| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
| |
h5committested.
|
| |
|
|
| |
h5committested.
|
| |
|
|
| |
h5committested.
|
| |
|
|
| |
h5committested.
|
| |
|
|
| |
h5committested.
|
| |
|
|
| |
h5committested.
|
| |
|
|
| |
h5committested.
|
| |
|
|
| |
h5committested.
|
| |
|
|
| |
h5committested.
|
| |
|
|
| |
h5committested.
|
| |
|
|
| |
h5committested.
|
| |
|
|
| |
h5committested.
|
| |
|
|
| |
h5committested.
|
| |
|
|
| |
h5committested.
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
| |
ended the program.
Solution: added exit(0) after usage(). This is not a good fix but will work
for now.
Test: jam.
|
| |
|
|
|
|
|
|
| |
The script execute the program by just $program which will not work if the
user does not have "." in the $PATH.
Changed it to "./$program" to avoid this problem.
Tested: koala.
|
| |
|
|
|
|
| |
1) H5O_load() in H5Ocache.c: when reading a block that is > spec_read siez, read the whole block in again and possibly decode the header.
2) H5F_accum_write() in H5Faccum.c: for a large write that is >= H5F_ACCUM_MAX_SIZE, flush the metadata in the accumulator first before the write.
Tests are added to test/ohdr.c and test/accum.c. h5committested.
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
| |
Replace H5Fflush in writer with H5Dflush and file close/reopen with
H5Drefresh. Use cases seem to run faster. Keeping the old code so that
later on, I may test the performance between the two.
Tested: koala.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Implemented use case 1.9: Appending n-1 dimensional planes
Adding the option "-y" which allows chunks to be thicker
(more than 1 plane). This supports use case 1.9 which have chunks
of multiple planes but writing is still appending one plane at a time.
Using -y with "use_append_chunk", would provide the use case
of writing a plane to a partial chunk.
Using -y with "use_append_mchunks", would provide the use case
of writing a plane to multiple partial chunks.
Also added the option -n which specifies how many planes to write
instead of the defaults of chunksize number of planes. This allows
the final dataset to be other than a cube.
test/use_append_mchunks.c:
test/use_common.c:
test/use.h:
test/use_append_chunk.c:
I also overhauled code in use_append_mchunks and use_append_chunk:
1. creating a set_parameter() to customize individual use case;
2. combine the create_uc_file(), read_uc_file() and write_uc_file()
of both use cases appending one plane per chunks and one plane per
multiple chunks. Moved these combined create/read/write functions
into use_common.c to make future modification and maintence easier.
test/test_usecases.sh.in:
Added a simple test to demonstarte how to use -y to run use case 1.9.
Tested: h5committest, except cmake, passed. Hand tested in Koala for various options.
|
| |
|
|
| |
read-write operation on the system.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Implemented use case 1.8 program--SWRM read/write multiple chunks at a time.
test/use_append_mchunks.c:
The program.
test/Makefile.am:
test/Makefile.in:
Added program use_append_mchunks.
test/test_usecases.sh.in:
Added simple tests for use_append_mchunks.
MANIFEST:
Updated for new file.
Tested: h5committest, except cmake, passed. Hand tested in Koala for various options.
|
| |
|
|
| |
Tested: h5committested.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When zlib is not available (use "configure --without-zlib"), some tests
failed to not using compression features.
Solution:
test/dsets.c:
Conditional blocked out use of zlib compression code.
tools/h5ls/testh5ls.sh.in:
Temporary commented out test code that uses datafile that has zlib
compressed data.
Tested: h5committest (cmake does not work but that is ignored for now.) Also,
hand tested in koala using "configure --without-zlib".
|
| | |
|
| | |
|
| | |
|
| |
|
|
| |
Tested: koala.
|
| |
|
|
| |
Tested: h5committest.
|
| |
|
|
|
|
|
|
|
|
| |
usecase1_7.c -> use_append_chunk;
usecase_common.c -> use_common.c;
usecase.h -> use.h.
MANIFEST updated with new names.
Tested: h5committest.
|
| |
|
|
|
|
|
|
|
|
|
| |
1. "-f filename" can create the test file somewhere other than
the current directory or <progname>.h5. This allows running tests in different
filesystems, for example.
2. "-l w|r" can launch only the writer or the reader (default does both).
This allows launching writer (includes the "create file") in one process and
launching the reader somewhere else. (The "-f" helps too.)
Tested: h5committested.
|
| |
|
|
| |
Tested: h5committest.
|
| |
|
|
|
| |
Resort the hl/test/* entries to be alphabetical order, easier to detect
duplicates like this.
|
| |
|
|
| |
h5committtested.
|
| |
|
|
| |
h5committested.
|
| |
|
|
| |
h5committested.
|
| |
|
|
| |
h5committested.
|
| |
|
|
| |
h5committested.
|
| |
|
|
| |
h5committested.
|
| |
|
|
| |
h5committested.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
Descriptions:
When removing object header messages, it is possible for object header
continuation messages to move to a different chunk. When this happens, flush
dependencies need to be updated to reflect the new structure. This change adds
code to update the flush dependencies, and a test for this.
Also fixed a bug where the flush dependency no the object header proxy was not
being destroyed when an object header chunk was deleted.
Tested: ummon
|
| |
|
|
| |
Tested: ummon
|
| |
|
|
|
|
|
|
|
|
|
| |
Description:
Added flush dependencies between object header chunks. Also added a new object
header proxy object, which is a child dependency of all object header chunks.
All objects which are child dependencies of an object header should set up a
flush dependency on the object header proxy, so the dependency applies to all
object header chunks.
Tested: ummon; jam (fails ph5diff -v, unrelated), koala, ostrich (h5committest)
|
| |
|
|
| |
Tested: ummon
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
multiple parents
Description:
Reworked how flush dependencies worked internally, allowing multiple flush
dependency parents, and removing the notion of flush dependency heights, instead
keeping track of the number of dirty descendents to determine if parents can be
flushed. Also removed the requirement that cache clients destroy flush
dependencies before eviction (this is now handled by the cache) and removed the
maximum number of passes (the cache should detect infinite loops elsewhere).
Added test cases for this.
Tested: durandal
|
| |
|
|
|
|
|
| |
Add missing skip list header to H5AC.c
Tested on:
jam (trivial change)
|
