| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
1) Provide facilities in cache to allow us to avoid a potential cache
consistency bug in the parallel case.
2) Clean up a off by one sanity checking bug.
3) Turn off execution of long running tests in debug mode.
Description:
1) In the parallel case, all writes to metadata must be collective,
but reads may not be. In pricipal, this allows us to different
contents in different caches. This isn't a problem as long as the
correct data is always on disk, but unless we can force certain
writes immediately, that need not be the case.
2) & 3) should need no further explanation.
Solution:
1) Add code allowing us to mark cache entries, and then force
these entries to be flushed at a later time.
Note that to actually avoid the bug, we will have to modify
existing code to use these new features.
2) & 3) should need no further explanation.
Platforms tested:
heping (serial debug and production)
committest (copper, sol, and heping). test failed on heping in the
c++ portion of the build, but at Quincey's siggestion, I am proceeding
with the checkin.
Misc. update:
|
|
|
|
|
|
|
|
|
| |
Description: Test conversion from native integer to derived floating-point type and convert back;
test conversion from derived floating-point to derived floating-point types and convert back. Fixed
a few minor bugs related to type conversion in the library.
Platforms tested: h5committest and fuss
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Description: The functions for user-define floating-point type, like
H5Tset_fields, H5Tset_offset, H5Tset_precision, H5Tset_size, have some
minor bugs. For error checking, the library didn't include offset value
somehow.
Solution: Corrected those bugs.
Platforms tested: h5committest and fuss
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix
Description:
Correctly retire the H5E_LEN setting, now that the FORTRAN and C++ APIs
have been corrected to not use it either.
Solution:
Pass in the string buffer length for FORTRAN
In the C++ API, call H5Eget_msg() in a manner similar to the way
H5Fget_name() is called.
Platforms tested:
Linux 2.4 (heping) w/FORTRAN & C++
Solaris 2.7 (arabica) w/FORTRAN & C++
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix
Description:
Calling HDvsnprintf() several times (should be twice) is incorrect
without calling va_end() and va_start() again.
Solution:
Call va_end() and va_start() before calling HDvsnprintf() again.
Platforms tested:
Linux 2.4/x86_64 (mir)
Too minor to require h5committest
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix
Description:
Fortran and C++ API failed to compile because they need to use
H5E_LEN which was removed.
Solution:
Restored the H5E_LEN definition.
A permenant solution is still needed.
Platforms tested:
H5committested. (heping, sol, and copper all compiled fine again but
sol had an error in testerror.sh which seemed to be a different
problem.
Also tested in tg-NSCA.
|
|
|
|
|
|
|
|
|
|
| |
Description: Intel compiler on Linux has some problem to convert long double to
unsigned int correctly.
Solution: Detect the problem in configure and define a macro to skip this test
if it happens.
Platforms tested: eirene and fuss. Simple change.
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix
Description:
Fix possible overrun in error description string by allocating large enough
string on the fly.
Platforms tested:
FreeBSD 4.10 (sleipnir)
Too minor to require h5committest
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
cleanup warning messages.
Description:
Compilers complained about local variable shadowed global variables.
The global variables d[], m[], nd, na were having the same names
as local function arguments. This was a left over when I broke the
original one big chunk of code into separated functions to go around
the Intel Compiler optimization error.
Solution:
Appended the global variables with suffix _g, thus
d[] => d_g[]
m[] => m_g[]
nd => nd_g
na => na_g
This is only an intermediate step. After the changes tested
out fine by all machines, need to rename the variables to more
meaningful names like dtype_array.
Platforms tested:
Tested in heping by saving the previous generated H5Tinit.c,
regenerated a new H5Tinit.c with the changed code, finally
compared the new and old versions were identical.
|
|
|
|
|
|
|
|
|
|
|
| |
Description: For FreeBSD (sleipnir), when GNU compilers do conversion from
unsigned long long to long double, the last 2 bytes of mantissa are lost.
The impact of precision loss isn't significant.
Solution: Detect this case on FreeBSD in configure, ignore it in dtypes.c
test instead of return failure.
Platforms tested: sleipnir, fuss, modi4. These systems are mainly concerned.
|
|
|
|
|
|
|
|
|
|
| |
Code cleanup
Description:
Remove obsolete support for Watcom C compiler.
Platforms tested:
None - too minor to require any.
|
|
|
|
|
|
|
|
|
|
| |
Description: For HP-UX 11.00, the compiler generates 'floating exception'
when converting 'long double' to most of integer types.
Solution: Define a macro for all other systems except HP-UX 11.00. Hard set
this macro to 'no' in config/hpux11.00 to skip this test for HP-UX 11.00.
Platforms tested: modi4, kelgia, fuss
|
|
|
|
|
|
|
|
|
|
| |
Description:
Solution:
Platforms tested:
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix
Description:
Print the thread ID in a little more portable of a fashion, disallowing
negative thread IDs.
Platforms tested:
Linux 2.4 (heping) w/threadsafe
Too Minor to require h5committest
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Remove feature
Description:
Retire threaded, balanced binary tree code from HDF5 use. Requiescat in
pace...
Also, regenerate dependencies files.
Platforms tested:
FreeBSD 4.10 (sleipnir) w/parallel
Too minor to require full h5committesting (the code is already
disconnected from everything except its tests)
|
|
|
|
|
|
|
|
|
| |
Description: 2 small things to correct:
1. the change in H5config.h.in was wept out by a later change.
2. long long isn't supported on Windows. It's long_long instead.
Platforms tested: fuss(linux 2.4)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Code cleanup
Description:
Transition FPHDF5 to use skip lists for various internal data structures,
instead of threaded, balanced binary trees.
Also, simplify some of FPHDF5 code some, etc.
Platforms tested:
FreeBSD 4.10 (sleipnir) w/parallel & fphdf5
Too minor to require h5committest
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
New feature
Description:
Add some additional features to the skip list code that was needed to
fully support all the features that the threaded, balanced binary tree code
has.
Also, updated the property list code to take advantage of a few of the
new features.
Platforms tested:
FreeBSD 4.10 (sleipnir)
Too minor to require h5committest
|
|
|
|
|
|
|
|
|
|
| |
Code cleanup
Description:
Change pablo mask to conform to the style used by the rest of the library
Platforms tested:
None needed - very, very minor
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix.
Description:
threadsafe test (ttsafe) would seg-fault if --disable-production --enable-debug
mode is used. Reason was that the fstack.nused field was not initialized
when created.
Solution:
Init new fstack.nused with 0.
Platforms tested:
Tested in mir, using development/debug mode.
Did not "h5committested" as change is trivial and limited to threadsafe mode.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
To add n-bit implementation inside HDF5.
Description:
This check-in for implementation is for debugging purpose.
N-bit filter should be used for the time being.
Solution:
Platforms tested:
sol 2.7, Linux 2.4, Aix 5.1
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Adding framework of N-bit filter
Description:
N-bit filter is required for NetCDF project.
To add N-bit filter, configure.in configure, Makefile.in under src and other
filter-related source code needs to be updated.
Currently, N-bit tests are turned off from the library. So the change will affect daily test.
Solution:
Platforms tested:
sol 2.7, linux 2.4, aix 5.1
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Description: Start to support software conversion between long double and
all integers.
Solution: No major changes to the algorithm. Changes to configure is to
exclude SGI for long double to integers test because there're too many problems
in their compiler.
Platforms tested: h5committest, modi4, fuss, Teragrid, arabica
Misc. update: RELEASE.txt
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix.
Description:
When building H5Z.c with Visual C++ 6.0, I got the following warning message:
H5Z.c(1241) : warning C4013: 'SZ_encoder_enabled' undefined; assuming extern returning int
Solution:
The header file szlib.h should be included in H5Z.c in order to use SZ_encoder_enabled() function.
Add the following script into H5Z.c:
#ifdef H5_HAVE_SZLIB_H
# include "szlib.h"
#endif
Platforms tested:
Heping (Linux)
Visual C++ 6.0 on Windows XP/2000.
.NET on Windows XP.
(Tested with SZIP-with-encoder and SZIP-without-encoder).
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
| |
Code cleanup
Description:
Convert chunk iteration code to use skip lists instead of threaded, balanced
binary trees.
Platforms tested:
FreeBSD 4.10 (sleipnir) w/parallel & szip
Too minor to require h5committest
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Code cleanup (sorta)
Description:
Transition the generic property list code from using the threaded, balanced
binary tree code (H5TB<foo>() routines) to use skip lists (H5SL<foo>() routines)
for internal management of properties, etc.
Platforms tested:
FreeBSD 4.10 (sleipnir)
Too minor to require h5committest
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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:
|
|
|
|
|
|
|
| |
Add new file
Description:
Add new PC-Lint/Flexelint configuration files
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fix a bug in the cache caused by a slightly over active sanity check.
Description:
When the #define H5C_DO_SANITY_CHECKS is TRUE, the macro
H5C__DLL_PRE_INSERT_SC is executed prior to inserting an
entry in the LRU list. The macro performs a variety of
sanity checks, and flags an error if any of the checks
fail.
Prior to this update, the macro used to check to see if
the target list had length 1 and size <= 0. The new
epoch marker entries used in the age out cache size
reduction algorithm have size zero -- making it possible
for this sanity check to fail incorrectly.
Note that cache sanity checks are disabled in the CVS
version of the code, so this bug and bug fix should be
invisible unless you are working with the cache, and
turn the sanity checks on.
Solution:
Removed the offending clause in H5C__DLL_PRE_INSERT_SC. Since
the size used in the macro is typically a size_t and thus
cannot have negative value, there was no point in changing
it to "size < 0".
Platforms tested:
Ran a serial test on heping.
Under the circumstances, I didn't feel the need for further
testing.
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Modify the cache code (H5C) to support automatic cache resizing to
adapt to the work load at run time.
Description:
Different applications require different sized caches to maintain
an acceptable hit rate. This set of changes attempts to provide the
ability to adjust to circumstances automatically.
Solution:
Added highly configurable code to allow the user to either set a
fixed cache size, or allow the cache to grow and shrink according to
conditions.
If enabled, cache size increases are triggered when the hit rate
drops below a user specified threshold in a user specified interval.
Cache size reductions (if enabled) are triggered when either the
hit rate exceeds some user specified threshold over a user specified
interval, when the cache contains "enough" entries that haven't been
accessed for a user specified interval, or some mix of the above.
See the header comments on the H5C_auto_size_ctl_t structure in
H5Cprivate.h for further details.
At present, the cache resize configuration options are not
accessible via the user API. Must add this.
Platforms tested:
h5committested, heping (serial), and copper (parallel)
Misc. update:
|
|
|
|
|
|
|
|
|
|
| |
Fix whitespace
Description:
Clean up whitespace a bit
Platforms tested:
None, just eyeballed, very minor
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
small bug fix
Description:
When checking whether the current chunk covers the irregular hyperslab,
the ending point of the chunk is not updated.
This will cause the wrong checking output and fail for some irregular hyperslab
selection.
Solution:
Updating the ending point of the chunk.
Platforms tested:
Linux 2.4 + parallel (too small to use h5committest)
Misc. update:
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Description: In file H5FDsec2.c and H5FDlog.c, there's statements like
H5_ASSIGN_OVERFLOW(file->eof,sb.st_size,off_t,haddr_t);
It assumes sb.st_size from h5_stat_t is of type off_t. But on Windows, it
has type __int64. So the H5_ASSIGN_OVERFLOW statement may cause problem.
Instead of trying to do switch between Windows and other systems in H5FDsec2.c
and H5FDlog.c, define a substituting type in H5private.h to fix the problem.
On Windows, this type is __int64; on other systems, it is off_t.
Platforms tested: fuss. Simple fix. Already did h5committest for v1.6.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
bug fix
Description:
The irregular hyperslab span_tree implementation inside
H5S_hyper_add_span_element_helper updating the high bound of the span, but forget updating the number of element of the span.
New routine obtain_derived_datatype catches the bug.
Solution:
Updating the number of element of the span as well.
Platforms tested:
Linux 2.4 and AIX 5.1, both with parallel enabled.
Misc. update:
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
New feature
Description:
Add callback iteration routine to skip lists.
Platforms tested:
FreeBSD 4.10 (sleipnir) w/parallel
Too minor to require h5committest
|
|
|
|
|
|
|
|
|
|
|
| |
New feature
Description:
Allow skip list keys to be strings (needed for property list support)
Platforms tested:
FreeBSD 4.10 (sleipnir)
Too minor to require h5committest
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
New internal feature
Description:
Add a "release" routine to the skip lists, which releases all the nodes
in the skip list, but keeps the skip list active.
Platforms tested:
FreeBSD 4.10 (sleipnir)
Linux 2.4 (verbena)
Too minor to require h5committest
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Code optimization
Description:
Retarget metadata cache code from using TBBT routines to using new skiplist
routines, for a reasonable speedup (when dealing with dirty metadata)
Platforms tested:
FreeBSD 4.10 (sleipnir) w/parallel
Solaris 2.7 (arabica)
Too minor to require h5committest
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix
Description:
Re-work array of 'forward' pointers in a way that makes the PGI compiler
more happy.
Solution:
Change from 'H5SL_node_t *forward[1]' to 'H5SL_node_t **forward' and
change how array of forward pointers is dealt with.
Platforms tested:
Linux 2.4 (verbena) w/PGI
FreeBSD 4.10 (sleipnir)
Configuration not tested w/h5committest
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix
Description:
Correct off-by-on error which could cause invalid # of levels to be used
for a skip list node, in rare circumstances.
Change from srandom()/random() to more portable srand()/rand() routines
for random # generation.
Pre-compute the probability factor for computing the level of new nodes,
for a minor speedup.
Platforms tested:
FreeBSD 4.10 (sleipnir)
IRIX64 6.5 (modi4)
Solaris 2.8 (sol)
AIX 5.1(?) (copper)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Code optimization
Description:
Begin de-orbital burn of threaded, balanced binary tree code use by
switching file open object code over to using them. Initial benchmarks show
that skip lists are 5-10x faster than our TBBT code...
Platforms tested:
FreeBSD 4.10 (sleipnir) w/parallel
Solaris 2.7 (arabica)
Too minor to require h5committest
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Code optimization
Description:
Revise mechanism for calling H5open/H5check to better avoid calling them
from internal libray code.
Only copy a datatype's group entry information if it has some.
Platforms tested:
FreeBSD 4.10 (sleipnir) w/parallel
Solaris 2.7 (arabica)
Too minor to require h5committest
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add new internal data structure
Description:
Add an implementation of skip lists to the library (see comment in
src/H5SL.c for references to the papers describing them) as a potential
replacement for our current threaded, balanced binary tree container.
Skip lists are much simpler to implement and should be faster to use.
Also, added new error codes to release branch, so bump the minor version
number to indicate that the library is no longer perfectly compatible with
the 1.6.3 release.
Platforms tested:
FreeBSD 4.10 (sleipnir) w/parallel
Solaris 2.7 (arabica)
Too minor to require further testing (the skip lists aren't actually
used by any library code yet)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Adding code for using MPI derived datatype to handle collective IO
Description:
No testing yet, won't affect the library.
Solution:
Platforms tested:
linux 2.4 + mpich 1.2.6
Aix 5.1 + mpcc_r
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Code optimization
Description:
Change how default allocation time is handled internally to the library,
to avoid some performance issues with property lists.
Platforms tested:
FreeBSD 4.10 (sleipnir) w/parallel
Solaris 2.7 (arabica)
Too minor to require h5committest
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Code optimization
Description:
Switch a few more malloc/free pairs over to using internal free list code,
to avoid abusing system memory allocator as badly.
Platforms tested:
FreeBSD 4.10 (sleipnir) w/parallel
Solaris 2.7 (arabica)
Too minor to require h5committest
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Code cleanup & optimization
Description:
Improve ADF/CGNS benchmark by reducing the number of internal attribute
copies made during creations, opens and writes.
Added new H5O_iterate() routine for iterating through messages of a certain
type in the object header (attributes are the only message currently that can
have multiple instances in the object header).
Cross-pollinated various minor code cleanups to reduce diffs between
branches.
Platforms tested:
FreeBSD 4.10 (sleipnir) w/parallel
Solaris 2.7 (arabica)
Too minor to require h5committest
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Code optimization
Description:
Avoid making as many copies of attribute information.
Also, be smarter about which properties we've seen before when copying
and closing property lists.
Fix memory leak of attribute data structures.
Platforms tested:
FreeBSD 4.10 (sleipnir) w/parallel
Solaris 2.7 (arabica)
Too minor to require h5committest
|