mirror/hdf5
2
0
Fork 0
mirror of https://github.com/HDFGroup/hdf5.git synced 2024-11-27 02:10:55 +08:00
Code Issues Packages Projects Releases Wiki Activity
389cc309c5
hdf5/test/set_extent.c

505 lines
18 KiB
C
Raw Normal View History

[svn-r6538] Purpose: Update Description: Updated the Copyright statement Platforms tested: Linux (This change is only in the comments, so I just check that the modules still compile) Misc. update:
2003-04-01 01:59:04 +08:00
/* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
[svn-r13253] Updated all C and C++ style source code files with the THG copyright notice. Tested platform: Kagiso only since it is only a comment block change. If it works in one machine, it should work in all, I hope. Still need to check the parallel build on copper.
2007-02-07 22:56:24 +08:00
* Copyright by The HDF Group. *
[svn-r6538] Purpose: Update Description: Updated the Copyright statement Platforms tested: Linux (This change is only in the comments, so I just check that the modules still compile) Misc. update:
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 *
[svn-r13253] Updated all C and C++ style source code files with the THG copyright notice. Tested platform: Kagiso only since it is only a comment block change. If it works in one machine, it should work in all, I hope. Still need to check the parallel build on copper.
2007-02-07 22:56:24 +08:00
* http://hdfgroup.org/HDF5/doc/Copyright.html. If you do not have *
* access to either file, you may request a copy from help@hdfgroup.org. *
[svn-r6538] Purpose: Update Description: Updated the Copyright statement Platforms tested: Linux (This change is only in the comments, so I just check that the modules still compile) Misc. update:
2003-04-01 01:59:04 +08:00
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
[svn-r5200] Purpose: Code Cleanup Description: Clean up compiler warnings from the last bunch of checkins Platforms tested: FreeBSD 4.5 (sleipnir)
2002-04-18 22:05:38 +08:00
/*
* Programmer: Pedro Vicente <pvn@ncsa.uiuc.edu>
* April 12, 2002
*
* Purpose: Tests the H5Dset_extent call
*/
[svn-r5154] Purpose: Renamed the all the functions "set_extend" to "set_extent". Renamed the test file accordingly
2002-04-10 01:31:21 +08:00
#include "hdf5.h"
#include "h5test.h"
/*-------------------------------------------------------------------------
*
[svn-r11245] Purpose: Code cleanup Description: Trim trailing whitespace, which is making 'diff'ing the two branches difficult. Solution: Ran this script in each directory: foreach f (*.[ch] *.cpp) sed 's/[[:blank:]]*$//' $f > sed.out && mv sed.out $f end Platforms tested: FreeBSD 4.11 (sleipnir) Too minor to require h5committest
2005-08-14 04:53:35 +08:00
* Tests the function H5Dset_extent. In the current version of the library
[svn-r5154] Purpose: Renamed the all the functions "set_extend" to "set_extent". Renamed the test file accordingly
2002-04-10 01:31:21 +08:00
* the dataset MUST be chunked.
*
*-------------------------------------------------------------------------
*/
#define RANK 2
[svn-r7232] Purpose: Bug fix Description: When a non-default indexed storage B-tree internal 'K' value is set by the user, the chunked datasets created in that file (until it is closed) use the user's 'K' value and the data can be accessed correctly, but the 'K' value is not stored in the file. However, once the file is closed and re-opened, the non-default 'K' value is lost and the data in the chunked datasets will not be able to be accessed correctly. Solution: Store the indexed storage B-tree internal 'K' value in the superblock. Platforms tested: FreeBSD 4.8 (sleipnir) h5committest
2003-07-16 22:56:58 +08:00
#define ISTORE_IK 64
[svn-r5154] Purpose: Renamed the all the functions "set_extend" to "set_extent". Renamed the test file accordingly
2002-04-10 01:31:21 +08:00
int main( void )
{
[svn-r5169] Purpose: test for the H5Dset_extent API function Description: there are 2 types of tests: Test H5Dset_extent with chunks on the raw data cache Test H5Dset_extent with chunks written to file Platforms tested: Windows 2000 SUN( arabica) LINUX (eirene) IRIX64 (modi4)
2002-04-12 05:30:42 +08:00
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
hid_t file_id;
hid_t dataset_id=(-1);
[svn-r11245] Purpose: Code cleanup Description: Trim trailing whitespace, which is making 'diff'ing the two branches difficult. Solution: Ran this script in each directory: foreach f (*.[ch] *.cpp) sed 's/[[:blank:]]*$//' $f > sed.out && mv sed.out $f end Platforms tested: FreeBSD 4.11 (sleipnir) Too minor to require h5committest
2005-08-14 04:53:35 +08:00
hid_t space_id=(-1);
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
hid_t plist_id=(-1);
[svn-r7232] Purpose: Bug fix Description: When a non-default indexed storage B-tree internal 'K' value is set by the user, the chunked datasets created in that file (until it is closed) use the user's 'K' value and the data can be accessed correctly, but the 'K' value is not stored in the file. However, once the file is closed and re-opened, the non-default 'K' value is lost and the data in the chunked datasets will not be able to be accessed correctly. Solution: Store the indexed storage B-tree internal 'K' value in the superblock. Platforms tested: FreeBSD 4.8 (sleipnir) h5committest
2003-07-16 22:56:58 +08:00
hid_t fcpl; /* File creation property list */
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
hsize_t dims[RANK] = { 90, 90 };
hsize_t dims_new[RANK] = { 70, 70 };
hsize_t dims_chunk[RANK] = { 20, 20 };
hsize_t dims_out[RANK];
hsize_t maxdims[RANK] = { H5S_UNLIMITED, H5S_UNLIMITED };
int data[ 90 ][ 90 ];
int buf1[ 70 ][ 70 ];
int buf2[ 90 ][ 90 ];
int i, j, n = 0;
int fillvalue = 1; /* Fill value for the dataset */
[svn-r5313] Purpose: added more tests to the H5set_extent function Description: there was a bug in the H5S_select_fill call when the fill value was not defined added 2 more set of tests that call H5set_extent without the fill value being defined there are now 4 sets of tests: Test H5Dset_extent with chunks on the raw data cache, with and without fill value Test H5Dset_extent with chunks written to file (b-tree routines), with and without fill value Platforms tested: w2000 linux
2002-05-01 23:54:56 +08:00
[svn-r5167] Purpose: turned temporarely off this test, while the rest of the code is not checked in
2002-04-11 23:54:39 +08:00
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
for( i = 0; i < 90; i++ )
for( j = 0; j < 90; j++ )
data[i][j] = n++;
/*-------------------------------------------------------------------------
* Test H5Dset_extent with chunks on the raw data cache
*-------------------------------------------------------------------------
*/
/* Create a new file using default properties. */
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if((file_id = H5Fcreate("set_extent_create.h5", H5F_ACC_TRUNC, H5P_DEFAULT, H5P_DEFAULT)) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
TESTING("extend dataset create with fill value");
/* Create the data space with unlimited dimensions. */
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if((space_id = H5Screate_simple(RANK, dims, maxdims)) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Modify dataset creation properties, i.e. enable chunking. */
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if((plist_id = H5Pcreate (H5P_DATASET_CREATE)) < 0) TEST_ERROR;
if(H5Pset_chunk(plist_id, RANK, dims_chunk) < 0) TEST_ERROR;
if(H5Pset_fill_value(plist_id, H5T_NATIVE_INT, &fillvalue) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/*-------------------------------------------------------------------------
* Create and write one dataset
*-------------------------------------------------------------------------
*/
/* Create a new dataset */
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if((dataset_id = H5Dcreate2(file_id , "Dataset1", H5T_NATIVE_INT, space_id, H5P_DEFAULT, plist_id, H5P_DEFAULT)) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Write the data. */
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if(H5Dwrite(dataset_id , H5T_NATIVE_INT, space_id, H5S_ALL, H5P_DEFAULT, data) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/*-------------------------------------------------------------------------
[svn-r11245] Purpose: Code cleanup Description: Trim trailing whitespace, which is making 'diff'ing the two branches difficult. Solution: Ran this script in each directory: foreach f (*.[ch] *.cpp) sed 's/[[:blank:]]*$//' $f > sed.out && mv sed.out $f end Platforms tested: FreeBSD 4.11 (sleipnir) Too minor to require h5committest
2005-08-14 04:53:35 +08:00
* Set new dimensions for the array; shrink it
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
*-------------------------------------------------------------------------
*/
/* Set new dimensions for the array. */
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if(H5Dset_extent(dataset_id , dims_new) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Get the space. */
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if((space_id = H5Dget_space(dataset_id)) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Get dimensions. */
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if(H5Sget_simple_extent_dims(space_id, dims_out, NULL) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if(dims_out[0] != dims_new[0]) TEST_ERROR;
if(dims_out[1] != dims_new[1]) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/*-------------------------------------------------------------------------
* Read
*-------------------------------------------------------------------------
*/
/* Read the new dataset. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if (H5Dread( dataset_id, H5T_NATIVE_INT, H5S_ALL, H5S_ALL, H5P_DEFAULT, buf1 ) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Compare the read array with the original array */
for( i = 0; i < (int)dims_out[0]; i++ )
for( j = 0; j < (int)dims_out[1]; j++ )
[svn-r8600] Purpose: Code optimization Description: Don't recompute the internal index value for looking up the chunk in the hash table, just use the value already computed from iterating through the chunks. Platforms tested: Solaris 2.7 (arabica) FreeBSD 4.9 (sleipnir) w/parallel
2004-06-01 03:59:59 +08:00
if ( buf1[i][j] != data[i][j] ) {
[svn-r13741] Description: Minor formatting & warning reduction code cleanups Tested on: Mac OS X/32 10.4.9 (amazon) Linux/32 2.6 (chicago) Linux/64 2.6 (chicago2)
2007-05-10 23:06:49 +08:00
printf("buf1[%d][%d] = %d\n", i, j, buf1[i][j]);
printf("data[%d][%d] = %d\n", i, j, data[i][j]);
[svn-r8600] Purpose: Code optimization Description: Don't recompute the internal index value for looking up the chunk in the hash table, just use the value already computed from iterating through the chunks. Platforms tested: Solaris 2.7 (arabica) FreeBSD 4.9 (sleipnir) w/parallel
2004-06-01 03:59:59 +08:00
TEST_ERROR;
} /* end if */
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/*-------------------------------------------------------------------------
[svn-r13741] Description: Minor formatting & warning reduction code cleanups Tested on: Mac OS X/32 10.4.9 (amazon) Linux/32 2.6 (chicago) Linux/64 2.6 (chicago2)
2007-05-10 23:06:49 +08:00
* Set new dimensions for the array; expand it back to original size
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
*-------------------------------------------------------------------------
*/
/* Set new dimensions for the array. */
[svn-r13741] Description: Minor formatting & warning reduction code cleanups Tested on: Mac OS X/32 10.4.9 (amazon) Linux/32 2.6 (chicago) Linux/64 2.6 (chicago2)
2007-05-10 23:06:49 +08:00
if(H5Dset_extent(dataset_id, dims) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Get the space. */
[svn-r13741] Description: Minor formatting & warning reduction code cleanups Tested on: Mac OS X/32 10.4.9 (amazon) Linux/32 2.6 (chicago) Linux/64 2.6 (chicago2)
2007-05-10 23:06:49 +08:00
if((space_id = H5Dget_space(dataset_id)) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Get dimensions. */
[svn-r13741] Description: Minor formatting & warning reduction code cleanups Tested on: Mac OS X/32 10.4.9 (amazon) Linux/32 2.6 (chicago) Linux/64 2.6 (chicago2)
2007-05-10 23:06:49 +08:00
if(H5Sget_simple_extent_dims(space_id, dims_out, NULL) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
[svn-r13741] Description: Minor formatting & warning reduction code cleanups Tested on: Mac OS X/32 10.4.9 (amazon) Linux/32 2.6 (chicago) Linux/64 2.6 (chicago2)
2007-05-10 23:06:49 +08:00
if(dims_out[0] != dims[0]) TEST_ERROR;
if(dims_out[1] != dims[1]) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/*-------------------------------------------------------------------------
* Read
*-------------------------------------------------------------------------
*/
/* Read the new dataset. */
[svn-r13741] Description: Minor formatting & warning reduction code cleanups Tested on: Mac OS X/32 10.4.9 (amazon) Linux/32 2.6 (chicago) Linux/64 2.6 (chicago2)
2007-05-10 23:06:49 +08:00
if(H5Dread(dataset_id, H5T_NATIVE_INT, H5S_ALL, H5S_ALL, H5P_DEFAULT, buf2) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Compare the read array with the original array */
[svn-r13741] Description: Minor formatting & warning reduction code cleanups Tested on: Mac OS X/32 10.4.9 (amazon) Linux/32 2.6 (chicago) Linux/64 2.6 (chicago2)
2007-05-10 23:06:49 +08:00
for(i = 0; i < (int)dims_out[0]; i++ )
for(j = 0; j < (int)dims_out[1]; j++ )
if(i >= 70 || j >= 70) {
if(buf2[i][j] != fillvalue) {
printf("buf1[%d][%d] = %d\n", i, j, buf1[i][j]);
printf("fillvalue = %d\n", fillvalue);
TEST_ERROR;
} /* end if */
} /* end if */
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
else {
[svn-r13741] Description: Minor formatting & warning reduction code cleanups Tested on: Mac OS X/32 10.4.9 (amazon) Linux/32 2.6 (chicago) Linux/64 2.6 (chicago2)
2007-05-10 23:06:49 +08:00
if(buf2[i][j] != data[i][j]) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
}
/*-------------------------------------------------------------------------
* Close/release resources
*-------------------------------------------------------------------------
*/
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if(H5Dclose(dataset_id) < 0) TEST_ERROR
if(H5Sclose(space_id) < 0) TEST_ERROR
if(H5Pclose(plist_id) < 0) TEST_ERROR
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
PASSED();
TESTING("extend dataset create without fill value");
/* Create the data space with unlimited dimensions. */
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if((space_id = H5Screate_simple(RANK, dims, maxdims)) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Modify dataset creation properties, i.e. enable chunking. */
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if((plist_id = H5Pcreate(H5P_DATASET_CREATE)) < 0) TEST_ERROR;
if(H5Pset_chunk(plist_id, RANK, dims_chunk) < 0) TEST_ERROR;
if(H5Pset_fill_time(plist_id, H5D_FILL_TIME_ALLOC) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/*-------------------------------------------------------------------------
* Create and write one dataset
*-------------------------------------------------------------------------
*/
/* Create a new dataset */
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if((dataset_id = H5Dcreate2(file_id , "Dataset2", H5T_NATIVE_INT, space_id, H5P_DEFAULT, plist_id, H5P_DEFAULT)) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Write the data. */
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if(H5Dwrite(dataset_id , H5T_NATIVE_INT, space_id, H5S_ALL, H5P_DEFAULT, data) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/*-------------------------------------------------------------------------
[svn-r11245] Purpose: Code cleanup Description: Trim trailing whitespace, which is making 'diff'ing the two branches difficult. Solution: Ran this script in each directory: foreach f (*.[ch] *.cpp) sed 's/[[:blank:]]*$//' $f > sed.out && mv sed.out $f end Platforms tested: FreeBSD 4.11 (sleipnir) Too minor to require h5committest
2005-08-14 04:53:35 +08:00
* Set new dimensions for the array; shrink it
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
*-------------------------------------------------------------------------
*/
/* Set new dimensions for the array. */
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if(H5Dset_extent(dataset_id , dims_new) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Get the space. */
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if((space_id = H5Dget_space(dataset_id)) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Get dimensions. */
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if(H5Sget_simple_extent_dims(space_id, dims_out, NULL) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if(dims_out[0] != dims_new[0]) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/*-------------------------------------------------------------------------
* Read
*-------------------------------------------------------------------------
*/
/* Read the new dataset. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if (H5Dread( dataset_id, H5T_NATIVE_INT, H5S_ALL, H5S_ALL, H5P_DEFAULT, buf1 ) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Compare the read array with the original array */
for( i = 0; i < (int)dims_out[0]; i++ )
for( j = 0; j < (int)dims_out[1]; j++ )
[svn-r8600] Purpose: Code optimization Description: Don't recompute the internal index value for looking up the chunk in the hash table, just use the value already computed from iterating through the chunks. Platforms tested: Solaris 2.7 (arabica) FreeBSD 4.9 (sleipnir) w/parallel
2004-06-01 03:59:59 +08:00
if ( buf1[i][j] != data[i][j] ) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/*-------------------------------------------------------------------------
[svn-r11245] Purpose: Code cleanup Description: Trim trailing whitespace, which is making 'diff'ing the two branches difficult. Solution: Ran this script in each directory: foreach f (*.[ch] *.cpp) sed 's/[[:blank:]]*$//' $f > sed.out && mv sed.out $f end Platforms tested: FreeBSD 4.11 (sleipnir) Too minor to require h5committest
2005-08-14 04:53:35 +08:00
* Set new dimensions for the array; expand it again
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
*-------------------------------------------------------------------------
*/
/* Set new dimensions for the array. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if (H5Dset_extent( dataset_id , dims ) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Get the space. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if ((space_id = H5Dget_space( dataset_id )) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Get dimensions. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if (H5Sget_simple_extent_dims( space_id, dims_out, NULL ) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
[svn-r8600] Purpose: Code optimization Description: Don't recompute the internal index value for looking up the chunk in the hash table, just use the value already computed from iterating through the chunks. Platforms tested: Solaris 2.7 (arabica) FreeBSD 4.9 (sleipnir) w/parallel
2004-06-01 03:59:59 +08:00
if ( dims_out[0] != dims[0] ) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/*-------------------------------------------------------------------------
* Read
*-------------------------------------------------------------------------
*/
/* Read the new dataset. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if (H5Dread( dataset_id, H5T_NATIVE_INT, H5S_ALL, H5S_ALL, H5P_DEFAULT, buf2 ) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Compare the read array with the original array */
for( i = 0; i < (int)dims_out[0]; i++ ) {
for( j = 0; j < (int)dims_out[1]; j++ ) {
if ( i >= 70 || j >= 70 ) {
[svn-r8600] Purpose: Code optimization Description: Don't recompute the internal index value for looking up the chunk in the hash table, just use the value already computed from iterating through the chunks. Platforms tested: Solaris 2.7 (arabica) FreeBSD 4.9 (sleipnir) w/parallel
2004-06-01 03:59:59 +08:00
if ( buf2[i][j] != 0 ) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
}
else {
[svn-r8600] Purpose: Code optimization Description: Don't recompute the internal index value for looking up the chunk in the hash table, just use the value already computed from iterating through the chunks. Platforms tested: Solaris 2.7 (arabica) FreeBSD 4.9 (sleipnir) w/parallel
2004-06-01 03:59:59 +08:00
if ( buf2[i][j] != data[i][j] ) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
}
}
[svn-r5154] Purpose: Renamed the all the functions "set_extend" to "set_extent". Renamed the test file accordingly
2002-04-10 01:31:21 +08:00
}
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/*-------------------------------------------------------------------------
* Close/release resources
*-------------------------------------------------------------------------
*/
H5Dclose( dataset_id );
H5Sclose( space_id );
H5Pclose( plist_id );
H5Fclose( file_id );
PASSED();
/*-------------------------------------------------------------------------
* Test H5Dset_extent with chunks written to file
*-------------------------------------------------------------------------
*/
[svn-r7232] Purpose: Bug fix Description: When a non-default indexed storage B-tree internal 'K' value is set by the user, the chunked datasets created in that file (until it is closed) use the user's 'K' value and the data can be accessed correctly, but the 'K' value is not stored in the file. However, once the file is closed and re-opened, the non-default 'K' value is lost and the data in the chunked datasets will not be able to be accessed correctly. Solution: Store the indexed storage B-tree internal 'K' value in the superblock. Platforms tested: FreeBSD 4.8 (sleipnir) h5committest
2003-07-16 22:56:58 +08:00
/* Create a file creation property list */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if((fcpl = H5Pcreate(H5P_FILE_CREATE)) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
[svn-r7232] Purpose: Bug fix Description: When a non-default indexed storage B-tree internal 'K' value is set by the user, the chunked datasets created in that file (until it is closed) use the user's 'K' value and the data can be accessed correctly, but the 'K' value is not stored in the file. However, once the file is closed and re-opened, the non-default 'K' value is lost and the data in the chunked datasets will not be able to be accessed correctly. Solution: Store the indexed storage B-tree internal 'K' value in the superblock. Platforms tested: FreeBSD 4.8 (sleipnir) h5committest
2003-07-16 22:56:58 +08:00
/* Set non-default indexed storage B-tree internal 'K' value */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if(H5Pset_istore_k(fcpl,ISTORE_IK) < 0) TEST_ERROR;
[svn-r7232] Purpose: Bug fix Description: When a non-default indexed storage B-tree internal 'K' value is set by the user, the chunked datasets created in that file (until it is closed) use the user's 'K' value and the data can be accessed correctly, but the 'K' value is not stored in the file. However, once the file is closed and re-opened, the non-default 'K' value is lost and the data in the chunked datasets will not be able to be accessed correctly. Solution: Store the indexed storage B-tree internal 'K' value in the superblock. Platforms tested: FreeBSD 4.8 (sleipnir) h5committest
2003-07-16 22:56:58 +08:00
/* Create a new file using properties. */
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if((file_id = H5Fcreate("set_extent_read.h5", H5F_ACC_TRUNC, fcpl, H5P_DEFAULT)) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
[svn-r7232] Purpose: Bug fix Description: When a non-default indexed storage B-tree internal 'K' value is set by the user, the chunked datasets created in that file (until it is closed) use the user's 'K' value and the data can be accessed correctly, but the 'K' value is not stored in the file. However, once the file is closed and re-opened, the non-default 'K' value is lost and the data in the chunked datasets will not be able to be accessed correctly. Solution: Store the indexed storage B-tree internal 'K' value in the superblock. Platforms tested: FreeBSD 4.8 (sleipnir) h5committest
2003-07-16 22:56:58 +08:00
/* Close property list */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if(H5Pclose(fcpl) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
TESTING("extend dataset read with fill value");
/* Create the data space with unlimited dimensions. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if((space_id = H5Screate_simple(RANK, dims, maxdims)) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Modify dataset creation properties, i.e. enable chunking. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if((plist_id = H5Pcreate(H5P_DATASET_CREATE)) < 0) TEST_ERROR;
if(H5Pset_chunk(plist_id, RANK, dims_chunk) < 0) TEST_ERROR;
if(H5Pset_fill_value(plist_id, H5T_NATIVE_INT, &fillvalue) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Create a new dataset within the file using cparms creation properties. */
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if((dataset_id = H5Dcreate2(file_id , "Dataset1", H5T_NATIVE_INT, space_id, H5P_DEFAULT, plist_id, H5P_DEFAULT)) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Write the data. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if(H5Dwrite(dataset_id , H5T_NATIVE_INT, space_id, H5S_ALL, H5P_DEFAULT, data) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Close/release resources. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if(H5Dclose(dataset_id) < 0) FAIL_STACK_ERROR
if(H5Sclose(space_id) < 0) FAIL_STACK_ERROR
if(H5Pclose(plist_id) < 0) FAIL_STACK_ERROR
if(H5Fclose(file_id) < 0) FAIL_STACK_ERROR
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Open the file */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if((file_id = H5Fopen("set_extent_read.h5", H5F_ACC_RDWR, H5P_DEFAULT)) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Open the dataset */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if((dataset_id = H5Dopen2(file_id , "Dataset1", H5P_DEFAULT)) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Set new dimensions for the array. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if(H5Dset_extent(dataset_id, dims_new) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Get the space. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if((space_id = H5Dget_space(dataset_id)) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Get dimensions. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if(H5Sget_simple_extent_dims(space_id, dims_out, NULL) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if(dims_out[0] != dims_new[0]) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Read the new dataset. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if(H5Dread(dataset_id, H5T_NATIVE_INT, H5S_ALL, H5S_ALL, H5P_DEFAULT, buf1) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Compare the read array with the original array */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
for(i = 0; i < (int)dims_out[0]; i++)
for(j = 0; j < (int)dims_out[1]; j++)
if(buf1[i][j] != data[i][j])
TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/*-------------------------------------------------------------------------
[svn-r11245] Purpose: Code cleanup Description: Trim trailing whitespace, which is making 'diff'ing the two branches difficult. Solution: Ran this script in each directory: foreach f (*.[ch] *.cpp) sed 's/[[:blank:]]*$//' $f > sed.out && mv sed.out $f end Platforms tested: FreeBSD 4.11 (sleipnir) Too minor to require h5committest
2005-08-14 04:53:35 +08:00
* Set new dimensions for the array; expand it again
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
*-------------------------------------------------------------------------
*/
/* Set new dimensions for the array. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if (H5Dset_extent( dataset_id , dims ) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Get the space. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if ((space_id = H5Dget_space( dataset_id )) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Get dimensions. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if (H5Sget_simple_extent_dims( space_id, dims_out, NULL ) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
[svn-r8600] Purpose: Code optimization Description: Don't recompute the internal index value for looking up the chunk in the hash table, just use the value already computed from iterating through the chunks. Platforms tested: Solaris 2.7 (arabica) FreeBSD 4.9 (sleipnir) w/parallel
2004-06-01 03:59:59 +08:00
if ( dims_out[0] != dims[0] ) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Read the new dataset. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if (H5Dread( dataset_id, H5T_NATIVE_INT, H5S_ALL, H5S_ALL, H5P_DEFAULT, buf2 ) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Compare the read array with the original array */
for( i = 0; i < (int)dims_out[0]; i++ ) {
for( j = 0; j < (int)dims_out[1]; j++ ) {
if ( i >= 70 || j >= 70 ) {
[svn-r8600] Purpose: Code optimization Description: Don't recompute the internal index value for looking up the chunk in the hash table, just use the value already computed from iterating through the chunks. Platforms tested: Solaris 2.7 (arabica) FreeBSD 4.9 (sleipnir) w/parallel
2004-06-01 03:59:59 +08:00
if ( buf2[i][j] != fillvalue ) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
}
else {
[svn-r8600] Purpose: Code optimization Description: Don't recompute the internal index value for looking up the chunk in the hash table, just use the value already computed from iterating through the chunks. Platforms tested: Solaris 2.7 (arabica) FreeBSD 4.9 (sleipnir) w/parallel
2004-06-01 03:59:59 +08:00
if ( buf2[i][j] != data[i][j] ) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
}
}
[svn-r5313] Purpose: added more tests to the H5set_extent function Description: there was a bug in the H5S_select_fill call when the fill value was not defined added 2 more set of tests that call H5set_extent without the fill value being defined there are now 4 sets of tests: Test H5Dset_extent with chunks on the raw data cache, with and without fill value Test H5Dset_extent with chunks written to file (b-tree routines), with and without fill value Platforms tested: w2000 linux
2002-05-01 23:54:56 +08:00
}
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/*-------------------------------------------------------------------------
* Close/release resources
*-------------------------------------------------------------------------
*/
H5Dclose( dataset_id );
H5Sclose( space_id );
PASSED();
TESTING("extend dataset read without fill value");
/* Create the data space with unlimited dimensions. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if((space_id = H5Screate_simple(RANK, dims, maxdims)) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Modify dataset creation properties, i.e. enable chunking. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if((plist_id = H5Pcreate(H5P_DATASET_CREATE)) < 0) TEST_ERROR;
if(H5Pset_chunk(plist_id, RANK, dims_chunk) < 0) TEST_ERROR;
if(H5Pset_fill_time(plist_id, H5D_FILL_TIME_ALLOC) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Create a new dataset within the file using cparms creation properties. */
[svn-r14199] Description: Add H5Dcreate to API versioned routines, replacing internal usage with H5Dcreate2 Fix thread-safe error stack initialization for API versioned error stack printing routines. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-12 00:24:11 +08:00
if((dataset_id = H5Dcreate2(file_id , "Dataset2", H5T_NATIVE_INT, space_id, H5P_DEFAULT, plist_id, H5P_DEFAULT)) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Write the data. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if(H5Dwrite(dataset_id , H5T_NATIVE_INT, space_id, H5S_ALL, H5P_DEFAULT, data) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Close/release resources. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if(H5Dclose(dataset_id) < 0) FAIL_STACK_ERROR
if(H5Sclose(space_id) < 0) FAIL_STACK_ERROR
if(H5Pclose(plist_id) < 0) FAIL_STACK_ERROR
if(H5Fclose(file_id) < 0) FAIL_STACK_ERROR
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Open the file */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if((file_id = H5Fopen("set_extent_read.h5", H5F_ACC_RDWR, H5P_DEFAULT)) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Open the dataset */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if((dataset_id = H5Dopen2(file_id , "Dataset2", H5P_DEFAULT)) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Set new dimensions for the array. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if(H5Dset_extent(dataset_id, dims_new) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Get the space. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if((space_id = H5Dget_space(dataset_id)) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Get dimensions. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if(H5Sget_simple_extent_dims(space_id, dims_out, NULL) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if(dims_out[0] != dims_new[0]) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Read the new dataset. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if(H5Dread(dataset_id, H5T_NATIVE_INT, H5S_ALL, H5S_ALL, H5P_DEFAULT, buf1) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Compare the read array with the original array */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
for(i = 0; i < (int)dims_out[0]; i++)
for(j = 0; j < (int)dims_out[1]; j++)
if(buf1[i][j] != data[i][j])
TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/*-------------------------------------------------------------------------
[svn-r11245] Purpose: Code cleanup Description: Trim trailing whitespace, which is making 'diff'ing the two branches difficult. Solution: Ran this script in each directory: foreach f (*.[ch] *.cpp) sed 's/[[:blank:]]*$//' $f > sed.out && mv sed.out $f end Platforms tested: FreeBSD 4.11 (sleipnir) Too minor to require h5committest
2005-08-14 04:53:35 +08:00
* Set new dimensions for the array; expand it again
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
*-------------------------------------------------------------------------
*/
/* Set new dimensions for the array. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if (H5Dset_extent( dataset_id , dims ) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Get the space. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if ((space_id = H5Dget_space( dataset_id )) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Get dimensions. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if (H5Sget_simple_extent_dims( space_id, dims_out, NULL ) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
[svn-r8600] Purpose: Code optimization Description: Don't recompute the internal index value for looking up the chunk in the hash table, just use the value already computed from iterating through the chunks. Platforms tested: Solaris 2.7 (arabica) FreeBSD 4.9 (sleipnir) w/parallel
2004-06-01 03:59:59 +08:00
if ( dims_out[0] != dims[0] ) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Read the new dataset. */
[svn-r14193] Description: Make H5Dopen versioned and change all internal usage to use H5Dopen2 Add simple regression test for H5Dopen1 Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode
2007-10-09 03:59:36 +08:00
if (H5Dread( dataset_id, H5T_NATIVE_INT, H5S_ALL, H5S_ALL, H5P_DEFAULT, buf2 ) < 0) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/* Compare the read array with the original array */
for( i = 0; i < (int)dims_out[0]; i++ ) {
for( j = 0; j < (int)dims_out[1]; j++ ) {
if ( i >= 70 || j >= 70 ) {
[svn-r8600] Purpose: Code optimization Description: Don't recompute the internal index value for looking up the chunk in the hash table, just use the value already computed from iterating through the chunks. Platforms tested: Solaris 2.7 (arabica) FreeBSD 4.9 (sleipnir) w/parallel
2004-06-01 03:59:59 +08:00
if ( buf2[i][j] != 0 ) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
}
else {
[svn-r8600] Purpose: Code optimization Description: Don't recompute the internal index value for looking up the chunk in the hash table, just use the value already computed from iterating through the chunks. Platforms tested: Solaris 2.7 (arabica) FreeBSD 4.9 (sleipnir) w/parallel
2004-06-01 03:59:59 +08:00
if ( buf2[i][j] != data[i][j] ) TEST_ERROR;
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
}
}
[svn-r5313] Purpose: added more tests to the H5set_extent function Description: there was a bug in the H5S_select_fill call when the fill value was not defined added 2 more set of tests that call H5set_extent without the fill value being defined there are now 4 sets of tests: Test H5Dset_extent with chunks on the raw data cache, with and without fill value Test H5Dset_extent with chunks written to file (b-tree routines), with and without fill value Platforms tested: w2000 linux
2002-05-01 23:54:56 +08:00
}
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
/*-------------------------------------------------------------------------
* Close/release resources
*-------------------------------------------------------------------------
*/
H5Dclose( dataset_id );
H5Sclose( space_id );
[svn-r5167] Purpose: turned temporarely off this test, while the rest of the code is not checked in
2002-04-11 23:54:39 +08:00
[svn-r5154] Purpose: Renamed the all the functions "set_extend" to "set_extent". Renamed the test file accordingly
2002-04-10 01:31:21 +08:00
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
H5Fclose( file_id );
[svn-r5313] Purpose: added more tests to the H5set_extent function Description: there was a bug in the H5S_select_fill call when the fill value was not defined added 2 more set of tests that call H5set_extent without the fill value being defined there are now 4 sets of tests: Test H5Dset_extent with chunks on the raw data cache, with and without fill value Test H5Dset_extent with chunks written to file (b-tree routines), with and without fill value Platforms tested: w2000 linux
2002-05-01 23:54:56 +08:00
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
PASSED();
[svn-r7109] Purpose: Code cleanup Description: Clean up varios compiler warnings flagged by SGI compiler and gcc 3.3 Platforms tested: FreeBSD 4.8 (sleipnir) h5committest
2003-06-26 10:10:33 +08:00
puts("All set_extent tests passed.");
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
return 0;
[svn-r8600] Purpose: Code optimization Description: Don't recompute the internal index value for looking up the chunk in the hash table, just use the value already computed from iterating through the chunks. Platforms tested: Solaris 2.7 (arabica) FreeBSD 4.9 (sleipnir) w/parallel
2004-06-01 03:59:59 +08:00
error:
[svn-r6825] Purpose: New feature/enhancement Description: Chunked datasets are handled poorly in several circumstances involving certain selections and chunks that are too large for the chunk cache and/or chunks with filters, causing the chunk to be read from disk multiple times. Solution: Rearrange raw data I/O infrastructure to handle chunked datasets in a much more friendly way by creating a selection in memory and on disk for each chunk in a chunked dataset and performing all of the I/O on that chunk at one time. There are still some scalability (the current code attempts to create a selection for all the chunks in the dataset, instead of just the chunks that are accessed, requiring portions of the istore.c and fillval.c tests to be commented out) and performance issues, but checking this in will allow the changes to be tested by a much wider audience while I address the remaining issues. Platforms tested: h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
2003-05-08 05:52:24 +08:00
H5Dclose( dataset_id );
H5Sclose( space_id );
H5Pclose( plist_id );
H5Fclose( file_id );
H5_FAILED();
return 1;
}
[svn-r5313] Purpose: added more tests to the H5set_extent function Description: there was a bug in the H5S_select_fill call when the fill value was not defined added 2 more set of tests that call H5set_extent without the fill value being defined there are now 4 sets of tests: Test H5Dset_extent with chunks on the raw data cache, with and without fill value Test H5Dset_extent with chunks written to file (b-tree routines), with and without fill value Platforms tested: w2000 linux
2002-05-01 23:54:56 +08:00
Reference in New Issue Copy Permalink