Code optimization
Description:
Further reduce the number of copies we make of a hyperslab selection for
chunked I/O, especially when we are only going to throw the old selection away
for a new one.
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir) w/parallel
Bug fixes
Description:
Updated dependencies
Fixed error with C++ compiler builds of main library
Added H5Pset_data_transform to MPE info
Platforms tested:
FreeBSD 4.9 (sleipnir)
Code cleanup
Description:
Reset pointers to heap info and actual heap after we have unlocked them,
to avoid accidentally referencing them after that point.
Platforms tested:
FreeBSD 4.9 (sleipnir)
too minor to require h5committest
Bug fix.
Description:
Fixed token_type enum struct in H5Ztrans.c to compile under windows.
Solution:
Prefixed all values in the struct with H5Z_XFORM_ to prevent name collisions
with pre-defined values.
Platforms tested:
h5committest'ed, but on arabica + no sol.
Bug fix.
Description:
Fixed handling of trivial data transform expressions (like 5/3 + 3) and some
data coversion fixes. Also added more tests to dtransform.c
Solution:
Added some more checks in the H5Z_xform_reduce_tree function to see if perhaps
the transform expression is complicated and is a non-trivial reduction.
Added tests for data conversion to dtransform as well as tests for a trivial
data transform expression.
Platforms tested:
h5committest'ed, except used arabica instead of sol and didn't do on copper
b/c no logon there. Problem noted with mtime test...doesn't appear to be
related to anything having to do with data transforms.
Description: H5T_reverse_order was declared as private function.
Solution: Change it to static local function in H5Tconv.c since it's not
used by any function in other file.
Platforms tested: RH 8(fuss)
Code cleanup
Description:
Refactored data transform code to reduce amount of symbols in the global
scope and also cleaned up & simplified the code a bit.
Platforms tested:
h5committest (minus copper, plus serial modi4)
FreeBSD 4.9 (sleipnir) w & w/o parallel
New Feature
Description:
Add the data transform function, H5Pset_transform().
Platforms tested:
"h5committested".
Copper was down. Ran parallel tests in sol instead.
Misc. update:
Cleanup source.
Description:
H5DZtrans.c is renamed as H5Ztrans.c.
H5DZ.c is not used any more.
Platforms tested:
No need since removing files only.
Misc. update:
Code optimization
Description:
Improved LRU algorithm for locating blocks of the correct size in the
block free-list
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir) w/parallel
too minor for h5committest
Code optimization
Description:
Reduce amount of information allocated/freed for chunk cache.
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir) w/paralell)
too minor to require h5committest
Code optimization
Description:
Eliminate the B-tree "split_ratios" as a parameter and pull it from the
DXPL instead.
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir) w/parallel
too minor to require h5committest
Fixing building Fortran DLL on windows
Description:
Fortran DLL needs to separate C stub DLL from Fortran DLL.
So we need to use another keyword H5_FCDLL to replace the old H5_DLL for
C stub library.
Solution:
Add another section at H5api_adpt.h to define a macro block for Fortran
C stub library.
Platforms tested:
Manual tests.
Copper is not accessible. Only test on arabica and eirene with fortran enabled.
Misc. update:
Code optimization
Description:
Eliminate redundant B-tree lookups for chunk information.
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir) w/parallel
too minor to require h5committest
Code optimization
Description:
Reduce the number of times the number of elements in a selection is
computed.
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir) w/parallel
too minor to require h5committest
Code optimization
Description:
Move the element size for the selection into the selection iterator instead
of always passing it as a parameter.
Also, eleminate another 64-bit multiply for "all" selections.
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir)
too minor to require h5committest
Code optimization
Description:
Stop straddling the line and always use elements instead of bytes to
determine how how I/O to perform on a selection.
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir)
too minor to require h5committest
Code optimization
Description:
Eliminate more 64-bit multiplies by remebering the size of contiguous
datasets as well as chunked datasets.
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir)
too minor to require h5committest
Code optimization
Description:
Eliminate more redundant 64-bit multiplies
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir)
too minor to require h5committest
Code optimization
Description:
Eliminate frivolous 64-bit multiply.
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir)
too minor to require h5committest
Code optimization
Description:
Remove extraneous memcpy()
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir)
too minor to require h5committest
Code optimization
Description:
Remove redundant error checking.
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir)
too minor to require h5committest
Code optimization
Description:
Avoid dividing the chunk coordinates at the top levels of the chunk I/O
routines, only to multiply them at the bottom of the routines.
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir)
too minor to require h5committest
Code cleanup/optimization
Description:
Don't clear the error stack when calling H5open(), it's unnecessary.
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir)
too minor to require h5committest
Code cleanup
Description:
Refactor code to eliminate duplication of setup & calls to H5B_find to
locate information about a chunk on disk.
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir)
too minor to require h5committest
Code optimization
Description:
Hoist some if statements out of a loop in vector comparision routines that
are called very frequently by chunking code.
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir)
too minor to require h5committest
Code optimization.
Description:
Use free-list pool of blocks for allocating chunk buffers, when there are
no filters in the pipeline. The avoids calling malloc/free so much.
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir)
too minor to require h5committest
Bug fix (sorta)
Description:
Add hack to allow the MS Visual Studio 6 compiler to build the library.
It cannot cast unsigned long long values to float or double values. So, add
another configuration macro to disable this conversion in the library. Just
the "hardware" conversion is disabled, so the library will still correctly
convert unsigned long long to float and double values, it will just happen
more slowly with the "software" conversion routine.
Platforms tested:
FreeBSD 4.9 (sleipnir) with "Windows" setting faked
inappropriate for h5committest
Description: H5Pset_type_conv_cb and H5Pget_type_conv_cb weren't put into
H5MPprivate.h. Another thing is the func type doesn't match the func field
in the struct in these functions.
Platforms tested: RH 8(fuss) and sleipnir
Bug fix (sorta)
Description:
The SGI machines have problems accurately (and consistently) converting
unsigned long values to float and double values, so put in a bit of a hack in
the datatype conversion test code to allow them to get "close enough". This
hack is enabled at configure time by a flag which should only be set on machines
with this problem.
Platforms tested:
FreeBSD 4.9 (sleipnir)
h5committest
Undo the mysterous changes.
Description:
Somehow something changed this file during snapshot release.
Undo the change.
Platforms tested:
No test since it is a simple editing.
Misc. update:
Code optimization
Description:
Instead of dynamicly allocating various arrays for various pieces of
information about a selection or selection iterator, just use fixed size
array of size H5S_MAX_RANK (as the rest of the library does).
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir)
h5committest
Code optimization
Description:
Make calculation of "all" selection sequence length information more
efficient.
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir)
h5comittest
Code optimization
Description:
Avoid clearing the error stack unless necessary.
Platforms tested:
Solaris 2.7 (arabica)
FreeBSD 4.9 (sleipnir)
too minor to require h5committest
Checkpoint checkin of FP bug fixes. FP is still quite
buggy, but I must go deal with other matters.
Description:
Fixed two major bugs:
1) H5FPserver.c was clobbering meta data in its care.
2) H5FPserver.c was allocating the same space multiple
times, causing both data and meta data corruption.
Also made minor fixes, added debugging code, and familiarized
myself with the FP code.
All development work with FP enabled was done on Eirene.
On this platform, FP now passes its test reliably with
up to 9 processes. At 10 processes it seg faults every
time. I haven't looked into this issue.
There are also several known locking bugs which have to
be fixed. However, they are of sufficiently low probability
that I didn't bother with them on this pass.
FP has not been tested with deletions -- this should be
done.
Also, need to test FP chunked I/O.
Solution:
1) Modified cache in H5FPserver.c to merge changes correctly.
Found and fixed a bug in H5TB.c in passing.
2) Multiple space allocation was caused by a race condition
with set eoa requests.
Most of these eoa requests appeared to be superfluous, so
I deleted them.
Those issued during the superblock read seemed necessary,
so I inserted a barrier at the end of the superblock read,
to prevent races with allocations.
Platforms tested:
h5committested
Description:
Fixed bug in H5TB_less(). It was returning the next largest node,
instead of the next smallest as advertised.
Added comments warning that H5TB_rem()s will occasionally delete a
node other than the one provided in its argument list.
Solution:
It was sufficient t invert two comparisons in H5TB_less().
Platforms tested:
h5committested
Code cleanup
Description:
Clean up lots of warnings based on those reported from the SGI compilers
as well as gcc.
Platforms tested:
SGI O3900, IRIX64 6.5 (Cheryl's SGI machine)
FreeBSD 4.9 (sleipnir) w/ & w/o parallel
h5committest
Code cleanup
Description:
Update null dataspace changes to try to write older version of dataspace
information whenever possible.
Refactor common code to only one location.
Allow I/O operations to succeed on null dataspaces.
Platforms tested:
FreeBSD 4.9 (sleipnir)
h5committest
