2003-04-01 01:59:04 +08:00
|
|
|
|
/* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
|
|
|
|
|
* Copyright by the Board of Trustees of the University of Illinois. *
|
|
|
|
|
* All rights reserved. *
|
|
|
|
|
* *
|
|
|
|
|
* This file is part of HDF5. The full HDF5 copyright notice, including *
|
|
|
|
|
* terms governing use, modification, and redistribution, is contained in *
|
|
|
|
|
* the files COPYING and Copyright.html. COPYING can be found at the root *
|
|
|
|
|
* of the source code distribution tree; Copyright.html can be found at the *
|
|
|
|
|
* root level of an installed copy of the electronic HDF5 document set and *
|
|
|
|
|
* is linked from the top-level documents page. It can also be found at *
|
|
|
|
|
* http://hdf.ncsa.uiuc.edu/HDF5/doc/Copyright.html. If you do not have *
|
|
|
|
|
* access to either file, you may request a copy from hdfhelp@ncsa.uiuc.edu. *
|
|
|
|
|
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
|
|
|
|
|
|
1998-07-31 00:38:21 +08:00
|
|
|
|
/*
|
|
|
|
|
* Programmer: Robb Matzke <matzke@llnl.gov>
|
|
|
|
|
* Thursday, July 30, 1998
|
|
|
|
|
*
|
|
|
|
|
* Purpose: Determines if the modification time message is working
|
|
|
|
|
* properly. Specifically, the code in H5O_mtime_decode() is
|
|
|
|
|
* very OS-dependent and this test tries to figure out if it's
|
|
|
|
|
* working properly.
|
|
|
|
|
*/
|
2001-04-04 02:09:16 +08:00
|
|
|
|
#include "h5test.h"
|
1998-11-03 01:58:28 +08:00
|
|
|
|
|
1998-11-24 04:40:35 +08:00
|
|
|
|
const char *FILENAME[] = {
|
|
|
|
|
"mtime",
|
|
|
|
|
NULL
|
|
|
|
|
};
|
1998-07-31 00:38:21 +08:00
|
|
|
|
|
[svn-r6252] Purpose:
Lots of performance improvements & a couple new internal API interfaces.
Description:
Performance Improvements:
- Cached file offset & length sizes in shared file struct, to avoid
constantly looking them up in the FCPL.
- Generic property improvements:
- Added "revision" number to generic property classes to speed
up comparisons.
- Changed method of storing properties from using a hash-table
to the TBBT routines in the library.
- Share the propery names between classes and the lists derived
from them.
- Removed redundant 'def_value' buffer from each property.
- Switching code to use a "copy on write" strategy for
properties in each list, where the properties in each list
are shared with the properties in the class, until a
property's value is changed in a list.
- Fixed error in layout code which was allocating too many buffers.
- Redefined public macros of the form (H5open()/H5check, <variable>)
internally to only be (<variable>), avoiding innumerable useless
calls to H5open() and H5check_version().
- Reuse already zeroed buffers in H5F_contig_fill instead of
constantly re-zeroing them.
- Don't write fill values if writing entire dataset.
- Use gettimeofday() system call instead of time() system when
checking the modification time of a dataset.
- Added reference counted string API and use it for tracking the
names of objects opening in a file (for the ID->name code).
- Removed redundant H5P_get() calls in B-tree routines.
- Redefine H5T datatype macros internally to the library, to avoid
calling H5check redundantly.
- Keep dataspace information for dataset locally instead of reading
from disk each time. Added new module to track open objects
in a file, to allow this (which will be useful eventually for
some FPH5 metadata caching issues).
- Remove H5AC_find macro which was inlining metadata cache lookups,
and call function instead.
- Remove redundant memset() calls from H5G_namei() routine.
- Remove redundant checking of object type when locating objects
in metadata cache and rely on the address only.
- Create default dataset object to use when default dataset creation
property list is used to create datasets, bypassing querying
for all the property list values.
- Use default I/O vector size when performing raw data with the
default dataset transfer property list, instead of querying for
I/O vector size.
- Remove H5P_DEFAULT internally to the library, replacing it with
more specific default property list based on the type of
property list needed.
- Remove redundant memset() calls in object header message (H5O*)
routines.
- Remove redunant memset() calls in data I/O routines.
- Split free-list allocation routines into malloc() and calloc()-
like routines, instead of one combined routine.
- Remove lots of indirection in H5O*() routines.
- Simplify metadata cache entry comparison routine (used when
flushing entire cache out).
- Only enable metadata cache statistics when H5AC_DEBUG is turned
on, instead of always tracking them.
- Simplify address comparison macro (H5F_addr_eq).
- Remove redundant metadata cache entry protections during dataset
creation by protecting the object header once and making all
the modifications necessary for the dataset creation before
unprotecting it.
- Reduce # of "number of element in extent" computations performed
by computing and storing the value during dataspace creation.
- Simplify checking for group location's file information, when file
has not been involving in file-mounting operations.
- Use binary encoding for modification time, instead of ASCII.
- Hoist H5HL_peek calls (to get information in a local heap)
out of loops in many group routine.
- Use static variable for iterators of selections, instead of
dynamically allocation them each time.
- Lookup & insert new entries in one step, avoiding traversing
group's B-tree twice.
- Fixed memory leak in H5Gget_objname_idx() routine (tangential to
performance improvements, but fixed along the way).
- Use free-list for reference counted strings.
- Don't bother copying object names into cached group entries,
since they are re-created when an object is opened.
The benchmark I used to measure these results created several thousand
small (2K) datasets in a file and wrote out the data for them. This is
Elena's "regular.c" benchmark.
These changes resulted in approximately ~4.3x speedup of the
development branch when compared to the previous code in the
development branch and ~1.4x speedup compared to the release
branch.
Additionally, these changes reduce the total memory used (code and
data) by the development branch by ~800KB, bringing the development
branch back into the same ballpark as the release branch.
I'll send out a more detailed description of the benchmark results
as a followup note.
New internal API routines:
Added "reference counted strings" API for tracking strings that get
used by multiple owners without duplicating the strings.
Added "ternary search tree" API for text->object mappings.
Platforms tested:
Tested h5committest {arabica (fortran), eirene (fortran, C++)
modi4 (parallel, fortran)}
Other platforms/configurations tested?
FreeBSD 4.7 (sleipnir) serial & parallel
Solaris 2.6 (baldric) serial
2003-01-10 01:20:03 +08:00
|
|
|
|
#define TESTFILE1 "tmtimeo.h5"
|
2003-06-14 04:30:43 +08:00
|
|
|
|
#define MTIME1 1055531866
|
[svn-r6252] Purpose:
Lots of performance improvements & a couple new internal API interfaces.
Description:
Performance Improvements:
- Cached file offset & length sizes in shared file struct, to avoid
constantly looking them up in the FCPL.
- Generic property improvements:
- Added "revision" number to generic property classes to speed
up comparisons.
- Changed method of storing properties from using a hash-table
to the TBBT routines in the library.
- Share the propery names between classes and the lists derived
from them.
- Removed redundant 'def_value' buffer from each property.
- Switching code to use a "copy on write" strategy for
properties in each list, where the properties in each list
are shared with the properties in the class, until a
property's value is changed in a list.
- Fixed error in layout code which was allocating too many buffers.
- Redefined public macros of the form (H5open()/H5check, <variable>)
internally to only be (<variable>), avoiding innumerable useless
calls to H5open() and H5check_version().
- Reuse already zeroed buffers in H5F_contig_fill instead of
constantly re-zeroing them.
- Don't write fill values if writing entire dataset.
- Use gettimeofday() system call instead of time() system when
checking the modification time of a dataset.
- Added reference counted string API and use it for tracking the
names of objects opening in a file (for the ID->name code).
- Removed redundant H5P_get() calls in B-tree routines.
- Redefine H5T datatype macros internally to the library, to avoid
calling H5check redundantly.
- Keep dataspace information for dataset locally instead of reading
from disk each time. Added new module to track open objects
in a file, to allow this (which will be useful eventually for
some FPH5 metadata caching issues).
- Remove H5AC_find macro which was inlining metadata cache lookups,
and call function instead.
- Remove redundant memset() calls from H5G_namei() routine.
- Remove redundant checking of object type when locating objects
in metadata cache and rely on the address only.
- Create default dataset object to use when default dataset creation
property list is used to create datasets, bypassing querying
for all the property list values.
- Use default I/O vector size when performing raw data with the
default dataset transfer property list, instead of querying for
I/O vector size.
- Remove H5P_DEFAULT internally to the library, replacing it with
more specific default property list based on the type of
property list needed.
- Remove redundant memset() calls in object header message (H5O*)
routines.
- Remove redunant memset() calls in data I/O routines.
- Split free-list allocation routines into malloc() and calloc()-
like routines, instead of one combined routine.
- Remove lots of indirection in H5O*() routines.
- Simplify metadata cache entry comparison routine (used when
flushing entire cache out).
- Only enable metadata cache statistics when H5AC_DEBUG is turned
on, instead of always tracking them.
- Simplify address comparison macro (H5F_addr_eq).
- Remove redundant metadata cache entry protections during dataset
creation by protecting the object header once and making all
the modifications necessary for the dataset creation before
unprotecting it.
- Reduce # of "number of element in extent" computations performed
by computing and storing the value during dataspace creation.
- Simplify checking for group location's file information, when file
has not been involving in file-mounting operations.
- Use binary encoding for modification time, instead of ASCII.
- Hoist H5HL_peek calls (to get information in a local heap)
out of loops in many group routine.
- Use static variable for iterators of selections, instead of
dynamically allocation them each time.
- Lookup & insert new entries in one step, avoiding traversing
group's B-tree twice.
- Fixed memory leak in H5Gget_objname_idx() routine (tangential to
performance improvements, but fixed along the way).
- Use free-list for reference counted strings.
- Don't bother copying object names into cached group entries,
since they are re-created when an object is opened.
The benchmark I used to measure these results created several thousand
small (2K) datasets in a file and wrote out the data for them. This is
Elena's "regular.c" benchmark.
These changes resulted in approximately ~4.3x speedup of the
development branch when compared to the previous code in the
development branch and ~1.4x speedup compared to the release
branch.
Additionally, these changes reduce the total memory used (code and
data) by the development branch by ~800KB, bringing the development
branch back into the same ballpark as the release branch.
I'll send out a more detailed description of the benchmark results
as a followup note.
New internal API routines:
Added "reference counted strings" API for tracking strings that get
used by multiple owners without duplicating the strings.
Added "ternary search tree" API for text->object mappings.
Platforms tested:
Tested h5committest {arabica (fortran), eirene (fortran, C++)
modi4 (parallel, fortran)}
Other platforms/configurations tested?
FreeBSD 4.7 (sleipnir) serial & parallel
Solaris 2.6 (baldric) serial
2003-01-10 01:20:03 +08:00
|
|
|
|
#define TESTFILE2 "tmtimen.h5"
|
|
|
|
|
#define MTIME2 1041606478
|
|
|
|
|
|
1998-07-31 00:38:21 +08:00
|
|
|
|
|
|
|
|
|
/*-------------------------------------------------------------------------
|
|
|
|
|
* Function: main
|
|
|
|
|
*
|
|
|
|
|
* Purpose: H5O_mtime_decode() test.
|
|
|
|
|
*
|
|
|
|
|
* Return: Success:
|
|
|
|
|
*
|
|
|
|
|
* Failure:
|
|
|
|
|
*
|
|
|
|
|
* Programmer: Robb Matzke
|
|
|
|
|
* Thursday, July 30, 1998
|
|
|
|
|
*
|
|
|
|
|
* Modifications:
|
[svn-r6252] Purpose:
Lots of performance improvements & a couple new internal API interfaces.
Description:
Performance Improvements:
- Cached file offset & length sizes in shared file struct, to avoid
constantly looking them up in the FCPL.
- Generic property improvements:
- Added "revision" number to generic property classes to speed
up comparisons.
- Changed method of storing properties from using a hash-table
to the TBBT routines in the library.
- Share the propery names between classes and the lists derived
from them.
- Removed redundant 'def_value' buffer from each property.
- Switching code to use a "copy on write" strategy for
properties in each list, where the properties in each list
are shared with the properties in the class, until a
property's value is changed in a list.
- Fixed error in layout code which was allocating too many buffers.
- Redefined public macros of the form (H5open()/H5check, <variable>)
internally to only be (<variable>), avoiding innumerable useless
calls to H5open() and H5check_version().
- Reuse already zeroed buffers in H5F_contig_fill instead of
constantly re-zeroing them.
- Don't write fill values if writing entire dataset.
- Use gettimeofday() system call instead of time() system when
checking the modification time of a dataset.
- Added reference counted string API and use it for tracking the
names of objects opening in a file (for the ID->name code).
- Removed redundant H5P_get() calls in B-tree routines.
- Redefine H5T datatype macros internally to the library, to avoid
calling H5check redundantly.
- Keep dataspace information for dataset locally instead of reading
from disk each time. Added new module to track open objects
in a file, to allow this (which will be useful eventually for
some FPH5 metadata caching issues).
- Remove H5AC_find macro which was inlining metadata cache lookups,
and call function instead.
- Remove redundant memset() calls from H5G_namei() routine.
- Remove redundant checking of object type when locating objects
in metadata cache and rely on the address only.
- Create default dataset object to use when default dataset creation
property list is used to create datasets, bypassing querying
for all the property list values.
- Use default I/O vector size when performing raw data with the
default dataset transfer property list, instead of querying for
I/O vector size.
- Remove H5P_DEFAULT internally to the library, replacing it with
more specific default property list based on the type of
property list needed.
- Remove redundant memset() calls in object header message (H5O*)
routines.
- Remove redunant memset() calls in data I/O routines.
- Split free-list allocation routines into malloc() and calloc()-
like routines, instead of one combined routine.
- Remove lots of indirection in H5O*() routines.
- Simplify metadata cache entry comparison routine (used when
flushing entire cache out).
- Only enable metadata cache statistics when H5AC_DEBUG is turned
on, instead of always tracking them.
- Simplify address comparison macro (H5F_addr_eq).
- Remove redundant metadata cache entry protections during dataset
creation by protecting the object header once and making all
the modifications necessary for the dataset creation before
unprotecting it.
- Reduce # of "number of element in extent" computations performed
by computing and storing the value during dataspace creation.
- Simplify checking for group location's file information, when file
has not been involving in file-mounting operations.
- Use binary encoding for modification time, instead of ASCII.
- Hoist H5HL_peek calls (to get information in a local heap)
out of loops in many group routine.
- Use static variable for iterators of selections, instead of
dynamically allocation them each time.
- Lookup & insert new entries in one step, avoiding traversing
group's B-tree twice.
- Fixed memory leak in H5Gget_objname_idx() routine (tangential to
performance improvements, but fixed along the way).
- Use free-list for reference counted strings.
- Don't bother copying object names into cached group entries,
since they are re-created when an object is opened.
The benchmark I used to measure these results created several thousand
small (2K) datasets in a file and wrote out the data for them. This is
Elena's "regular.c" benchmark.
These changes resulted in approximately ~4.3x speedup of the
development branch when compared to the previous code in the
development branch and ~1.4x speedup compared to the release
branch.
Additionally, these changes reduce the total memory used (code and
data) by the development branch by ~800KB, bringing the development
branch back into the same ballpark as the release branch.
I'll send out a more detailed description of the benchmark results
as a followup note.
New internal API routines:
Added "reference counted strings" API for tracking strings that get
used by multiple owners without duplicating the strings.
Added "ternary search tree" API for text->object mappings.
Platforms tested:
Tested h5committest {arabica (fortran), eirene (fortran, C++)
modi4 (parallel, fortran)}
Other platforms/configurations tested?
FreeBSD 4.7 (sleipnir) serial & parallel
Solaris 2.6 (baldric) serial
2003-01-10 01:20:03 +08:00
|
|
|
|
* Added checks for old and new modification time messages
|
|
|
|
|
* in pre-created datafiles (generated with gen_old_mtime.c and
|
|
|
|
|
* gen_new_mtime.c).
|
|
|
|
|
* Quincey Koziol
|
|
|
|
|
* Friday, January 3, 2003
|
1998-07-31 00:38:21 +08:00
|
|
|
|
*
|
|
|
|
|
*-------------------------------------------------------------------------
|
|
|
|
|
*/
|
|
|
|
|
int
|
|
|
|
|
main(void)
|
|
|
|
|
{
|
1998-11-24 04:40:35 +08:00
|
|
|
|
hid_t fapl, file, space, dset;
|
1998-07-31 00:38:21 +08:00
|
|
|
|
hsize_t size[1] = {2};
|
|
|
|
|
time_t now;
|
|
|
|
|
struct tm *tm;
|
1998-08-31 21:46:47 +08:00
|
|
|
|
H5G_stat_t sb1, sb2;
|
1998-11-26 01:21:21 +08:00
|
|
|
|
signed char buf1[32], buf2[32];
|
|
|
|
|
char filename[1024];
|
1998-07-31 00:38:21 +08:00
|
|
|
|
|
1998-11-24 04:40:35 +08:00
|
|
|
|
h5_reset();
|
|
|
|
|
fapl = h5_fileaccess();
|
|
|
|
|
|
|
|
|
|
TESTING("modification time messages");
|
1998-07-31 00:38:21 +08:00
|
|
|
|
|
|
|
|
|
/* Create the file, create a dataset, then close the file */
|
1998-11-24 04:40:35 +08:00
|
|
|
|
h5_fixname(FILENAME[0], fapl, filename, sizeof filename);
|
|
|
|
|
if ((file=H5Fcreate(filename, H5F_ACC_TRUNC, H5P_DEFAULT, fapl))<0)
|
|
|
|
|
return 1;
|
|
|
|
|
if ((space=H5Screate_simple(1, size, NULL))<0) return 1;
|
1998-11-26 01:21:21 +08:00
|
|
|
|
if ((dset=H5Dcreate(file, "dset", H5T_NATIVE_SCHAR, space, H5P_DEFAULT))<0)
|
1998-11-24 04:40:35 +08:00
|
|
|
|
return 1;
|
1998-07-31 00:38:21 +08:00
|
|
|
|
now = time(NULL);
|
|
|
|
|
if (H5Dclose(dset)<0) return 1;
|
|
|
|
|
if (H5Sclose(space)<0) return 1;
|
|
|
|
|
if (H5Fclose(file)<0) return 1;
|
|
|
|
|
|
1998-08-31 21:46:47 +08:00
|
|
|
|
/*
|
|
|
|
|
* Open the file and get the modification time. We'll test the new
|
1998-11-24 04:40:35 +08:00
|
|
|
|
* H5Gget_objinfo() arguments too: being able to stat something without
|
|
|
|
|
* knowing its name.
|
1998-08-31 21:46:47 +08:00
|
|
|
|
*/
|
1998-11-24 04:40:35 +08:00
|
|
|
|
h5_fixname(FILENAME[0], fapl, filename, sizeof filename);
|
|
|
|
|
if ((file = H5Fopen(filename, H5F_ACC_RDONLY, fapl))<0) return 1;
|
1998-09-03 05:16:24 +08:00
|
|
|
|
if (H5Gget_objinfo(file, "dset", TRUE, &sb1)<0) return 1;
|
1998-08-31 21:46:47 +08:00
|
|
|
|
if ((dset=H5Dopen(file, "dset"))<0) return 1;
|
1998-09-03 05:16:24 +08:00
|
|
|
|
if (H5Gget_objinfo(dset, ".", TRUE, &sb2)<0) return 1;
|
1998-08-31 21:46:47 +08:00
|
|
|
|
if (H5Dclose(dset)<0) return 1;
|
1998-07-31 00:38:21 +08:00
|
|
|
|
if (H5Fclose(file)<0) return 1;
|
|
|
|
|
|
1998-09-03 05:16:24 +08:00
|
|
|
|
/* Compare times from the two ways of calling H5Gget_objinfo() */
|
2003-08-09 03:25:18 +08:00
|
|
|
|
if (sb1.objno!=sb2.objno || sb1.mtime!=sb2.mtime) {
|
2001-01-26 01:03:29 +08:00
|
|
|
|
H5_FAILED();
|
1998-11-24 04:40:35 +08:00
|
|
|
|
puts(" Calling H5Gget_objinfo() with the dataset ID returned");
|
|
|
|
|
puts(" different values than calling it with a file and dataset");
|
|
|
|
|
puts(" name.");
|
1998-08-31 21:46:47 +08:00
|
|
|
|
return 1;
|
|
|
|
|
}
|
|
|
|
|
|
1998-07-31 00:38:21 +08:00
|
|
|
|
/* Compare times -- they must be within 60 seconds of one another */
|
1998-08-31 21:46:47 +08:00
|
|
|
|
if (0==sb1.mtime) {
|
1998-11-24 04:40:35 +08:00
|
|
|
|
SKIPPED();
|
|
|
|
|
puts(" The modification time could not be decoded on this OS.");
|
|
|
|
|
puts(" Modification times will be mantained in the file but");
|
|
|
|
|
puts(" cannot be queried on this system. See H5O_mtime_decode().");
|
1998-09-03 23:30:11 +08:00
|
|
|
|
return 0;
|
1998-08-31 21:46:47 +08:00
|
|
|
|
} else if (fabs(HDdifftime(now, sb1.mtime))>60.0) {
|
2001-01-26 01:03:29 +08:00
|
|
|
|
H5_FAILED();
|
1998-08-31 21:46:47 +08:00
|
|
|
|
tm = localtime(&(sb1.mtime));
|
1999-02-25 23:40:27 +08:00
|
|
|
|
strftime((char*)buf1, sizeof buf1, "%Y-%m-%d %H:%M:%S", tm);
|
1998-07-31 00:38:21 +08:00
|
|
|
|
tm = localtime(&now);
|
1999-02-25 23:40:27 +08:00
|
|
|
|
strftime((char*)buf2, sizeof buf2, "%Y-%m-%d %H:%M:%S", tm);
|
1998-11-24 04:40:35 +08:00
|
|
|
|
printf(" got: %s\n ans: %s\n", buf1, buf2);
|
1998-07-31 00:38:21 +08:00
|
|
|
|
return 1;
|
|
|
|
|
}
|
[svn-r6252] Purpose:
Lots of performance improvements & a couple new internal API interfaces.
Description:
Performance Improvements:
- Cached file offset & length sizes in shared file struct, to avoid
constantly looking them up in the FCPL.
- Generic property improvements:
- Added "revision" number to generic property classes to speed
up comparisons.
- Changed method of storing properties from using a hash-table
to the TBBT routines in the library.
- Share the propery names between classes and the lists derived
from them.
- Removed redundant 'def_value' buffer from each property.
- Switching code to use a "copy on write" strategy for
properties in each list, where the properties in each list
are shared with the properties in the class, until a
property's value is changed in a list.
- Fixed error in layout code which was allocating too many buffers.
- Redefined public macros of the form (H5open()/H5check, <variable>)
internally to only be (<variable>), avoiding innumerable useless
calls to H5open() and H5check_version().
- Reuse already zeroed buffers in H5F_contig_fill instead of
constantly re-zeroing them.
- Don't write fill values if writing entire dataset.
- Use gettimeofday() system call instead of time() system when
checking the modification time of a dataset.
- Added reference counted string API and use it for tracking the
names of objects opening in a file (for the ID->name code).
- Removed redundant H5P_get() calls in B-tree routines.
- Redefine H5T datatype macros internally to the library, to avoid
calling H5check redundantly.
- Keep dataspace information for dataset locally instead of reading
from disk each time. Added new module to track open objects
in a file, to allow this (which will be useful eventually for
some FPH5 metadata caching issues).
- Remove H5AC_find macro which was inlining metadata cache lookups,
and call function instead.
- Remove redundant memset() calls from H5G_namei() routine.
- Remove redundant checking of object type when locating objects
in metadata cache and rely on the address only.
- Create default dataset object to use when default dataset creation
property list is used to create datasets, bypassing querying
for all the property list values.
- Use default I/O vector size when performing raw data with the
default dataset transfer property list, instead of querying for
I/O vector size.
- Remove H5P_DEFAULT internally to the library, replacing it with
more specific default property list based on the type of
property list needed.
- Remove redundant memset() calls in object header message (H5O*)
routines.
- Remove redunant memset() calls in data I/O routines.
- Split free-list allocation routines into malloc() and calloc()-
like routines, instead of one combined routine.
- Remove lots of indirection in H5O*() routines.
- Simplify metadata cache entry comparison routine (used when
flushing entire cache out).
- Only enable metadata cache statistics when H5AC_DEBUG is turned
on, instead of always tracking them.
- Simplify address comparison macro (H5F_addr_eq).
- Remove redundant metadata cache entry protections during dataset
creation by protecting the object header once and making all
the modifications necessary for the dataset creation before
unprotecting it.
- Reduce # of "number of element in extent" computations performed
by computing and storing the value during dataspace creation.
- Simplify checking for group location's file information, when file
has not been involving in file-mounting operations.
- Use binary encoding for modification time, instead of ASCII.
- Hoist H5HL_peek calls (to get information in a local heap)
out of loops in many group routine.
- Use static variable for iterators of selections, instead of
dynamically allocation them each time.
- Lookup & insert new entries in one step, avoiding traversing
group's B-tree twice.
- Fixed memory leak in H5Gget_objname_idx() routine (tangential to
performance improvements, but fixed along the way).
- Use free-list for reference counted strings.
- Don't bother copying object names into cached group entries,
since they are re-created when an object is opened.
The benchmark I used to measure these results created several thousand
small (2K) datasets in a file and wrote out the data for them. This is
Elena's "regular.c" benchmark.
These changes resulted in approximately ~4.3x speedup of the
development branch when compared to the previous code in the
development branch and ~1.4x speedup compared to the release
branch.
Additionally, these changes reduce the total memory used (code and
data) by the development branch by ~800KB, bringing the development
branch back into the same ballpark as the release branch.
I'll send out a more detailed description of the benchmark results
as a followup note.
New internal API routines:
Added "reference counted strings" API for tracking strings that get
used by multiple owners without duplicating the strings.
Added "ternary search tree" API for text->object mappings.
Platforms tested:
Tested h5committest {arabica (fortran), eirene (fortran, C++)
modi4 (parallel, fortran)}
Other platforms/configurations tested?
FreeBSD 4.7 (sleipnir) serial & parallel
Solaris 2.6 (baldric) serial
2003-01-10 01:20:03 +08:00
|
|
|
|
PASSED();
|
1998-07-31 00:38:21 +08:00
|
|
|
|
|
[svn-r6252] Purpose:
Lots of performance improvements & a couple new internal API interfaces.
Description:
Performance Improvements:
- Cached file offset & length sizes in shared file struct, to avoid
constantly looking them up in the FCPL.
- Generic property improvements:
- Added "revision" number to generic property classes to speed
up comparisons.
- Changed method of storing properties from using a hash-table
to the TBBT routines in the library.
- Share the propery names between classes and the lists derived
from them.
- Removed redundant 'def_value' buffer from each property.
- Switching code to use a "copy on write" strategy for
properties in each list, where the properties in each list
are shared with the properties in the class, until a
property's value is changed in a list.
- Fixed error in layout code which was allocating too many buffers.
- Redefined public macros of the form (H5open()/H5check, <variable>)
internally to only be (<variable>), avoiding innumerable useless
calls to H5open() and H5check_version().
- Reuse already zeroed buffers in H5F_contig_fill instead of
constantly re-zeroing them.
- Don't write fill values if writing entire dataset.
- Use gettimeofday() system call instead of time() system when
checking the modification time of a dataset.
- Added reference counted string API and use it for tracking the
names of objects opening in a file (for the ID->name code).
- Removed redundant H5P_get() calls in B-tree routines.
- Redefine H5T datatype macros internally to the library, to avoid
calling H5check redundantly.
- Keep dataspace information for dataset locally instead of reading
from disk each time. Added new module to track open objects
in a file, to allow this (which will be useful eventually for
some FPH5 metadata caching issues).
- Remove H5AC_find macro which was inlining metadata cache lookups,
and call function instead.
- Remove redundant memset() calls from H5G_namei() routine.
- Remove redundant checking of object type when locating objects
in metadata cache and rely on the address only.
- Create default dataset object to use when default dataset creation
property list is used to create datasets, bypassing querying
for all the property list values.
- Use default I/O vector size when performing raw data with the
default dataset transfer property list, instead of querying for
I/O vector size.
- Remove H5P_DEFAULT internally to the library, replacing it with
more specific default property list based on the type of
property list needed.
- Remove redundant memset() calls in object header message (H5O*)
routines.
- Remove redunant memset() calls in data I/O routines.
- Split free-list allocation routines into malloc() and calloc()-
like routines, instead of one combined routine.
- Remove lots of indirection in H5O*() routines.
- Simplify metadata cache entry comparison routine (used when
flushing entire cache out).
- Only enable metadata cache statistics when H5AC_DEBUG is turned
on, instead of always tracking them.
- Simplify address comparison macro (H5F_addr_eq).
- Remove redundant metadata cache entry protections during dataset
creation by protecting the object header once and making all
the modifications necessary for the dataset creation before
unprotecting it.
- Reduce # of "number of element in extent" computations performed
by computing and storing the value during dataspace creation.
- Simplify checking for group location's file information, when file
has not been involving in file-mounting operations.
- Use binary encoding for modification time, instead of ASCII.
- Hoist H5HL_peek calls (to get information in a local heap)
out of loops in many group routine.
- Use static variable for iterators of selections, instead of
dynamically allocation them each time.
- Lookup & insert new entries in one step, avoiding traversing
group's B-tree twice.
- Fixed memory leak in H5Gget_objname_idx() routine (tangential to
performance improvements, but fixed along the way).
- Use free-list for reference counted strings.
- Don't bother copying object names into cached group entries,
since they are re-created when an object is opened.
The benchmark I used to measure these results created several thousand
small (2K) datasets in a file and wrote out the data for them. This is
Elena's "regular.c" benchmark.
These changes resulted in approximately ~4.3x speedup of the
development branch when compared to the previous code in the
development branch and ~1.4x speedup compared to the release
branch.
Additionally, these changes reduce the total memory used (code and
data) by the development branch by ~800KB, bringing the development
branch back into the same ballpark as the release branch.
I'll send out a more detailed description of the benchmark results
as a followup note.
New internal API routines:
Added "reference counted strings" API for tracking strings that get
used by multiple owners without duplicating the strings.
Added "ternary search tree" API for text->object mappings.
Platforms tested:
Tested h5committest {arabica (fortran), eirene (fortran, C++)
modi4 (parallel, fortran)}
Other platforms/configurations tested?
FreeBSD 4.7 (sleipnir) serial & parallel
Solaris 2.6 (baldric) serial
2003-01-10 01:20:03 +08:00
|
|
|
|
/* Check opening existing file with old-style modification time information
|
|
|
|
|
* and make certain that the time is correct
|
|
|
|
|
*/
|
|
|
|
|
TESTING("accessing old modification time messages");
|
|
|
|
|
|
|
|
|
|
{
|
|
|
|
|
char testfile[512]="";
|
|
|
|
|
char *srcdir = HDgetenv("srcdir");
|
|
|
|
|
if (srcdir && ((HDstrlen(srcdir) + strlen(TESTFILE1) + 1) < sizeof(testfile))){
|
|
|
|
|
HDstrcpy(testfile, srcdir);
|
|
|
|
|
HDstrcat(testfile, "/");
|
|
|
|
|
}
|
|
|
|
|
HDstrcat(testfile, TESTFILE1);
|
|
|
|
|
file = H5Fopen(testfile, H5F_ACC_RDONLY, H5P_DEFAULT);
|
|
|
|
|
if (file >= 0){
|
|
|
|
|
if(H5Gget_objinfo(file, "/Dataset1", TRUE, &sb1)<0)
|
|
|
|
|
return 1;
|
|
|
|
|
if(sb1.mtime!=MTIME1) return 1;
|
|
|
|
|
if (H5Fclose(file)<0) return 1;
|
|
|
|
|
}
|
|
|
|
|
else {
|
|
|
|
|
printf("***cannot open the pre-created old modification test file (%s)\n",
|
|
|
|
|
testfile);
|
|
|
|
|
return 1;
|
|
|
|
|
} /* end else */
|
|
|
|
|
}
|
1998-11-24 04:40:35 +08:00
|
|
|
|
PASSED();
|
[svn-r6252] Purpose:
Lots of performance improvements & a couple new internal API interfaces.
Description:
Performance Improvements:
- Cached file offset & length sizes in shared file struct, to avoid
constantly looking them up in the FCPL.
- Generic property improvements:
- Added "revision" number to generic property classes to speed
up comparisons.
- Changed method of storing properties from using a hash-table
to the TBBT routines in the library.
- Share the propery names between classes and the lists derived
from them.
- Removed redundant 'def_value' buffer from each property.
- Switching code to use a "copy on write" strategy for
properties in each list, where the properties in each list
are shared with the properties in the class, until a
property's value is changed in a list.
- Fixed error in layout code which was allocating too many buffers.
- Redefined public macros of the form (H5open()/H5check, <variable>)
internally to only be (<variable>), avoiding innumerable useless
calls to H5open() and H5check_version().
- Reuse already zeroed buffers in H5F_contig_fill instead of
constantly re-zeroing them.
- Don't write fill values if writing entire dataset.
- Use gettimeofday() system call instead of time() system when
checking the modification time of a dataset.
- Added reference counted string API and use it for tracking the
names of objects opening in a file (for the ID->name code).
- Removed redundant H5P_get() calls in B-tree routines.
- Redefine H5T datatype macros internally to the library, to avoid
calling H5check redundantly.
- Keep dataspace information for dataset locally instead of reading
from disk each time. Added new module to track open objects
in a file, to allow this (which will be useful eventually for
some FPH5 metadata caching issues).
- Remove H5AC_find macro which was inlining metadata cache lookups,
and call function instead.
- Remove redundant memset() calls from H5G_namei() routine.
- Remove redundant checking of object type when locating objects
in metadata cache and rely on the address only.
- Create default dataset object to use when default dataset creation
property list is used to create datasets, bypassing querying
for all the property list values.
- Use default I/O vector size when performing raw data with the
default dataset transfer property list, instead of querying for
I/O vector size.
- Remove H5P_DEFAULT internally to the library, replacing it with
more specific default property list based on the type of
property list needed.
- Remove redundant memset() calls in object header message (H5O*)
routines.
- Remove redunant memset() calls in data I/O routines.
- Split free-list allocation routines into malloc() and calloc()-
like routines, instead of one combined routine.
- Remove lots of indirection in H5O*() routines.
- Simplify metadata cache entry comparison routine (used when
flushing entire cache out).
- Only enable metadata cache statistics when H5AC_DEBUG is turned
on, instead of always tracking them.
- Simplify address comparison macro (H5F_addr_eq).
- Remove redundant metadata cache entry protections during dataset
creation by protecting the object header once and making all
the modifications necessary for the dataset creation before
unprotecting it.
- Reduce # of "number of element in extent" computations performed
by computing and storing the value during dataspace creation.
- Simplify checking for group location's file information, when file
has not been involving in file-mounting operations.
- Use binary encoding for modification time, instead of ASCII.
- Hoist H5HL_peek calls (to get information in a local heap)
out of loops in many group routine.
- Use static variable for iterators of selections, instead of
dynamically allocation them each time.
- Lookup & insert new entries in one step, avoiding traversing
group's B-tree twice.
- Fixed memory leak in H5Gget_objname_idx() routine (tangential to
performance improvements, but fixed along the way).
- Use free-list for reference counted strings.
- Don't bother copying object names into cached group entries,
since they are re-created when an object is opened.
The benchmark I used to measure these results created several thousand
small (2K) datasets in a file and wrote out the data for them. This is
Elena's "regular.c" benchmark.
These changes resulted in approximately ~4.3x speedup of the
development branch when compared to the previous code in the
development branch and ~1.4x speedup compared to the release
branch.
Additionally, these changes reduce the total memory used (code and
data) by the development branch by ~800KB, bringing the development
branch back into the same ballpark as the release branch.
I'll send out a more detailed description of the benchmark results
as a followup note.
New internal API routines:
Added "reference counted strings" API for tracking strings that get
used by multiple owners without duplicating the strings.
Added "ternary search tree" API for text->object mappings.
Platforms tested:
Tested h5committest {arabica (fortran), eirene (fortran, C++)
modi4 (parallel, fortran)}
Other platforms/configurations tested?
FreeBSD 4.7 (sleipnir) serial & parallel
Solaris 2.6 (baldric) serial
2003-01-10 01:20:03 +08:00
|
|
|
|
|
|
|
|
|
/* Check opening existing file with new-style modification time information
|
|
|
|
|
* and make certain that the time is correct
|
|
|
|
|
*/
|
|
|
|
|
TESTING("accessing new modification time messages");
|
|
|
|
|
|
|
|
|
|
{
|
|
|
|
|
char testfile[512]="";
|
|
|
|
|
char *srcdir = HDgetenv("srcdir");
|
|
|
|
|
if (srcdir && ((HDstrlen(srcdir) + strlen(TESTFILE2) + 1) < sizeof(testfile))){
|
|
|
|
|
HDstrcpy(testfile, srcdir);
|
|
|
|
|
HDstrcat(testfile, "/");
|
|
|
|
|
}
|
|
|
|
|
HDstrcat(testfile, TESTFILE2);
|
|
|
|
|
file = H5Fopen(testfile, H5F_ACC_RDONLY, H5P_DEFAULT);
|
|
|
|
|
if (file >= 0){
|
|
|
|
|
if(H5Gget_objinfo(file, "/Dataset1", TRUE, &sb2)<0)
|
|
|
|
|
return 1;
|
|
|
|
|
if(sb2.mtime!=MTIME2) return 1;
|
|
|
|
|
if (H5Fclose(file)<0) return 1;
|
|
|
|
|
}
|
|
|
|
|
else {
|
|
|
|
|
printf("***cannot open the pre-created old modification test file (%s)\n",
|
|
|
|
|
testfile);
|
|
|
|
|
return 1;
|
|
|
|
|
} /* end else */
|
|
|
|
|
}
|
|
|
|
|
PASSED();
|
|
|
|
|
|
|
|
|
|
/* All looks good */
|
1998-11-24 04:40:35 +08:00
|
|
|
|
puts("All modification time tests passed.");
|
2000-09-10 08:08:27 +08:00
|
|
|
|
h5_cleanup(FILENAME, fapl);
|
1998-07-31 00:38:21 +08:00
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|