderr, "%s: not implemented\n", argv[0]);
return 1;
}
./src/H5public.h
Removed definition for `ssize_t' since this is now handled by
configure in H5config.h with a #define.
./MANIFEST
./testpar/Makefile.in [NEW]
Added a makefile for parallel tests. As soon as we get this
working properly we can remove the other four makefiles in
that directory and maintain just one.
New feature
Problem:
Source code must be modified prior to compiling to allow full
filenames to contain preferred user directories. Source must be
recompiled everytime the preferred user directory changes.
Solution:
Instead, allow file prefixes preferred by user to be passed at
runtime using the arguments "-f <prefix>". The "-f" to signal the
next argument is to be a file prefix, and the "<prefix>" to be the
actual prefix used.
Platform tested:
ASCI Red
The purpose of the change is to correct the H5 filenames declared in
testphdf5.c when used on the Intel TFLOP PFS system.
Problem:
The ROMIO library complains of the H5 filenames having no "piofs:",
"pfs:", "ufs:", or "nfs:" prefixes on the filenames. The only type
currently recognized are the "pfs:" prefixes, and the files must
reside in the Intel TFLOP PFS system: /pfs.
Solution: (REQUIRED)
The compile flag -DPOOMA_ARCH=TFLOP is added to the list of CFLAGS in
Makefile.ascired located in the parallel test directory. This compile
flag will allow a special set of filenames to be used when compiling
testphdf5.c on the Intel TFLOP. In addition, instructions to editing
the filenames in testphdf5.c to reside in the user's own /pfs directory
are included in INSTALL.ascired.
Platform tested:
Intel TFLOP
changed to use system supported MPI library. The path to it is
changed slightly too. Updated all the paths variables.
Also updated to reflect the new program source files structure.
Platform tested: Intel Red
big testphdf5.c file. Broke it into modules of related routines.
testphdf5.c -- main routine and global variables initialization plus
some ulitility routines.
t_file.c -- tests of file operations using parallel I/O.
t_dset.c -- tests of datasets operations.
Added routines to fill in test data and to verify data read back.
Would verify data and report errors when detected.
Added verbose flag to control output "volume"--be more verbose
only if it is run with -v option.
----------------------
./configure.in
./src/H5Fprivate.h
./src/H5Fsec2.c
We now detect and use lseek64() on systems that have it (e.g.,
Irix64) and are able to generate >2GB files on Irix XFS file
systems (and anything else that supports large files). This
change also removed some warning messages from the Irix `-64'
compiler.
> $ ls -l istore.h5
> -rw-r--r-- 1 matzke meshtv 8605436856 Feb 17 14:03 istore.h5
./configure.in
./src/H5Fprivate.h
./src/H5Fstdio.h
We now detect and use fseek64() on systems that have it (e.g.,
Irix64) and are able to generate >2GB files on Irix XFS file
systems (and anything else that supports large files). This
change also removed some warning messages from the Iris `-64'
compiler.
./src/H5E.c
./src/H5Epublic.h
Added the H5E_OVERFLOW error to signal file address overflow.
./src/H5Fpublic.h
./examples/h5_chunk_read.c
./examples/h5_compound.c
./examples/h5_extend_write.c
./examples/h5_group.c
./examples/h5_read.c
./examples/h5_write.c
./html/Datasets.html
./html/Files.html
./html/H5.api.html
./html/H5.intro.html
./html/H5.sample_code.html
./html/ph5example.c
./html/review1.html
./test/cmpd_dset.c
./test/dsets.c
./test/extend.c
./test/tfile.c
./test/th5p.c
./test/theap.c
./test/tohdr.c
./test/tstab.c
./testpar/phdf5.c
Renamed file access constants to follow the naming scheme.
Also changed the base names a little to be more accurate as to
what they do. The old names H5ACC_WRITE and H5ACC_OVERWRITE
will temporarily work.
H5ACC_DEFAULT --> H5F_ACC_RDONLY for H5Fopen()
H5ACC_DEFAULT --> H5F_ACC_EXCL for H5Fcreate()
H5ACC_WRITE --> H5F_ACC_RDWR
H5ACC_OVERWRITE --> H5F_ACC_TRUNC
Albert or Kim: The H5ACC_INDEPENDENT and H5ACC_COLLECTIVE
macros in H5Fpublic.h should be an enum typedef and have names
more like H5F_MPIO_INDEPENDENT and H5F_MPIO_COLLECTIVE. Also
change the access_mode argument of H5Cset_mpio().
H5Fcreate() and H5Fopen() are more strict now about which
flags are acceptable for the operation.
./src/H5Fprivate.h
./src/H5F.c
./src/H5C.c
Changed the file access template to make it more general. A
union contains a struct for each type of low-level driver and
the default template is initialized at run-time.
./src/H5Fpublic.h
./src/H5F.c
Added H5Fget_access_template() and cleaned up
H5Fget_create_template().
./src/H5C.c
The H5Cset_mpi() no longer trashes the file access template
when an error is detected. We check for errors and *then*
update the file access template.
Added H5C_close() so Albert and Kim have a place to release
the MPI communicator and info from the file access property
list.
Kim or Albert: I notice in H5Cset_mpi() you copy the MPI
communicator. Do you need to do something similar in
H5C_copy()?
./src/H5F.c
Added more error checking for the file creation and access
property lists because it used to be possible to make the
library dump core by swapping the creation and access property
list ID numbers of H5Fcreate().
./test/istore.c
./test/tfile.c
./test/th5p.c
./test/theap.c
./test/tohdr.c
./test/tstab.c
./testpar/testphdf5.c
One must pass H5C_DEFAULT as the file creation or access
property list in order to get the default property list. It
is no longer possible to pass zero or any other arbitrary bad
object ID.
./src/H5Fcore.c
./src/H5Ffamly.c
./src/H5Flow.c
./src/H5Fmpio.c
./src/H5Fsec2.c
./src/H5Fsplit.c
./src/H5Fstdio.c
./src/H5MF.c
The file access property list is passed to all H5F_low_...()
functions and to the drivers.
./src/H5Fcore.c
The block size can be set at run time on a per-file basis
instead of at compile time across all files. The "5000 items in
a group test" now takes 1.6 seconds.
./src/H5private.h
Removed inclusion of mpi.h and mpio.h since they're included
from H5public.h.
./src/H5Cpublic.h
./src/H5C.c
Added H5Cset_stdio(), H5Cset_sec2(), H5Cset_core(),
H5Cset_split(), and H5Cset_family() in addition to the
H5Cset_mpio() that Kim and Albert already wrote. We still
need the H5Cget_driver() and an H5Cget...() counterpart for
each of those functions. The split and family drivers still
need a little work but I'm checking this in anyway.
