hdf5/src/H5Sselect.c

2395 lines
89 KiB
C
Raw Normal View History

/* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* Copyright by The HDF Group. *
* Copyright by the Board of Trustees of the University of Illinois. *
* All rights reserved. *
* *
* This file is part of HDF5. The full HDF5 copyright notice, including *
* terms governing use, modification, and redistribution, is contained in *
* the files COPYING and Copyright.html. COPYING can be found at the root *
* of the source code distribution tree; Copyright.html can be found at the *
* root level of an installed copy of the electronic HDF5 document set and *
* is linked from the top-level documents page. It can also be found at *
* http://hdfgroup.org/HDF5/doc/Copyright.html. If you do not have *
* access to either file, you may request a copy from help@hdfgroup.org. *
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
/* Programmer: Quincey Koziol <koziol@ncsa.uiuc.ued>
* Friday, May 29, 1998
*
* Purpose: Dataspace selection functions.
*/
#define H5S_PACKAGE /*suppress error about including H5Spkg */
#include "H5private.h" /* Generic Functions */
#include "H5Dprivate.h" /* Datasets */
#include "H5Eprivate.h" /* Error handling */
#include "H5FLprivate.h" /* Free Lists */
#include "H5Iprivate.h" /* IDs */
#include "H5MMprivate.h" /* Memory management */
#include "H5Spkg.h" /* Dataspaces */
#include "H5VMprivate.h" /* Vector and array functions */
#include "H5WBprivate.h" /* Wrapped Buffers */
/* Local functions */
#ifdef LATER
static herr_t H5S_select_iter_block(const H5S_sel_iter_t *iter, hsize_t *start, hsize_t *end);
static htri_t H5S_select_iter_has_next_block(const H5S_sel_iter_t *iter);
static herr_t H5S_select_iter_next_block(H5S_sel_iter_t *iter);
#endif /* LATER */
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/*--------------------------------------------------------------------------
NAME
H5S_select_offset
PURPOSE
Set the selection offset for a datapace
USAGE
herr_t H5S_select_offset(space, offset)
H5S_t *space; IN/OUT: Dataspace object to set selection offset
const hssize_t *offset; IN: Offset to position the selection at
RETURNS
Non-negative on success/Negative on failure
DESCRIPTION
Sets the selection offset for the dataspace
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
Only works for simple dataspaces currently
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
herr_t
H5S_select_offset(H5S_t *space, const hssize_t *offset)
{
FUNC_ENTER_NOAPI_NOINIT_NOERR
/* Check args */
HDassert(space);
HDassert(0 < space->extent.rank && space->extent.rank <= H5S_MAX_RANK);
HDassert(offset);
/* Copy the offset over */
HDmemcpy(space->select.offset, offset, sizeof(hssize_t)*space->extent.rank);
/* Indicate that the offset was changed */
space->select.offset_changed = TRUE;
FUNC_LEAVE_NOAPI(SUCCEED)
} /* H5S_select_offset() */
/*--------------------------------------------------------------------------
NAME
H5S_select_copy
PURPOSE
Copy a selection from one dataspace to another
USAGE
herr_t H5S_select_copy(dst, src)
H5S_t *dst; OUT: Pointer to the destination dataspace
H5S_t *src; IN: Pointer to the source dataspace
RETURNS
Non-negative on success/Negative on failure
DESCRIPTION
Copies all the selection information (include offset) from the source
dataspace to the destination dataspace.
If the SHARE_SELECTION flag is set, then the selection can be shared
between the source and destination dataspaces. (This should only occur in
situations where the destination dataspace will immediately change to a new
selection)
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
herr_t
H5S_select_copy (H5S_t *dst, const H5S_t *src, hbool_t share_selection)
{
herr_t ret_value; /* return value */
FUNC_ENTER_NOAPI(FAIL)
/* Check args */
HDassert(dst);
HDassert(src);
/* Copy regular fields */
dst->select=src->select;
/* Perform correct type of copy based on the type of selection */
if((ret_value=(*src->select.type->copy)(dst,src,share_selection))<0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTCOPY, FAIL, "can't copy selection specific information")
done:
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_copy() */
1998-07-23 18:29:44 -05:00
/*-------------------------------------------------------------------------
* Function: H5S_select_release
*
* Purpose: Releases all memory associated with a dataspace selection.
*
* Return: Non-negative on success/Negative on failure
*
* Programmer: Quincey Koziol
* Friday, May 30, 2003
*
* Note: This routine participates in the "Inlining C function pointers"
* pattern, don't call it directly, use the appropriate macro
* defined in H5Sprivate.h.
*
* Modifications:
*
*-------------------------------------------------------------------------
*/
herr_t
H5S_select_release(H5S_t *ds)
{
herr_t ret_value; /* Return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
HDassert(ds);
/* Call the selection type's release function */
ret_value=(*ds->select.type->release)(ds);
FUNC_LEAVE_NOAPI(ret_value)
} /* end H5S_select_release() */
/*-------------------------------------------------------------------------
* Function: H5S_select_get_seq_list
*
* Purpose: Retrieves the next sequence of offset/length pairs for an
* iterator on a dataspace
*
* Return: Non-negative on success/Negative on failure
*
* Programmer: Quincey Koziol
* Tuesday, May 18, 2004
*
* Note: This routine participates in the "Inlining C function pointers"
* pattern, don't call it directly, use the appropriate macro
* defined in H5Sprivate.h.
*
* Modifications:
*
*-------------------------------------------------------------------------
*/
herr_t
H5S_select_get_seq_list(const H5S_t *space, unsigned flags,
H5S_sel_iter_t *iter, size_t maxseq, size_t maxbytes,
size_t *nseq, size_t *nbytes, hsize_t *off, size_t *len)
{
herr_t ret_value; /* Return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
HDassert(space);
/* Call the selection type's get_seq_list function */
ret_value = (*space->select.type->get_seq_list)(space, flags, iter, maxseq, maxbytes, nseq, nbytes, off, len);
FUNC_LEAVE_NOAPI(ret_value)
} /* end H5S_select_get_seq_list() */
/*-------------------------------------------------------------------------
* Function: H5S_select_serial_size
*
* Purpose: Determines the number of bytes required to store the current
* selection
*
* Return: Non-negative on success/Negative on failure
*
* Programmer: Quincey Koziol
* Tuesday, May 18, 2004
*
* Note: This routine participates in the "Inlining C function pointers"
* pattern, don't call it directly, use the appropriate macro
* defined in H5Sprivate.h.
*
* Modifications:
*
*-------------------------------------------------------------------------
*/
hssize_t
H5S_select_serial_size(const H5S_t *space)
{
hssize_t ret_value; /* Return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
HDassert(space);
/* Call the selection type's serial_size function */
ret_value=(*space->select.type->serial_size)(space);
FUNC_LEAVE_NOAPI(ret_value)
} /* end H5S_select_serial_size() */
/*--------------------------------------------------------------------------
NAME
H5S_select_serialize
PURPOSE
Serialize the selection for a dataspace into a buffer
USAGE
herr_t H5S_select_serialize(space, p)
const H5S_t *space; IN: Dataspace with selection to serialize
uint8_t **p; OUT: Pointer to buffer to put serialized
selection. Will be advanced to end of
serialized selection.
RETURNS
Non-negative on success/Negative on failure
DESCRIPTION
Calls the appropriate dataspace selection callback to serialize the
current selection into a buffer.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
This routine participates in the "Inlining C function pointers"
pattern, don't call it directly, use the appropriate macro
defined in H5Sprivate.h.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
herr_t
H5S_select_serialize(const H5S_t *space, uint8_t **p)
{
herr_t ret_value=SUCCEED; /* Return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
HDassert(space);
HDassert(p);
/* Call the selection type's serialize function */
ret_value=(*space->select.type->serialize)(space,p);
FUNC_LEAVE_NOAPI(ret_value)
} /* end H5S_select_serialize() */
/*--------------------------------------------------------------------------
NAME
H5Sget_select_npoints
PURPOSE
Get the number of elements in current selection
USAGE
hssize_t H5Sget_select_npoints(dsid)
hid_t dsid; IN: Dataspace ID of selection to query
RETURNS
Non-negative on success/Negative on failure
DESCRIPTION
Returns the number of elements in current selection for dataspace.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
hssize_t
H5Sget_select_npoints(hid_t spaceid)
{
H5S_t *space; /* Dataspace to modify selection of */
hssize_t ret_value; /* return value */
FUNC_ENTER_API(FAIL)
H5TRACE1("Hs", "i", spaceid);
/* Check args */
if(NULL == (space = (H5S_t *)H5I_object_verify(spaceid, H5I_DATASPACE)))
HGOTO_ERROR(H5E_ARGS, H5E_BADTYPE, FAIL, "not a dataspace")
ret_value = (hssize_t)H5S_GET_SELECT_NPOINTS(space);
done:
FUNC_LEAVE_API(ret_value)
} /* H5Sget_select_npoints() */
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/*--------------------------------------------------------------------------
NAME
H5S_get_select_npoints
PURPOSE
Get the number of elements in current selection
USAGE
hssize_t H5Sget_select_npoints(space)
H5S_t *space; IN: Dataspace of selection to query
RETURNS
The number of elements in selection on success, 0 on failure
DESCRIPTION
Returns the number of elements in current selection for dataspace.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
This routine participates in the "Inlining C function pointers"
pattern, don't call it directly, use the appropriate macro
defined in H5Sprivate.h.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
hssize_t
H5S_get_select_npoints(const H5S_t *space)
{
FUNC_ENTER_NOAPI_NOINIT_NOERR
/* Check args */
HDassert(space);
FUNC_LEAVE_NOAPI((hssize_t)space->select.num_elem)
} /* H5S_get_select_npoints() */
/*--------------------------------------------------------------------------
NAME
H5Sselect_valid
PURPOSE
Check whether the selection fits within the extent, with the current
offset defined.
USAGE
htri_t H5Sselect_void(dsid)
hid_t dsid; IN: Dataspace ID to query
RETURNS
TRUE if the selection fits within the extent, FALSE if it does not and
Negative on an error.
DESCRIPTION
Determines if the current selection at the current offet fits within the
extent for the dataspace.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
EXAMPLES
REVISION LOG
Christian Chilan 01/17/2007
Changed the error return value from 0 to FAIL.
--------------------------------------------------------------------------*/
htri_t
[svn-r620] Changes since 19980825 ---------------------- ./MANIFEST ./src/H5R.c [NEW] ./src/H5Rprivate.h [NEW] ./src/H5Rpublic.h [NEW] ./src/Makefile.in ./src/hdf5.h ./test/ragged.c [NEW] Preliminary support for 2d ragged arrays for Mark Miller and Jim Reus. Not fully implemented yet. The test is not actually part of `make test' because we still have some memory problems. ./src/H5E.c ./src/H5Epublic.h Added H5E_RAGGED as a major error number. ./bin/release Checks the MANIFEST file against `svf ls' on systems that have it. ./bin/trace Fixed a bug that caused arguments of type `void *x[]' to not be handled. ./src/H5.c Removed unused variables and changed a couple types to fix compiler warnings. Added tracing support for ragged array object ID's and arrays of pointers. ./src/H5D.c H5Dcreate() will complain if either of the property lists are invalid (instead of using the default). ./src/H5D.c ./src/H5Dprivate.h Split H5Dget_space() into an API and internal function so it can be called from the new ragged array layer. ./src/H5Fistore.c Fixed warnings about unsigned vs. signed comparisons. ./src/H5Flow.c Fixed a warning about a variable being shadowed in the MPI-IO stuff. ./src/H5Iprivate.h ./src/H5Ipublic.h Added the H5_RAGGED atom group. ./src/H5Shyper.c Fixed some freeing-free-memory errors that resulted when certain arrays were freed but the pointers were left in the data structures. I simply set the pointers to null after they were freed. ./src/H5Sprivate.h ./src/H5Sselect.c Split the H5Sselect_hyperslab() function into an API and a private function so it could be called from the ragged array layer. Added H5S_SEL_ERROR and H5S_SEL_N to the switch statements to get rid or compiler warnings. ./src/H5Tconv.c Removed a misleading comment. ./test/bittests.c Fixed a warning about a printf(). ./test/cmpd_dset.c Fixed warnings about unused variables because of test #11 being commented out. ./bin/trace Shortened the right margin for the output to allow room for the `);' at the end of the TRACE() macros.
1998-08-27 11:48:50 -05:00
H5Sselect_valid(hid_t spaceid)
{
H5S_t *space; /* Dataspace to modify selection of */
htri_t ret_value; /* return value */
FUNC_ENTER_API(FAIL)
H5TRACE1("t", "i", spaceid);
/* Check args */
if(NULL == (space = (H5S_t *)H5I_object_verify(spaceid, H5I_DATASPACE)))
HGOTO_ERROR(H5E_ARGS, H5E_BADTYPE, FAIL, "not a dataspace")
ret_value = H5S_SELECT_VALID(space);
done:
FUNC_LEAVE_API(ret_value)
} /* H5Sselect_valid() */
/*--------------------------------------------------------------------------
NAME
H5S_select_valid
PURPOSE
Check whether the selection fits within the extent, with the current
offset defined.
USAGE
htri_t H5S_select_void(space)
H5S_t *space; IN: Dataspace to query
RETURNS
TRUE if the selection fits within the extent, FALSE if it does not and
Negative on an error.
DESCRIPTION
Determines if the current selection at the current offet fits within the
extent for the dataspace.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
This routine participates in the "Inlining C function pointers"
pattern, don't call it directly, use the appropriate macro
defined in H5Sprivate.h.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
htri_t
H5S_select_valid(const H5S_t *space)
{
htri_t ret_value; /* Return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
HDassert(space);
ret_value = (*space->select.type->is_valid)(space);
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_valid() */
/*--------------------------------------------------------------------------
NAME
H5S_select_deserialize
PURPOSE
Deserialize the current selection from a user-provided buffer into a real
selection in the dataspace.
USAGE
herr_t H5S_select_deserialize(space, p)
H5S_t **space; IN/OUT: Dataspace pointer to place
selection into. Will be allocated if not
provided.
uint8 **p; OUT: Pointer to buffer holding serialized
selection. Will be advanced to end of
serialized selection.
RETURNS
Non-negative on success/Negative on failure
DESCRIPTION
Deserializes the current selection into a buffer. (Primarily for retrieving
from disk). This routine just hands off to the appropriate routine for each
type of selection. The format of the serialized information is shown in
the H5S_select_serialize() header.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
herr_t
H5S_select_deserialize(H5S_t **space, const uint8_t **p)
{
H5S_t *tmp_space; /* Pointer to actual dataspace to use, either
*space or a newly allocated one */
uint32_t sel_type; /* Pointer to the selection type */
uint32_t version; /* Version number */
uint8_t flags = 0; /* Flags */
herr_t ret_value=FAIL; /* return value */
FUNC_ENTER_NOAPI(FAIL)
HDassert(space);
/* Allocate space if not provided */
if(!*space) {
if(NULL == (tmp_space = H5S_create(H5S_SIMPLE)))
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTCREATE, FAIL, "can't create dataspace")
} /* end if */
else
tmp_space = *space;
/* Decode selection type */
UINT32DECODE(*p, sel_type);
/* Decode version */
UINT32DECODE(*p, version);
if(version >= (uint32_t)2) {
/* Decode flags */
flags = *(*p)++;
/* Check for unknown flags */
if(flags & ~H5S_SELECT_FLAG_BITS)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTLOAD, FAIL, "unknown flag for selection")
/* Skip over the remainder of the header */
*p += 4;
} /* end if */
else
/* Skip over the remainder of the header */
*p += 8;
/* Decode and check or patch rank for point and hyperslab selections */
if((sel_type == H5S_SEL_POINTS) || (sel_type == H5S_SEL_HYPERSLABS)) {
uint32_t rank; /* Rank of dataspace */
/* Decode the rank of the point selection */
UINT32DECODE(*p,rank);
if(!*space) {
hsize_t dims[H5S_MAX_RANK];
/* Patch the rank of the allocated dataspace */
(void)HDmemset(dims, 0, (size_t)rank * sizeof(dims[0]));
if(H5S_set_extent_simple(tmp_space, rank, dims, NULL) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTINIT, FAIL, "can't set dimensions")
} /* end if */
else
/* Verify the rank of the provided dataspace */
if(rank != tmp_space->extent.rank)
HGOTO_ERROR(H5E_DATASPACE, H5E_BADRANGE, FAIL, "rank of serialized selection does not match dataspace")
} /* end if */
/* Make routine for selection type */
switch(sel_type) {
case H5S_SEL_POINTS: /* Sequence of points selected */
ret_value = (*H5S_sel_point->deserialize)(tmp_space, version, flags, p);
break;
case H5S_SEL_HYPERSLABS: /* Hyperslab selection defined */
ret_value = (*H5S_sel_hyper->deserialize)(tmp_space, version, flags, p);
break;
case H5S_SEL_ALL: /* Entire extent selected */
ret_value = (*H5S_sel_all->deserialize)(tmp_space, version, flags, p);
break;
case H5S_SEL_NONE: /* Nothing selected */
ret_value = (*H5S_sel_none->deserialize)(tmp_space, version, flags, p);
break;
default:
break;
}
if(ret_value < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTLOAD, FAIL, "can't deserialize selection")
/* Return space to the caller if allocated */
if(!*space)
*space = tmp_space;
done:
/* Free temporary space if not passed to caller (only happens on error) */
if(!*space && tmp_space)
if(H5S_close(tmp_space) < 0)
HDONE_ERROR(H5E_DATASPACE, H5E_CANTFREE, FAIL, "can't close dataspace")
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_deserialize() */
/*--------------------------------------------------------------------------
NAME
H5Sget_select_bounds
PURPOSE
Gets the bounding box containing the selection.
USAGE
herr_t H5S_get_select_bounds(space, start, end)
hid_t dsid; IN: Dataspace ID of selection to query
hsize_t start[]; OUT: Starting coordinate of bounding box
hsize_t end[]; OUT: Opposite coordinate of bounding box
RETURNS
Non-negative on success, negative on failure
DESCRIPTION
Retrieves the bounding box containing the current selection and places
it into the user's buffers. The start and end buffers must be large
enough to hold the dataspace rank number of coordinates. The bounding box
exactly contains the selection, ie. if a 2-D element selection is currently
defined with the following points: (4,5), (6,8) (10,7), the bounding box
with be (4, 5), (10, 8). Calling this function on a "none" selection
returns fail.
The bounding box calculations _does_ include the current offset of the
selection within the dataspace extent.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
This routine participates in the "Inlining C function pointers"
pattern, don't call it directly, use the appropriate macro
defined in H5Sprivate.h.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
herr_t
H5Sget_select_bounds(hid_t spaceid, hsize_t start[], hsize_t end[])
{
H5S_t *space; /* Dataspace to modify selection of */
herr_t ret_value; /* return value */
FUNC_ENTER_API(FAIL)
H5TRACE3("e", "i*h*h", spaceid, start, end);
/* Check args */
if(start == NULL || end == NULL)
HGOTO_ERROR(H5E_ARGS, H5E_BADVALUE, FAIL, "invalid pointer")
if(NULL == (space = (H5S_t *)H5I_object_verify(spaceid, H5I_DATASPACE)))
HGOTO_ERROR(H5E_ARGS, H5E_BADTYPE, FAIL, "not a dataspace")
ret_value = H5S_SELECT_BOUNDS(space, start, end);
done:
FUNC_LEAVE_API(ret_value)
} /* H5Sget_select_bounds() */
/*--------------------------------------------------------------------------
NAME
H5S_get_select_bounds
PURPOSE
Gets the bounding box containing the selection.
USAGE
herr_t H5S_get_select_bounds(space, start, end)
H5S_t *space; IN: Dataspace ID of selection to query
hsize_t *start; OUT: Starting coordinate of bounding box
hsize_t *end; OUT: Opposite coordinate of bounding box
RETURNS
Non-negative on success, negative on failure
DESCRIPTION
Retrieves the bounding box containing the current selection and places
it into the user's buffers. The start and end buffers must be large
enough to hold the dataspace rank number of coordinates. The bounding box
exactly contains the selection, ie. if a 2-D element selection is currently
defined with the following points: (4,5), (6,8) (10,7), the bounding box
with be (4, 5), (10, 8). Calling this function on a "none" selection
returns fail.
The bounding box calculations _does_ include the current offset of the
selection within the dataspace extent.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
herr_t
H5S_get_select_bounds(const H5S_t *space, hsize_t *start, hsize_t *end)
{
herr_t ret_value; /* return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
/* Check args */
HDassert(space);
HDassert(start);
HDassert(end);
ret_value = (*space->select.type->bounds)(space,start,end);
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_get_select_bounds() */
/*--------------------------------------------------------------------------
NAME
H5S_get_select_offset
PURPOSE
Gets the linear offset of the first element for the selection.
USAGE
herr_t H5S_get_select_offset(space, offset)
const H5S_t *space; IN: Dataspace pointer of selection to query
hsize_t *offset; OUT: Linear offset of first element in selection
RETURNS
Non-negative on success, negative on failure
DESCRIPTION
Retrieves the linear offset (in "units" of elements) of the first element
selected within the dataspace.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
The offset calculation _does_ include the current offset of the
selection within the dataspace extent.
Calling this function on a "none" selection returns fail.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
herr_t
H5S_get_select_offset(const H5S_t *space, hsize_t *offset)
{
herr_t ret_value; /* return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
/* Check args */
HDassert(space);
HDassert(offset);
ret_value = (*space->select.type->offset)(space, offset);
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_get_select_offset() */
/*--------------------------------------------------------------------------
NAME
H5S_get_select_unlim_dim
PURPOSE
Gets the unlimited dimension in the selection, or -1 if there is no
unlimited dimension.
USAGE
int H5S_get_select_unlim_dim(space)
const H5S_t *space; IN: Dataspace pointer of selection to query
RETURNS
Unlimited dimension in the selection, or -1 if there is no unlimited
dimension (never fails)
DESCRIPTION
Gets the unlimited dimension in the selection, or -1 if there is no
unlimited dimension.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
Currently only implemented for hyperslab selections, all others
simply return -1.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
int
H5S_get_select_unlim_dim(const H5S_t *space)
{
herr_t ret_value; /* return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
/* Check args */
HDassert(space);
ret_value = (*space->select.type->unlim_dim)(space);
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_get_select_unlim_dim() */
/*--------------------------------------------------------------------------
NAME
H5S_get_select_num_elem_non_unlim
PURPOSE
Gets the number of elements in the non-unlimited dimensions
USAGE
herr_t H5S_get_select_num_elem_non_unlim(space,num_elem_non_unlim)
H5S_t *space; IN: Dataspace pointer to check
hsize_t *num_elem_non_unlim; OUT: Number of elements in the non-unlimited dimensions
RETURNS
Non-negative on success/Negative on failure
DESCRIPTION
Returns the number of elements in a slice through the non-unlimited
dimensions of the selection. Fails if the selection has no unlimited
dimension.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
herr_t
H5S_get_select_num_elem_non_unlim(const H5S_t *space,
hsize_t *num_elem_non_unlim)
{
herr_t ret_value = SUCCEED; /* return value */
FUNC_ENTER_NOAPI(FAIL)
/* Check args */
HDassert(space);
HDassert(num_elem_non_unlim);
/* Check for selection callback */
if(!space->select.type->num_elem_non_unlim)
HGOTO_ERROR(H5E_DATASPACE, H5E_UNSUPPORTED, FAIL, "selection type has no num_elem_non_unlim callback")
/* Make selection callback */
if((*space->select.type->num_elem_non_unlim)(space, num_elem_non_unlim) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTCOUNT, FAIL, "can't get number of elements in non-unlimited dimension")
done:
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_get_select_unlim_dim() */
/*--------------------------------------------------------------------------
NAME
H5S_select_is_contiguous
PURPOSE
Determines if a selection is contiguous in the dataspace
USAGE
htri_t H5S_select_is_contiguous(space)
const H5S_t *space; IN: Dataspace of selection to query
RETURNS
Non-negative (TRUE/FALSE) on success, negative on failure
DESCRIPTION
Checks the selection to determine if the points to iterated over will be
contiguous in the particular dataspace.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
This routine participates in the "Inlining C function pointers"
pattern, don't call it directly, use the appropriate macro
defined in H5Sprivate.h.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
htri_t
H5S_select_is_contiguous(const H5S_t *space)
{
herr_t ret_value; /* return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
/* Check args */
HDassert(space);
ret_value = (*space->select.type->is_contiguous)(space);
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_is_contiguous() */
/*--------------------------------------------------------------------------
NAME
H5S_select_is_single
PURPOSE
Determines if a selection is a single block in the dataspace
USAGE
htri_t H5S_select_is_single(space)
const H5S_t *space; IN: Dataspace of selection to query
RETURNS
Non-negative (TRUE/FALSE) on success, negative on failure
DESCRIPTION
Checks the selection to determine if it occupies a single block in the
particular dataspace.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
This routine participates in the "Inlining C function pointers"
pattern, don't call it directly, use the appropriate macro
defined in H5Sprivate.h.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
htri_t
H5S_select_is_single(const H5S_t *space)
{
herr_t ret_value; /* return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
/* Check args */
HDassert(space);
ret_value = (*space->select.type->is_single)(space);
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_is_single() */
/*--------------------------------------------------------------------------
NAME
H5S_select_is_regular
PURPOSE
Determines if a selection is "regular" in the dataspace
USAGE
htri_t H5S_select_is_regular(space)
const H5S_t *space; IN: Dataspace of selection to query
RETURNS
Non-negative (TRUE/FALSE) on success, negative on failure
DESCRIPTION
Checks the selection to determine if it is "regular" (i.e. a single
block or a strided pattern) in the particular dataspace.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
This routine participates in the "Inlining C function pointers"
pattern, don't call it directly, use the appropriate macro
defined in H5Sprivate.h.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
htri_t
H5S_select_is_regular(const H5S_t *space)
{
herr_t ret_value; /* return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
/* Check args */
HDassert(space);
ret_value = (*space->select.type->is_regular)(space);
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_is_regular() */
/*--------------------------------------------------------------------------
NAME
H5S_select_adjust_u
PURPOSE
Adjust a selection by subtracting an offset
USAGE
herr_t H5S_select_adjust_u(space, offset)
H5S_t *space; IN/OUT: Pointer to dataspace to adjust
const hsize_t *offset; IN: Offset to subtract
RETURNS
Non-negative on success, negative on failure
DESCRIPTION
Moves a selection by subtracting an offset from it.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
This routine participates in the "Inlining C function pointers"
pattern, don't call it directly, use the appropriate macro
defined in H5Sprivate.h.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
herr_t
H5S_select_adjust_u(H5S_t *space, const hsize_t *offset)
{
herr_t ret_value; /* return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
/* Check args */
HDassert(space);
ret_value = (*space->select.type->adjust_u)(space, offset);
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_adjust_u() */
/*--------------------------------------------------------------------------
NAME
H5S_select_project_scalar
PURPOSE
Project a single element selection for a scalar dataspace
USAGE
herr_t H5S_select_project_scalar(space, offset)
const H5S_t *space; IN: Pointer to dataspace to project
hsize_t *offset; IN/OUT: Offset of projected point
RETURNS
Non-negative on success, negative on failure
DESCRIPTION
Projects a selection of a single element into a scalar dataspace, computing
the offset of the element in the original selection.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
This routine participates in the "Inlining C function pointers"
pattern, don't call it directly, use the appropriate macro
defined in H5Sprivate.h.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
herr_t
H5S_select_project_scalar(const H5S_t *space, hsize_t *offset)
{
herr_t ret_value; /* Return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
/* Check args */
HDassert(space);
HDassert(offset);
ret_value = (*space->select.type->project_scalar)(space, offset);
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_project_scalar() */
/*--------------------------------------------------------------------------
NAME
H5S_select_project_simple
PURPOSE
Project a selection onto/into a dataspace of different rank
USAGE
herr_t H5S_select_project_simple(space, new_space, offset)
const H5S_t *space; IN: Pointer to dataspace to project
H5S_t *new_space; IN/OUT: Pointer to dataspace projected onto
hsize_t *offset; IN/OUT: Offset of projected point
RETURNS
Non-negative on success, negative on failure
DESCRIPTION
Projects a selection onto/into a simple dataspace, computing
the offset of the first element in the original selection.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
This routine participates in the "Inlining C function pointers"
pattern, don't call it directly, use the appropriate macro
defined in H5Sprivate.h.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
herr_t
H5S_select_project_simple(const H5S_t *space, H5S_t *new_space, hsize_t *offset)
{
herr_t ret_value; /* Return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
/* Check args */
HDassert(space);
HDassert(new_space);
HDassert(offset);
ret_value = (*space->select.type->project_simple)(space, new_space, offset);
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_project_simple() */
/*--------------------------------------------------------------------------
NAME
H5S_select_iter_init
PURPOSE
Initializes iteration information for a selection.
USAGE
herr_t H5S_select_iter_init(sel_iter, space, elmt_size)
H5S_sel_iter_t *sel_iter; OUT: Selection iterator to initialize.
H5S_t *space; IN: Dataspace object containing selection to
iterate over
size_t elmt_size; IN: Size of elements in the selection
RETURNS
Non-negative on success, negative on failure.
DESCRIPTION
Initialize the selection iterator object to point to the first element
in the dataspace's selection.
--------------------------------------------------------------------------*/
herr_t
H5S_select_iter_init(H5S_sel_iter_t *sel_iter, const H5S_t *space, size_t elmt_size)
{
herr_t ret_value; /* Return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
/* Check args */
[svn-r20536] Description: Clean up various warnings & code formatting issues. Bring changes from Coverity branch to trunk: r20085: Purpose: Fix coverity issue 793 Description: Modified H5S_hyper_project_simple_higher() to free the entire span list in new_space on failure. r20091: This is a fix for coverity bug #1683. Changed the two printfs to use %lu (unsigned long) for printing "dset_size". r20162: Purpose: Fix coverity issue 785 Description: Modified H5T_enum_nameof() to free "name" on failure if it was allocated. Also clarified some code in H5S_hyper_rebuild_helper(). r20189: Addressed coverity defect 783. H5SL_new_node() in H5SL.c was failing to free space allocated in its first alloc if the second alloc failed. Added a call to H5FL_FREE to address this issue. This is purely to keep coverity happy -- if this code is ever triggered, we have much larger problems. Note that this fix will trigger an unused return value complaint from coverity next week. r20190: Fixed Coverity issues 1561 1565 and 1678 (UNUSED_VALUES) by moving checks of return values to after the function call. r20191: Fixed coverity issues 643 644 and 1678 (CHECKED_RETURN). r20232: Addressed coverity issues 923-925. Replaced calls to sprintf with calls to HDsnprintf. r20233: Fix coverity issue 662. Don't try to sort 0 attributes in H5Aint.c. r20234: Fix coverity issue 664. Check for NULL before dereferencing in H5Gdeprec.c. r20271: Purpose: Fix coverity issue 784 Description: Modified H5_debug_mask() to keep a list of files opened for use as a debugging output stream, and modified H5_term_library to close these files on exit. r20272: addressed coverity issues 838 & 955. Issue was use of strcpy() -- existing code was safe, but modified to use strncpy() to keep coverity happy. r20273: Addresed coverity issues 1388 and 1389. Initialized sel_iter->type to NULL in H5S_select_iter_init. r20275: Purpose: Fix valgrind issue in mf.c Description: Fixed bug (incomplete if statement) in test_mf_fs_alloc_free() so the retrieved node gets freed. Tested on: FreeBSD/32 6.3 (duty) in debug mode FreeBSD/64 6.3 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (jam) w/PGI compilers, w/default API=1.8.x, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (amani) w/Intel compilers, w/default API=1.6.x, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, w/threadsafe, in production mode Linux/PPC 2.6 (heiwa) w/C++ & FORTRAN, w/threadsafe, in debug mode
2011-04-17 13:57:07 -05:00
HDassert(sel_iter);
HDassert(space);
/* Initialize common information */
/* Save the dataspace's rank */
[svn-r20536] Description: Clean up various warnings & code formatting issues. Bring changes from Coverity branch to trunk: r20085: Purpose: Fix coverity issue 793 Description: Modified H5S_hyper_project_simple_higher() to free the entire span list in new_space on failure. r20091: This is a fix for coverity bug #1683. Changed the two printfs to use %lu (unsigned long) for printing "dset_size". r20162: Purpose: Fix coverity issue 785 Description: Modified H5T_enum_nameof() to free "name" on failure if it was allocated. Also clarified some code in H5S_hyper_rebuild_helper(). r20189: Addressed coverity defect 783. H5SL_new_node() in H5SL.c was failing to free space allocated in its first alloc if the second alloc failed. Added a call to H5FL_FREE to address this issue. This is purely to keep coverity happy -- if this code is ever triggered, we have much larger problems. Note that this fix will trigger an unused return value complaint from coverity next week. r20190: Fixed Coverity issues 1561 1565 and 1678 (UNUSED_VALUES) by moving checks of return values to after the function call. r20191: Fixed coverity issues 643 644 and 1678 (CHECKED_RETURN). r20232: Addressed coverity issues 923-925. Replaced calls to sprintf with calls to HDsnprintf. r20233: Fix coverity issue 662. Don't try to sort 0 attributes in H5Aint.c. r20234: Fix coverity issue 664. Check for NULL before dereferencing in H5Gdeprec.c. r20271: Purpose: Fix coverity issue 784 Description: Modified H5_debug_mask() to keep a list of files opened for use as a debugging output stream, and modified H5_term_library to close these files on exit. r20272: addressed coverity issues 838 & 955. Issue was use of strcpy() -- existing code was safe, but modified to use strncpy() to keep coverity happy. r20273: Addresed coverity issues 1388 and 1389. Initialized sel_iter->type to NULL in H5S_select_iter_init. r20275: Purpose: Fix valgrind issue in mf.c Description: Fixed bug (incomplete if statement) in test_mf_fs_alloc_free() so the retrieved node gets freed. Tested on: FreeBSD/32 6.3 (duty) in debug mode FreeBSD/64 6.3 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (jam) w/PGI compilers, w/default API=1.8.x, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (amani) w/Intel compilers, w/default API=1.6.x, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, w/threadsafe, in production mode Linux/PPC 2.6 (heiwa) w/C++ & FORTRAN, w/threadsafe, in debug mode
2011-04-17 13:57:07 -05:00
sel_iter->rank = space->extent.rank;
/* Point to the dataspace dimensions, if there are any */
if(sel_iter->rank > 0)
sel_iter->dims = space->extent.size;
else
sel_iter->dims = NULL;
/* Save the element size */
[svn-r20536] Description: Clean up various warnings & code formatting issues. Bring changes from Coverity branch to trunk: r20085: Purpose: Fix coverity issue 793 Description: Modified H5S_hyper_project_simple_higher() to free the entire span list in new_space on failure. r20091: This is a fix for coverity bug #1683. Changed the two printfs to use %lu (unsigned long) for printing "dset_size". r20162: Purpose: Fix coverity issue 785 Description: Modified H5T_enum_nameof() to free "name" on failure if it was allocated. Also clarified some code in H5S_hyper_rebuild_helper(). r20189: Addressed coverity defect 783. H5SL_new_node() in H5SL.c was failing to free space allocated in its first alloc if the second alloc failed. Added a call to H5FL_FREE to address this issue. This is purely to keep coverity happy -- if this code is ever triggered, we have much larger problems. Note that this fix will trigger an unused return value complaint from coverity next week. r20190: Fixed Coverity issues 1561 1565 and 1678 (UNUSED_VALUES) by moving checks of return values to after the function call. r20191: Fixed coverity issues 643 644 and 1678 (CHECKED_RETURN). r20232: Addressed coverity issues 923-925. Replaced calls to sprintf with calls to HDsnprintf. r20233: Fix coverity issue 662. Don't try to sort 0 attributes in H5Aint.c. r20234: Fix coverity issue 664. Check for NULL before dereferencing in H5Gdeprec.c. r20271: Purpose: Fix coverity issue 784 Description: Modified H5_debug_mask() to keep a list of files opened for use as a debugging output stream, and modified H5_term_library to close these files on exit. r20272: addressed coverity issues 838 & 955. Issue was use of strcpy() -- existing code was safe, but modified to use strncpy() to keep coverity happy. r20273: Addresed coverity issues 1388 and 1389. Initialized sel_iter->type to NULL in H5S_select_iter_init. r20275: Purpose: Fix valgrind issue in mf.c Description: Fixed bug (incomplete if statement) in test_mf_fs_alloc_free() so the retrieved node gets freed. Tested on: FreeBSD/32 6.3 (duty) in debug mode FreeBSD/64 6.3 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (jam) w/PGI compilers, w/default API=1.8.x, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (amani) w/Intel compilers, w/default API=1.6.x, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, w/threadsafe, in production mode Linux/PPC 2.6 (heiwa) w/C++ & FORTRAN, w/threadsafe, in debug mode
2011-04-17 13:57:07 -05:00
sel_iter->elmt_size = elmt_size;
/* Call initialization routine for selection type */
[svn-r20536] Description: Clean up various warnings & code formatting issues. Bring changes from Coverity branch to trunk: r20085: Purpose: Fix coverity issue 793 Description: Modified H5S_hyper_project_simple_higher() to free the entire span list in new_space on failure. r20091: This is a fix for coverity bug #1683. Changed the two printfs to use %lu (unsigned long) for printing "dset_size". r20162: Purpose: Fix coverity issue 785 Description: Modified H5T_enum_nameof() to free "name" on failure if it was allocated. Also clarified some code in H5S_hyper_rebuild_helper(). r20189: Addressed coverity defect 783. H5SL_new_node() in H5SL.c was failing to free space allocated in its first alloc if the second alloc failed. Added a call to H5FL_FREE to address this issue. This is purely to keep coverity happy -- if this code is ever triggered, we have much larger problems. Note that this fix will trigger an unused return value complaint from coverity next week. r20190: Fixed Coverity issues 1561 1565 and 1678 (UNUSED_VALUES) by moving checks of return values to after the function call. r20191: Fixed coverity issues 643 644 and 1678 (CHECKED_RETURN). r20232: Addressed coverity issues 923-925. Replaced calls to sprintf with calls to HDsnprintf. r20233: Fix coverity issue 662. Don't try to sort 0 attributes in H5Aint.c. r20234: Fix coverity issue 664. Check for NULL before dereferencing in H5Gdeprec.c. r20271: Purpose: Fix coverity issue 784 Description: Modified H5_debug_mask() to keep a list of files opened for use as a debugging output stream, and modified H5_term_library to close these files on exit. r20272: addressed coverity issues 838 & 955. Issue was use of strcpy() -- existing code was safe, but modified to use strncpy() to keep coverity happy. r20273: Addresed coverity issues 1388 and 1389. Initialized sel_iter->type to NULL in H5S_select_iter_init. r20275: Purpose: Fix valgrind issue in mf.c Description: Fixed bug (incomplete if statement) in test_mf_fs_alloc_free() so the retrieved node gets freed. Tested on: FreeBSD/32 6.3 (duty) in debug mode FreeBSD/64 6.3 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (jam) w/PGI compilers, w/default API=1.8.x, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (amani) w/Intel compilers, w/default API=1.6.x, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, w/threadsafe, in production mode Linux/PPC 2.6 (heiwa) w/C++ & FORTRAN, w/threadsafe, in debug mode
2011-04-17 13:57:07 -05:00
ret_value = (*space->select.type->iter_init)(sel_iter, space);
HDassert(sel_iter->type);
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_iter_init() */
/*--------------------------------------------------------------------------
NAME
H5S_select_iter_coords
PURPOSE
Get the coordinates of the current iterator position
USAGE
herr_t H5S_select_iter_coords(sel_iter,coords)
H5S_sel_iter_t *sel_iter; IN: Selection iterator to query
hsize_t *coords; OUT: Array to place iterator coordinates in
RETURNS
Non-negative on success, negative on failure.
DESCRIPTION
The current location of the iterator within the selection is placed in
the COORDS array.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
This routine participates in the "Inlining C function pointers"
pattern, don't call it directly, use the appropriate macro
defined in H5Sprivate.h.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
herr_t
H5S_select_iter_coords(const H5S_sel_iter_t *sel_iter, hsize_t *coords)
{
herr_t ret_value; /* return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
/* Check args */
HDassert(sel_iter);
HDassert(coords);
/* Call iter_coords routine for selection type */
ret_value = (*sel_iter->type->iter_coords)(sel_iter,coords);
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_iter_coords() */
#ifdef LATER
/*--------------------------------------------------------------------------
NAME
H5S_select_iter_block
PURPOSE
Get the block of the current iterator position
USAGE
herr_t H5S_select_iter_block(sel_iter,start,end)
const H5S_sel_iter_t *sel_iter; IN: Selection iterator to query
hsize_t *start; OUT: Array to place iterator start block coordinates
hsize_t *end; OUT: Array to place iterator end block coordinates
RETURNS
Non-negative on success, negative on failure.
DESCRIPTION
The current location of the iterator within the selection is placed in
the COORDS array.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
This routine participates in the "Inlining C function pointers"
pattern, don't call it directly, use the appropriate macro
defined in H5Sprivate.h.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
static herr_t
H5S_select_iter_block(const H5S_sel_iter_t *iter, hsize_t *start, hsize_t *end)
{
herr_t ret_value; /* return value */
FUNC_ENTER_NOAPI_NOINIT_NOINIT_NOERR
/* Check args */
HDassert(iter);
HDassert(start);
HDassert(end);
/* Call iter_block routine for selection type */
ret_value = (*iter->type->iter_block)(iter,start,end);
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_iter_block() */
#endif /* LATER */
/*--------------------------------------------------------------------------
NAME
H5S_select_iter_nelmts
PURPOSE
Get the number of elements left to iterate over in selection
USAGE
hssize_t H5S_select_iter_nelmts(sel_iter)
H5S_sel_iter_t *sel_iter; IN: Selection iterator to query
RETURNS
The number of elements in selection on success, 0 on failure
DESCRIPTION
Returns the number of elements in current selection for dataspace.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
This routine participates in the "Inlining C function pointers"
pattern, don't call it directly, use the appropriate macro
defined in H5Sprivate.h.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
hsize_t
H5S_select_iter_nelmts(const H5S_sel_iter_t *sel_iter)
{
hsize_t ret_value; /* return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
/* Check args */
HDassert(sel_iter);
/* Call iter_nelmts routine for selection type */
ret_value = (*sel_iter->type->iter_nelmts)(sel_iter);
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_iter_nelmts() */
#ifdef LATER
/*--------------------------------------------------------------------------
NAME
H5S_select_iter_has_next_block
PURPOSE
Check if there is another block available in the selection iterator
USAGE
htri_t H5S_select_iter_has_next_block(sel_iter)
const H5S_sel_iter_t *sel_iter; IN: Selection iterator to query
RETURNS
Non-negative on success, negative on failure.
DESCRIPTION
Check if there is another block available to advance to in the selection
iterator.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
This routine participates in the "Inlining C function pointers"
pattern, don't call it directly, use the appropriate macro
defined in H5Sprivate.h.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
static htri_t
H5S_select_iter_has_next_block(const H5S_sel_iter_t *iter)
{
herr_t ret_value; /* return value */
FUNC_ENTER_NOAPI_NOINIT_NOINIT_NOERR
/* Check args */
HDassert(iter);
/* Call iter_has_next_block routine for selection type */
ret_value = (*iter->type->iter_has_next_block)(iter);
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_iter_has_next_block() */
#endif /* LATER */
/*--------------------------------------------------------------------------
NAME
H5S_select_iter_next
PURPOSE
Advance selection iterator to next element
USAGE
herr_t H5S_select_iter_next(iter, nelem)
H5S_sel_iter_t *iter; IN/OUT: Selection iterator to change
size_t nelem; IN: Number of elements to advance by
RETURNS
Non-negative on success, negative on failure.
DESCRIPTION
Move the current element for the selection iterator to the NELEM'th next
element in the selection.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
This routine participates in the "Inlining C function pointers"
pattern, don't call it directly, use the appropriate macro
defined in H5Sprivate.h.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
herr_t
H5S_select_iter_next(H5S_sel_iter_t *iter, size_t nelem)
{
herr_t ret_value; /* return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
/* Check args */
HDassert(iter);
HDassert(nelem>0);
/* Call iter_next routine for selection type */
ret_value = (*iter->type->iter_next)(iter,nelem);
/* Decrement the number of elements left in selection */
iter->elmt_left-=nelem;
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_iter_next() */
#ifdef LATER
/*--------------------------------------------------------------------------
NAME
H5S_select_iter_next_block
PURPOSE
Advance selection iterator to next block
USAGE
herr_t H5S_select_iter_next_block(iter)
H5S_sel_iter_t *iter; IN/OUT: Selection iterator to change
RETURNS
Non-negative on success, negative on failure.
DESCRIPTION
Move the current element for the selection iterator to the next
block in the selection.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
Doesn't maintain the 'elmt_left' field of the selection iterator.
This routine participates in the "Inlining C function pointers"
pattern, don't call it directly, use the appropriate macro
defined in H5Sprivate.h.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
static herr_t
H5S_select_iter_next_block(H5S_sel_iter_t *iter)
{
herr_t ret_value; /* return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
/* Check args */
HDassert(iter);
/* Call iter_next_block routine for selection type */
ret_value = (*iter->type->iter_next_block)(iter);
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_iter_next_block() */
#endif /* LATER */
/*--------------------------------------------------------------------------
NAME
H5S_select_iter_release
PURPOSE
Release a selection iterator's resources.
USAGE
hssize_t H5S_select_iter_release(sel_iter)
H5S_sel_iter_t *sel_iter; IN: Selection iterator to query
RETURNS
The number of elements in selection on success, 0 on failure
DESCRIPTION
Returns the number of elements in current selection for dataspace.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
This routine participates in the "Inlining C function pointers"
pattern, don't call it directly, use the appropriate macro
defined in H5Sprivate.h.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
herr_t
H5S_select_iter_release(H5S_sel_iter_t *sel_iter)
{
herr_t ret_value; /* return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
/* Check args */
HDassert(sel_iter);
/* Call selection type-specific release routine */
ret_value = (*sel_iter->type->iter_release)(sel_iter);
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_iter_release() */
/*--------------------------------------------------------------------------
NAME
H5S_select_iterate
PURPOSE
Iterate over the selected elements in a memory buffer.
USAGE
herr_t H5S_select_iterate(buf, type_id, space, operator, operator_data)
void *buf; IN/OUT: Buffer containing elements to iterate over
hid_t type_id; IN: Datatype ID of BUF array.
H5S_t *space; IN: Dataspace object containing selection to iterate over
H5D_operator_t op; IN: Function pointer to the routine to be
called for each element in BUF iterated over.
void *operator_data; IN/OUT: Pointer to any user-defined data
associated with the operation.
RETURNS
Returns the return value of the last operator if it was non-zero, or zero
if all elements were processed. Otherwise returns a negative value.
DESCRIPTION
Iterates over the selected elements in a memory buffer, calling the user's
callback function for each element. The selection in the dataspace is
modified so that any elements already iterated over are removed from the
selection if the iteration is interrupted (by the H5D_operator_t function
returning non-zero) in the "middle" of the iteration and may be re-started
by the user where it left off.
NOTE: Until "subtracting" elements from a selection is implemented,
the selection is not modified.
--------------------------------------------------------------------------*/
herr_t
H5S_select_iterate(void *buf, hid_t type_id, const H5S_t *space, H5D_operator_t op,
void *operator_data)
{
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
H5T_t *dt; /* Datatype structure */
[svn-r6252] Purpose: Lots of performance improvements & a couple new internal API interfaces. Description: Performance Improvements: - Cached file offset & length sizes in shared file struct, to avoid constantly looking them up in the FCPL. - Generic property improvements: - Added "revision" number to generic property classes to speed up comparisons. - Changed method of storing properties from using a hash-table to the TBBT routines in the library. - Share the propery names between classes and the lists derived from them. - Removed redundant 'def_value' buffer from each property. - Switching code to use a "copy on write" strategy for properties in each list, where the properties in each list are shared with the properties in the class, until a property's value is changed in a list. - Fixed error in layout code which was allocating too many buffers. - Redefined public macros of the form (H5open()/H5check, <variable>) internally to only be (<variable>), avoiding innumerable useless calls to H5open() and H5check_version(). - Reuse already zeroed buffers in H5F_contig_fill instead of constantly re-zeroing them. - Don't write fill values if writing entire dataset. - Use gettimeofday() system call instead of time() system when checking the modification time of a dataset. - Added reference counted string API and use it for tracking the names of objects opening in a file (for the ID->name code). - Removed redundant H5P_get() calls in B-tree routines. - Redefine H5T datatype macros internally to the library, to avoid calling H5check redundantly. - Keep dataspace information for dataset locally instead of reading from disk each time. Added new module to track open objects in a file, to allow this (which will be useful eventually for some FPH5 metadata caching issues). - Remove H5AC_find macro which was inlining metadata cache lookups, and call function instead. - Remove redundant memset() calls from H5G_namei() routine. - Remove redundant checking of object type when locating objects in metadata cache and rely on the address only. - Create default dataset object to use when default dataset creation property list is used to create datasets, bypassing querying for all the property list values. - Use default I/O vector size when performing raw data with the default dataset transfer property list, instead of querying for I/O vector size. - Remove H5P_DEFAULT internally to the library, replacing it with more specific default property list based on the type of property list needed. - Remove redundant memset() calls in object header message (H5O*) routines. - Remove redunant memset() calls in data I/O routines. - Split free-list allocation routines into malloc() and calloc()- like routines, instead of one combined routine. - Remove lots of indirection in H5O*() routines. - Simplify metadata cache entry comparison routine (used when flushing entire cache out). - Only enable metadata cache statistics when H5AC_DEBUG is turned on, instead of always tracking them. - Simplify address comparison macro (H5F_addr_eq). - Remove redundant metadata cache entry protections during dataset creation by protecting the object header once and making all the modifications necessary for the dataset creation before unprotecting it. - Reduce # of "number of element in extent" computations performed by computing and storing the value during dataspace creation. - Simplify checking for group location's file information, when file has not been involving in file-mounting operations. - Use binary encoding for modification time, instead of ASCII. - Hoist H5HL_peek calls (to get information in a local heap) out of loops in many group routine. - Use static variable for iterators of selections, instead of dynamically allocation them each time. - Lookup & insert new entries in one step, avoiding traversing group's B-tree twice. - Fixed memory leak in H5Gget_objname_idx() routine (tangential to performance improvements, but fixed along the way). - Use free-list for reference counted strings. - Don't bother copying object names into cached group entries, since they are re-created when an object is opened. The benchmark I used to measure these results created several thousand small (2K) datasets in a file and wrote out the data for them. This is Elena's "regular.c" benchmark. These changes resulted in approximately ~4.3x speedup of the development branch when compared to the previous code in the development branch and ~1.4x speedup compared to the release branch. Additionally, these changes reduce the total memory used (code and data) by the development branch by ~800KB, bringing the development branch back into the same ballpark as the release branch. I'll send out a more detailed description of the benchmark results as a followup note. New internal API routines: Added "reference counted strings" API for tracking strings that get used by multiple owners without duplicating the strings. Added "ternary search tree" API for text->object mappings. Platforms tested: Tested h5committest {arabica (fortran), eirene (fortran, C++) modi4 (parallel, fortran)} Other platforms/configurations tested? FreeBSD 4.7 (sleipnir) serial & parallel Solaris 2.6 (baldric) serial
2003-01-09 12:20:03 -05:00
H5S_sel_iter_t iter; /* Selection iteration info */
hbool_t iter_init = FALSE; /* Selection iteration info has been initialized */
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
hssize_t nelmts; /* Number of elements in selection */
hsize_t space_size[H5O_LAYOUT_NDIMS]; /* Dataspace size */
size_t max_elem; /* Maximum number of elements allowed in sequences */
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
size_t elmt_size; /* Datatype size */
unsigned ndims; /* Number of dimensions in dataspace */
herr_t user_ret = 0; /* User's return value */
herr_t ret_value = SUCCEED; /* Return value */
FUNC_ENTER_NOAPI(FAIL)
/* Check args */
HDassert(buf);
HDassert(H5I_DATATYPE == H5I_get_type(type_id));
HDassert(space);
HDassert(op);
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Get the datatype size */
if(NULL == (dt = (H5T_t *)H5I_object_verify(type_id, H5I_DATATYPE)))
HGOTO_ERROR(H5E_ARGS, H5E_BADTYPE, FAIL, "not an valid base datatype")
if(0 == (elmt_size = H5T_get_size(dt)))
HGOTO_ERROR(H5E_DATATYPE, H5E_BADSIZE, FAIL, "datatype size invalid")
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Initialize iterator */
if(H5S_select_iter_init(&iter, space, elmt_size) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTINIT, FAIL, "unable to initialize selection iterator")
iter_init = TRUE; /* Selection iteration info has been initialized */
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Get the number of elements in selection */
if((nelmts = (hssize_t)H5S_GET_SELECT_NPOINTS(space)) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTCOUNT, FAIL, "can't get number of elements selected")
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Get the rank of the dataspace */
ndims = space->extent.rank;
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
if(ndims > 0) {
/* Copy the size of the space */
HDassert(space->extent.size);
HDmemcpy(space_size, space->extent.size, ndims * sizeof(hsize_t));
} /* end if */
space_size[ndims] = elmt_size;
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Compute the maximum number of bytes required */
H5_CHECKED_ASSIGN(max_elem, size_t, nelmts, hssize_t);
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Loop, while elements left in selection */
while(max_elem > 0 && user_ret == 0) {
hsize_t off[H5D_IO_VECTOR_SIZE]; /* Array to store sequence offsets */
size_t len[H5D_IO_VECTOR_SIZE]; /* Array to store sequence lengths */
size_t nelem; /* Number of elements used in sequences */
size_t nseq; /* Number of sequences generated */
size_t curr_seq; /* Current sequence being worked on */
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Get the sequences of bytes */
if(H5S_SELECT_GET_SEQ_LIST(space, 0, &iter, (size_t)H5D_IO_VECTOR_SIZE, max_elem, &nseq, &nelem, off, len) < 0)
HGOTO_ERROR(H5E_INTERNAL, H5E_UNSUPPORTED, FAIL, "sequence length generation failed")
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Loop, while sequences left to process */
for(curr_seq = 0; curr_seq < nseq && user_ret == 0; curr_seq++) {
hsize_t curr_off; /* Current offset within sequence */
size_t curr_len; /* Length of bytes left to process in sequence */
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Get the current offset */
curr_off = off[curr_seq];
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Get the number of bytes in sequence */
curr_len = len[curr_seq];
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Loop, while bytes left in sequence */
while(curr_len > 0 && user_ret == 0) {
hsize_t coords[H5O_LAYOUT_NDIMS]; /* Coordinates of element in dataspace */
hsize_t tmp_off; /* Temporary offset within sequence */
uint8_t *loc; /* Current element location in buffer */
int i; /* Local Index variable */
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Compute the coordinate from the offset */
for(i = (int)ndims, tmp_off = curr_off; i >= 0; i--) {
coords[i] = tmp_off % space_size[i];
tmp_off /= space_size[i];
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
} /* end for */
/* Get the location within the user's buffer */
loc = (unsigned char *)buf + curr_off;
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Call user's callback routine */
user_ret = (*op)(loc, type_id, ndims, coords, operator_data);
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Increment offset in dataspace */
curr_off += elmt_size;
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Decrement number of bytes left in sequence */
curr_len -= elmt_size;
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
} /* end while */
} /* end for */
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Decrement number of elements left to process */
max_elem -= nelem;
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
} /* end while */
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Set return value */
ret_value = user_ret;
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
done:
/* Release selection iterator */
if(iter_init && H5S_SELECT_ITER_RELEASE(&iter) < 0)
HDONE_ERROR(H5E_DATASPACE, H5E_CANTRELEASE, FAIL, "unable to release selection iterator")
FUNC_LEAVE_NOAPI(ret_value)
} /* end H5S_select_iterate() */
/*--------------------------------------------------------------------------
NAME
H5Sget_select_type
PURPOSE
Retrieve the type of selection in a dataspace
USAGE
H5S_sel_type H5Sget_select_type(space_id)
hid_t space_id; IN: Dataspace object to query
RETURNS
Non-negative on success/Negative on failure. Return value is from the
set of values in the H5S_sel_type enumerated type.
DESCRIPTION
This function retrieves the type of selection currently defined for
a dataspace.
--------------------------------------------------------------------------*/
H5S_sel_type
H5Sget_select_type(hid_t space_id)
{
H5S_t *space; /* dataspace to modify */
H5S_sel_type ret_value; /* Return value */
FUNC_ENTER_API(H5S_SEL_ERROR)
H5TRACE1("St", "i", space_id);
/* Check args */
if(NULL == (space = (H5S_t *)H5I_object_verify(space_id, H5I_DATASPACE)))
HGOTO_ERROR(H5E_ATOM, H5E_BADATOM, H5S_SEL_ERROR, "not a dataspace")
/* Set return value */
ret_value = H5S_GET_SELECT_TYPE(space);
done:
FUNC_LEAVE_API(ret_value)
} /* end H5Sget_select_type() */
/*--------------------------------------------------------------------------
NAME
H5S_get_select_type
PURPOSE
Retrieve the type of selection in a dataspace
USAGE
H5S_sel_type H5Sget_select_type(space)
const H5S_t *space; IN: Dataspace object to query
RETURNS
Non-negative on success/Negative on failure. Return value is from the
set of values in the H5S_sel_type enumerated type.
DESCRIPTION
This function retrieves the type of selection currently defined for
a dataspace.
COMMENTS
This routine participates in the "Inlining C function pointers"
pattern, don't call it directly, use the appropriate macro
defined in H5Sprivate.h.
--------------------------------------------------------------------------*/
H5S_sel_type
H5S_get_select_type(const H5S_t *space)
{
H5S_sel_type ret_value; /* Return value */
FUNC_ENTER_NOAPI_NOINIT_NOERR
/* Check args */
HDassert(space);
/* Set return value */
ret_value=H5S_GET_SELECT_TYPE(space);
FUNC_LEAVE_NOAPI(ret_value)
} /* end H5S_get_select_type() */
/*--------------------------------------------------------------------------
NAME
H5S_select_shape_same
PURPOSE
Check if two selections are the same shape
USAGE
htri_t H5S_select_shape_same(space1, space2)
const H5S_t *space1; IN: 1st Dataspace pointer to compare
const H5S_t *space2; IN: 2nd Dataspace pointer to compare
RETURNS
TRUE/FALSE/FAIL
DESCRIPTION
Checks to see if the current selection in the dataspaces are the same
dimensionality and shape.
This is primarily used for reading the entire selection in one swoop.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
Assumes that there is only a single "block" for hyperslab selections.
EXAMPLES
REVISION LOG
Modified function to view identical shapes with different dimensions
as being the same under some circumstances.
--------------------------------------------------------------------------*/
htri_t
H5S_select_shape_same(const H5S_t *space1, const H5S_t *space2)
{
H5S_sel_iter_t iter_a; /* Selection a iteration info */
H5S_sel_iter_t iter_b; /* Selection b iteration info */
hbool_t iter_a_init = 0; /* Selection a iteration info has been initialized */
hbool_t iter_b_init = 0; /* Selection b iteration info has been initialized */
htri_t ret_value = TRUE; /* Return value */
FUNC_ENTER_NOAPI(FAIL)
/* Check args */
HDassert(space1);
HDassert(space2);
/* Special case for one or both dataspaces being scalar */
if(space1->extent.rank == 0 || space2->extent.rank == 0) {
/* Check for different number of elements selected */
if(H5S_GET_SELECT_NPOINTS(space1) != H5S_GET_SELECT_NPOINTS(space2))
HGOTO_DONE(FALSE)
} /* end if */
else {
const H5S_t *space_a; /* Dataspace with larger rank */
const H5S_t *space_b; /* Dataspace with smaller rank */
unsigned space_a_rank; /* Number of dimensions of dataspace A */
unsigned space_b_rank; /* Number of dimensions of dataspace B */
/* need to be able to handle spaces of different rank:
*
* To simplify logic, let space_a point to the element of the set
* {space1, space2} with the largest rank or space1 if the ranks
* are identical.
*
* Similarly, let space_b point to the element of {space1, space2}
* with the smallest rank, or space2 if they are identical.
*
* Let: space_a_rank be the rank of space_a,
* space_b_rank be the rank of space_b,
* delta_rank = space_a_rank - space_b_rank.
*
* Set all this up below.
*/
if(space1->extent.rank >= space2->extent.rank) {
space_a = space1;
space_a_rank = space_a->extent.rank;
space_b = space2;
space_b_rank = space_b->extent.rank;
} /* end if */
else {
space_a = space2;
space_a_rank = space_a->extent.rank;
space_b = space1;
space_b_rank = space_b->extent.rank;
} /* end else */
HDassert(space_a_rank >= space_b_rank);
HDassert(space_b_rank > 0);
/* Check for different number of elements selected */
if(H5S_GET_SELECT_NPOINTS(space_a) != H5S_GET_SELECT_NPOINTS(space_b))
HGOTO_DONE(FALSE)
/* Check for "easy" cases before getting into generalized block iteration code */
if((H5S_GET_SELECT_TYPE(space_a) == H5S_SEL_ALL) && (H5S_GET_SELECT_TYPE(space_b) == H5S_SEL_ALL)) {
hsize_t dims1[H5O_LAYOUT_NDIMS]; /* End point of selection block in dataspace #1 */
hsize_t dims2[H5O_LAYOUT_NDIMS]; /* End point of selection block in dataspace #2 */
int space_a_dim; /* Current dimension in dataspace A */
int space_b_dim; /* Current dimension in dataspace B */
if(H5S_get_simple_extent_dims(space_a, dims1, NULL) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTGET, FAIL, "unable to get dimensionality")
if(H5S_get_simple_extent_dims(space_b, dims2, NULL) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTGET, FAIL, "unable to get dimensionality")
space_a_dim = (int)space_a_rank - 1;
space_b_dim = (int)space_b_rank - 1;
/* recall that space_a_rank >= space_b_rank.
*
* In the following while loop, we test to see if space_a and space_b
* have identical size in all dimensions they have in common.
*/
while(space_b_dim >= 0) {
if(dims1[space_a_dim] != dims2[space_b_dim])
HGOTO_DONE(FALSE)
space_a_dim--;
space_b_dim--;
} /* end while */
/* Since we are selecting the entire spaces, we must also verify that space_a
* has size 1 in all dimensions that it does not share with space_b.
*/
while(space_a_dim >= 0) {
if(dims1[space_a_dim] != 1)
HGOTO_DONE(FALSE)
space_a_dim--;
} /* end while */
} /* end if */
else if((H5S_GET_SELECT_TYPE(space1) == H5S_SEL_NONE) || (H5S_GET_SELECT_TYPE(space2) == H5S_SEL_NONE)) {
HGOTO_DONE(TRUE)
} /* end if */
else if((H5S_GET_SELECT_TYPE(space_a) == H5S_SEL_HYPERSLABS && space_a->select.sel_info.hslab->diminfo_valid)
&& (H5S_GET_SELECT_TYPE(space_b) == H5S_SEL_HYPERSLABS && space_b->select.sel_info.hslab->diminfo_valid)) {
int space_a_dim; /* Current dimension in dataspace A */
int space_b_dim; /* Current dimension in dataspace B */
space_a_dim = (int)space_a_rank - 1;
space_b_dim = (int)space_b_rank - 1;
/* check that the shapes are the same in the common dimensions, and that
* block == 1 in all dimensions that appear only in space_a.
*/
while(space_b_dim >= 0) {
if(space_a->select.sel_info.hslab->opt_diminfo[space_a_dim].stride !=
space_b->select.sel_info.hslab->opt_diminfo[space_b_dim].stride)
HGOTO_DONE(FALSE)
if(space_a->select.sel_info.hslab->opt_diminfo[space_a_dim].count !=
space_b->select.sel_info.hslab->opt_diminfo[space_b_dim].count)
HGOTO_DONE(FALSE)
if(space_a->select.sel_info.hslab->opt_diminfo[space_a_dim].block !=
space_b->select.sel_info.hslab->opt_diminfo[space_b_dim].block)
HGOTO_DONE(FALSE)
space_a_dim--;
space_b_dim--;
} /* end while */
while(space_a_dim >= 0) {
if(space_a->select.sel_info.hslab->opt_diminfo[space_a_dim].block != 1)
HGOTO_DONE(FALSE)
space_a_dim--;
} /* end while */
} /* end if */
/* Iterate through all the blocks in the selection */
else {
hsize_t start_a[H5O_LAYOUT_NDIMS]; /* Start point of selection block in dataspace a */
hsize_t start_b[H5O_LAYOUT_NDIMS]; /* Start point of selection block in dataspace b */
hsize_t end_a[H5O_LAYOUT_NDIMS]; /* End point of selection block in dataspace a */
hsize_t end_b[H5O_LAYOUT_NDIMS]; /* End point of selection block in dataspace b */
hsize_t off_a[H5O_LAYOUT_NDIMS]; /* Offset of selection a blocks */
hsize_t off_b[H5O_LAYOUT_NDIMS]; /* Offset of selection b blocks */
hbool_t first_block = TRUE; /* Flag to indicate the first block */
/* Initialize iterator for each dataspace selection
* Use '0' for element size instead of actual element size to indicate
* that the selection iterator shouldn't be "flattened", since we
* aren't actually going to be doing I/O with the iterators.
*/
if(H5S_select_iter_init(&iter_a, space_a, (size_t)0) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTINIT, FAIL, "unable to initialize selection iterator a")
iter_a_init = 1;
if(H5S_select_iter_init(&iter_b, space_b, (size_t)0) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTINIT, FAIL, "unable to initialize selection iterator b")
iter_b_init = 1;
/* Iterate over all the blocks in each selection */
while(1) {
int space_a_dim; /* Current dimension in dataspace A */
int space_b_dim; /* Current dimension in dataspace B */
htri_t status_a, status_b; /* Status from next block checks */
/* Get the current block for each selection iterator */
if(H5S_SELECT_ITER_BLOCK(&iter_a, start_a, end_a) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTGET, FAIL, "unable to get iterator block a")
if(H5S_SELECT_ITER_BLOCK(&iter_b, start_b, end_b) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTGET, FAIL, "unable to get iterator block b")
space_a_dim = (int)space_a_rank - 1;
space_b_dim = (int)space_b_rank - 1;
/* The first block only compares the sizes and sets the
* relative offsets for later blocks
*/
if(first_block) {
/* If the block sizes in the common dimensions from
* each selection don't match, get out
*/
while(space_b_dim >= 0) {
if((end_a[space_a_dim] - start_a[space_a_dim]) !=
(end_b[space_b_dim] - start_b[space_b_dim]))
HGOTO_DONE(FALSE)
/* Set the relative locations of the selections */
off_a[space_a_dim] = start_a[space_a_dim];
off_b[space_b_dim] = start_b[space_b_dim];
space_a_dim--;
space_b_dim--;
} /* end while */
/* similarly, if the block size in any dimension that appears only
* in space_a is not equal to 1, get out.
*/
while(space_a_dim >= 0) {
if((end_a[space_a_dim] - start_a[space_a_dim]) != 0)
HGOTO_DONE(FALSE)
/* Set the relative locations of the selections */
off_a[space_a_dim] = start_a[space_a_dim];
space_a_dim--;
} /* end while */
/* Reset "first block" flag */
first_block = FALSE;
} /* end if */
/* Check over the blocks for each selection */
else {
/* for dimensions that space_a and space_b have in common: */
while(space_b_dim >= 0) {
/* Check if the blocks are in the same relative location */
if((start_a[space_a_dim] - off_a[space_a_dim]) !=
(start_b[space_b_dim] - off_b[space_b_dim]))
HGOTO_DONE(FALSE)
/* If the block sizes from each selection doesn't match, get out */
if((end_a[space_a_dim] - start_a[space_a_dim]) !=
(end_b[space_b_dim] - start_b[space_b_dim]))
HGOTO_DONE(FALSE)
space_a_dim--;
space_b_dim--;
} /* end while */
/* For dimensions that appear only in space_a: */
while(space_a_dim >= 0) {
/* If the block size isn't 1, get out */
if((end_a[space_a_dim] - start_a[space_a_dim]) != 0)
HGOTO_DONE(FALSE)
space_a_dim--;
} /* end while */
} /* end else */
/* Check if we are able to advance to the next selection block */
if((status_a = H5S_SELECT_ITER_HAS_NEXT_BLOCK(&iter_a)) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTNEXT, FAIL, "unable to check iterator block a")
if((status_b = H5S_SELECT_ITER_HAS_NEXT_BLOCK(&iter_b)) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTNEXT, FAIL, "unable to check iterator block b")
/* Did we run out of blocks at the same time? */
if((status_a == FALSE) && (status_b == FALSE))
break;
else if(status_a != status_b)
HGOTO_DONE(FALSE)
else {
/* Advance to next block in selection iterators */
if(H5S_SELECT_ITER_NEXT_BLOCK(&iter_a) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTNEXT, FAIL, "unable to advance to next iterator block a")
if(H5S_SELECT_ITER_NEXT_BLOCK(&iter_b) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTNEXT, FAIL, "unable to advance to next iterator block b")
} /* end else */
} /* end while */
} /* end else */
} /* end else */
done:
if(iter_a_init)
if(H5S_SELECT_ITER_RELEASE(&iter_a) < 0)
HDONE_ERROR(H5E_DATASPACE, H5E_CANTRELEASE, FAIL, "unable to release selection iterator a")
if(iter_b_init)
if(H5S_SELECT_ITER_RELEASE(&iter_b) < 0)
HDONE_ERROR(H5E_DATASPACE, H5E_CANTRELEASE, FAIL, "unable to release selection iterator b")
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_shape_same() */
/*--------------------------------------------------------------------------
NAME
H5S_select_construct_projection
PURPOSE
Given a dataspace a of rank n with some selection, construct a new
dataspace b of rank m (m != n), with the selection in a being
topologically identical to that in b (as verified by
H5S_select_shape_same().
This function exists, as some I/O code chokes of topologically
identical selections with different ranks. At least to begin
with, we will deal with the issue by constructing projections
of the memory dataspace with ranks equaling those of the file
dataspace.
Note that if m > n, it is possible that the starting point in the
buffer associated with the memory dataspace will have to be
adjusted to match the projected dataspace. If the buf parameter
is not NULL, the function must return an adjusted buffer base
address in *adj_buf_ptr.
USAGE
htri_t H5S_select_construct_projection(base_space,
new_space_ptr,
new_space_rank,
buf,
adj_buf_ptr)
const H5S_t *base_space; IN: Ptr to Dataspace to project
H5S_t ** new_space_ptr; OUT: Ptr to location in which to return
the address of the projected space
int new_space_rank; IN: Rank of the projected space.
const void * buf; IN: Base address of the buffer
associated with the base space.
May be NULL.
void ** adj_buf_ptr; OUT: If buf != NULL, store the base
address of the section of buf
that is described by *new_space_ptr
in *adj_buf_ptr.
RETURNS
Non-negative on success/Negative on failure.
DESCRIPTION
Construct a new dataspace and associated selection which is a
projection of the supplied dataspace and associated selection into
the specified rank. Return it in *new_space_ptr.
If buf is supplied, computes the base address of the projected
selection in buf, and stores the base address in *adj_buf_ptr.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
The selection in the supplied base_space has thickness 1 in all
dimensions greater than new_space_rank. Note that here we count
dimensions from the fastest changing coordinate to the slowest
changing changing coordinate.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
herr_t
H5S_select_construct_projection(const H5S_t *base_space, H5S_t **new_space_ptr,
unsigned new_space_rank, const void *buf, void const **adj_buf_ptr, hsize_t element_size)
{
H5S_t * new_space = NULL; /* New dataspace constructed */
hsize_t base_space_dims[H5S_MAX_RANK]; /* Current dimensions of base dataspace */
hsize_t base_space_maxdims[H5S_MAX_RANK]; /* Maximum dimensions of base dataspace */
int sbase_space_rank; /* Signed # of dimensions of base dataspace */
unsigned base_space_rank; /* # of dimensions of base dataspace */
hsize_t projected_space_element_offset = 0; /* Offset of selected element in projected buffer */
herr_t ret_value = SUCCEED; /* Return value */
FUNC_ENTER_NOAPI(FAIL)
/* Sanity checks */
HDassert(base_space != NULL);
HDassert((H5S_GET_EXTENT_TYPE(base_space) == H5S_SCALAR) || (H5S_GET_EXTENT_TYPE(base_space) == H5S_SIMPLE));
HDassert(new_space_ptr != NULL);
HDassert((new_space_rank != 0) || (H5S_GET_SELECT_NPOINTS(base_space) <= 1));
HDassert(new_space_rank <= H5S_MAX_RANK);
HDassert((buf == NULL) || (adj_buf_ptr != NULL));
HDassert(element_size > 0 );
/* Get the extent info for the base dataspace */
if((sbase_space_rank = H5S_get_simple_extent_dims(base_space, base_space_dims, base_space_maxdims)) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTGET, FAIL, "unable to get dimensionality of base space")
base_space_rank = (unsigned)sbase_space_rank;
HDassert(base_space_rank != new_space_rank);
/* Check if projected space is scalar */
if(new_space_rank == 0) {
hssize_t npoints; /* Number of points selected */
/* Retreve the number of elements selected */
if((npoints = (hssize_t)H5S_GET_SELECT_NPOINTS(base_space)) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTGET, FAIL, "unable to get number of points selected")
HDassert(npoints <= 1);
/* Create new scalar dataspace */
if(NULL == (new_space = H5S_create(H5S_SCALAR)))
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTCREATE, FAIL, "unable to create scalar dataspace")
/* No need to register the dataspace(i.e. get an ID) as
* we will just be discarding it shortly.
*/
/* Selection for the new space will be either all or
* none, depending on whether the base space has 0 or
* 1 elements selected.
*
* Observe that the base space can't have more than
* one selected element, since its selection has the
* same shape as the file dataspace, and that data
* space is scalar.
*/
if(1 == npoints) {
/* Assuming that the selection in the base dataspace is not
* empty, we must compute the offset of the selected item in
* the buffer associated with the base dataspace.
*
* Since the new space rank is zero, we know that the
* the base space must have rank at least 1 -- and
* hence it is a simple dataspace. However, the
* selection, may be either point, hyperspace, or all.
*
*/
if(H5S_SELECT_PROJECT_SCALAR(base_space, &projected_space_element_offset) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTSET, FAIL, "unable to project scalar selection")
} /* end if */
else {
HDassert(0 == npoints);
if(H5S_select_none(new_space) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTDELETE, FAIL, "can't delete default selection")
} /* end else */
} /* end if */
else { /* projected space must be simple */
hsize_t new_space_dims[H5S_MAX_RANK]; /* Current dimensions for new dataspace */
hsize_t new_space_maxdims[H5S_MAX_RANK];/* Maximum dimensions for new dataspace */
unsigned rank_diff; /* Difference in ranks */
/* Set up the dimensions of the new, projected dataspace.
*
* How we do this depends on whether we are projecting up into
* increased dimensions, or down into a reduced number of
* dimensions.
*
* If we are projecting up (the first half of the following
* if statement), we copy the dimensions of the base data
* space into the fastest changing dimensions of the new
* projected dataspace, and set the remaining dimensions to
* one.
*
* If we are projecting down (the second half of the following
* if statement), we just copy the dimensions with the most
* quickly changing dimensions into the dims for the projected
* data set.
*
* This works, because H5S_select_shape_same() will return
* true on selections of different rank iff:
*
* 1) the selection in the lower rank dataspace matches that
* in the dimensions with the fastest changing indicies in
* the larger rank dataspace, and
*
* 2) the selection has thickness 1 in all ranks that appear
* only in the higher rank dataspace (i.e. those with
* more slowly changing indicies).
*/
if(new_space_rank > base_space_rank) {
hsize_t tmp_dim_size = 1; /* Temporary dimension value, for filling arrays */
/* we must copy the dimensions of the base space into
* the fastest changing dimensions of the new space,
* and set the remaining dimensions to 1
*/
rank_diff = new_space_rank - base_space_rank;
H5VM_array_fill(new_space_dims, &tmp_dim_size, sizeof(tmp_dim_size), rank_diff);
H5VM_array_fill(new_space_maxdims, &tmp_dim_size, sizeof(tmp_dim_size), rank_diff);
HDmemcpy(&new_space_dims[rank_diff], base_space_dims, sizeof(new_space_dims[0]) * base_space_rank);
HDmemcpy(&new_space_maxdims[rank_diff], base_space_maxdims, sizeof(new_space_maxdims[0]) * base_space_rank);
} /* end if */
else { /* new_space_rank < base_space_rank */
/* we must copy the fastest changing dimension of the
* base space into the dimensions of the new space.
*/
rank_diff = base_space_rank - new_space_rank;
HDmemcpy(new_space_dims, &base_space_dims[rank_diff], sizeof(new_space_dims[0]) * new_space_rank);
HDmemcpy(new_space_maxdims, &base_space_maxdims[rank_diff], sizeof(new_space_maxdims[0]) * new_space_rank);
} /* end else */
/* now have the new space rank and dimensions set up --
* so we can create the new simple dataspace.
*/
if(NULL == (new_space = H5S_create_simple(new_space_rank, new_space_dims, new_space_maxdims)))
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTCREATE, FAIL, "can't create simple dataspace")
/* No need to register the dataspace(i.e. get an ID) as
* we will just be discarding it shortly.
*/
/* If we get this far, we have successfully created the projected
* dataspace. We must now project the selection in the base
* dataspace into the projected dataspace.
*/
if(H5S_SELECT_PROJECT_SIMPLE(base_space, new_space, &projected_space_element_offset) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTSET, FAIL, "unable to project simple selection")
/* If we get this far, we have created the new dataspace, and projected
* the selection in the base dataspace into the new dataspace.
*
* If the base dataspace is simple, check to see if the
* offset_changed flag on the base selection has been set -- if so,
* project the offset into the new dataspace and set the
* offset_changed flag.
*/
if(H5S_GET_EXTENT_TYPE(base_space) == H5S_SIMPLE && base_space->select.offset_changed) {
if(new_space_rank > base_space_rank) {
HDmemset(new_space->select.offset, 0, sizeof(new_space->select.offset[0]) * rank_diff);
HDmemcpy(&new_space->select.offset[rank_diff], base_space->select.offset, sizeof(new_space->select.offset[0]) * base_space_rank);
} /* end if */
else
HDmemcpy(new_space->select.offset, &base_space->select.offset[rank_diff], sizeof(new_space->select.offset[0]) * new_space_rank);
/* Propagate the offset changed flag into the new dataspace. */
new_space->select.offset_changed = TRUE;
} /* end if */
} /* end else */
/* If we have done the projection correctly, the following assertion
* should hold.
*/
HDassert(TRUE == H5S_select_shape_same(base_space, new_space));
/* load the address of the new space into *new_space_ptr */
*new_space_ptr = new_space;
/* now adjust the buffer if required */
if(buf != NULL) {
if(new_space_rank < base_space_rank) {
/* a bit of pointer magic here:
*
* Since we can't do pointer arithmetic on void pointers, we first
* cast buf to a pointer to byte -- i.e. uint8_t.
*
* We then multiply the projected space element offset we
* calculated earlier by the supplied element size, add this
* value to the type cast buf pointer, cast the result back
* to a pointer to void, and assign the result to *adj_buf_ptr.
*/
*adj_buf_ptr = (const void *)(((const uint8_t *)buf) +
((size_t)(projected_space_element_offset * element_size)));
} /* end if */
else
/* No adjustment necessary */
*adj_buf_ptr = buf;
} /* end if */
done:
/* Cleanup on error */
if(ret_value < 0) {
if(new_space && H5S_close(new_space) < 0)
HDONE_ERROR(H5E_DATASPACE, H5E_CANTRELEASE, FAIL, "unable to release dataspace")
} /* end if */
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_construct_projection() */
/*--------------------------------------------------------------------------
NAME
H5S_select_fill
PURPOSE
Fill a selection in memory with a value
USAGE
herr_t H5S_select_fill(fill,fill_size,space,buf)
const void *fill; IN: Pointer to fill value to use
size_t fill_size; IN: Size of elements in memory buffer & size of
fill value
H5S_t *space; IN: Dataspace describing memory buffer &
containing selection to use.
void *buf; IN/OUT: Memory buffer to fill selection in
RETURNS
Non-negative on success/Negative on failure.
DESCRIPTION
Use the selection in the dataspace to fill elements in a memory buffer.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
The memory buffer elements are assumed to have the same datatype as the
fill value being placed into them.
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
herr_t
H5S_select_fill(const void *fill, size_t fill_size, const H5S_t *space, void *_buf)
{
[svn-r6252] Purpose: Lots of performance improvements & a couple new internal API interfaces. Description: Performance Improvements: - Cached file offset & length sizes in shared file struct, to avoid constantly looking them up in the FCPL. - Generic property improvements: - Added "revision" number to generic property classes to speed up comparisons. - Changed method of storing properties from using a hash-table to the TBBT routines in the library. - Share the propery names between classes and the lists derived from them. - Removed redundant 'def_value' buffer from each property. - Switching code to use a "copy on write" strategy for properties in each list, where the properties in each list are shared with the properties in the class, until a property's value is changed in a list. - Fixed error in layout code which was allocating too many buffers. - Redefined public macros of the form (H5open()/H5check, <variable>) internally to only be (<variable>), avoiding innumerable useless calls to H5open() and H5check_version(). - Reuse already zeroed buffers in H5F_contig_fill instead of constantly re-zeroing them. - Don't write fill values if writing entire dataset. - Use gettimeofday() system call instead of time() system when checking the modification time of a dataset. - Added reference counted string API and use it for tracking the names of objects opening in a file (for the ID->name code). - Removed redundant H5P_get() calls in B-tree routines. - Redefine H5T datatype macros internally to the library, to avoid calling H5check redundantly. - Keep dataspace information for dataset locally instead of reading from disk each time. Added new module to track open objects in a file, to allow this (which will be useful eventually for some FPH5 metadata caching issues). - Remove H5AC_find macro which was inlining metadata cache lookups, and call function instead. - Remove redundant memset() calls from H5G_namei() routine. - Remove redundant checking of object type when locating objects in metadata cache and rely on the address only. - Create default dataset object to use when default dataset creation property list is used to create datasets, bypassing querying for all the property list values. - Use default I/O vector size when performing raw data with the default dataset transfer property list, instead of querying for I/O vector size. - Remove H5P_DEFAULT internally to the library, replacing it with more specific default property list based on the type of property list needed. - Remove redundant memset() calls in object header message (H5O*) routines. - Remove redunant memset() calls in data I/O routines. - Split free-list allocation routines into malloc() and calloc()- like routines, instead of one combined routine. - Remove lots of indirection in H5O*() routines. - Simplify metadata cache entry comparison routine (used when flushing entire cache out). - Only enable metadata cache statistics when H5AC_DEBUG is turned on, instead of always tracking them. - Simplify address comparison macro (H5F_addr_eq). - Remove redundant metadata cache entry protections during dataset creation by protecting the object header once and making all the modifications necessary for the dataset creation before unprotecting it. - Reduce # of "number of element in extent" computations performed by computing and storing the value during dataspace creation. - Simplify checking for group location's file information, when file has not been involving in file-mounting operations. - Use binary encoding for modification time, instead of ASCII. - Hoist H5HL_peek calls (to get information in a local heap) out of loops in many group routine. - Use static variable for iterators of selections, instead of dynamically allocation them each time. - Lookup & insert new entries in one step, avoiding traversing group's B-tree twice. - Fixed memory leak in H5Gget_objname_idx() routine (tangential to performance improvements, but fixed along the way). - Use free-list for reference counted strings. - Don't bother copying object names into cached group entries, since they are re-created when an object is opened. The benchmark I used to measure these results created several thousand small (2K) datasets in a file and wrote out the data for them. This is Elena's "regular.c" benchmark. These changes resulted in approximately ~4.3x speedup of the development branch when compared to the previous code in the development branch and ~1.4x speedup compared to the release branch. Additionally, these changes reduce the total memory used (code and data) by the development branch by ~800KB, bringing the development branch back into the same ballpark as the release branch. I'll send out a more detailed description of the benchmark results as a followup note. New internal API routines: Added "reference counted strings" API for tracking strings that get used by multiple owners without duplicating the strings. Added "ternary search tree" API for text->object mappings. Platforms tested: Tested h5committest {arabica (fortran), eirene (fortran, C++) modi4 (parallel, fortran)} Other platforms/configurations tested? FreeBSD 4.7 (sleipnir) serial & parallel Solaris 2.6 (baldric) serial
2003-01-09 12:20:03 -05:00
H5S_sel_iter_t iter; /* Selection iteration info */
hbool_t iter_init = 0; /* Selection iteration info has been initialized */
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
hssize_t nelmts; /* Number of elements in selection */
size_t max_elem; /* Total number of elements in selection */
herr_t ret_value = SUCCEED; /* Return value */
FUNC_ENTER_NOAPI(FAIL)
/* Check args */
HDassert(fill);
HDassert(fill_size > 0);
HDassert(space);
HDassert(_buf);
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Initialize iterator */
if(H5S_select_iter_init(&iter, space, fill_size) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTINIT, FAIL, "unable to initialize selection iterator")
iter_init = 1; /* Selection iteration info has been initialized */
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Get the number of elements in selection */
if((nelmts = (hssize_t)H5S_GET_SELECT_NPOINTS(space)) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTCOUNT, FAIL, "can't get number of elements selected")
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Compute the number of bytes to process */
H5_CHECKED_ASSIGN(max_elem, size_t, nelmts, hssize_t);
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Loop, while elements left in selection */
while(max_elem > 0) {
hsize_t off[H5D_IO_VECTOR_SIZE]; /* Array to store sequence offsets */
size_t len[H5D_IO_VECTOR_SIZE]; /* Array to store sequence lengths */
size_t nseq; /* Number of sequences generated */
size_t curr_seq; /* Current sequnce being worked on */
size_t nelem; /* Number of elements used in sequences */
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Get the sequences of bytes */
if(H5S_SELECT_GET_SEQ_LIST(space, 0, &iter, (size_t)H5D_IO_VECTOR_SIZE, max_elem, &nseq, &nelem, off, len) < 0)
HGOTO_ERROR(H5E_INTERNAL, H5E_UNSUPPORTED, FAIL, "sequence length generation failed")
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Loop over sequences */
for(curr_seq = 0; curr_seq < nseq; curr_seq++) {
uint8_t *buf; /* Current location in buffer */
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Get offset in memory buffer */
buf = (uint8_t *)_buf + off[curr_seq];
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
/* Fill each sequence in memory with fill value */
HDassert((len[curr_seq] % fill_size) == 0);
H5VM_array_fill(buf, fill, fill_size, (len[curr_seq] / fill_size));
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
} /* end for */
/* Decrement number of elements left to process */
max_elem -= nelem;
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
} /* end while */
done:
/* Release resouces */
if(iter_init && H5S_SELECT_ITER_RELEASE(&iter) < 0)
HDONE_ERROR(H5E_DATASPACE, H5E_CANTRELEASE, FAIL, "unable to release selection iterator")
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
FUNC_LEAVE_NOAPI(ret_value)
[svn-r5834] Purpose: Large code cleanup/re-write Description: This is phase 1 of the data I/O re-architecture, with the following changes: - Changed the selection drivers to not actually do any I/O, they only generate the sequences of offset/length pairs needed for the I/O (or memory access, in the case of iterating or filling a selection in a memory buffer) - Wrote more abstract I/O routines which get the sequence of offset/ length pairs for each selection and access perform the I/O or memory access. Benefits of this change include: - Removed ~3400 lines of quite redundant code, with corresponding reduction in the size of library binary. - Any selection can now directly access memory when performing I/O, if no type conversions are required, instead of just "regular" hyperslab and 'all' selections, which speeds up I/O. - Sped up I/O for hyperslab selections which have contiguous lower dimensions by "flattening" them out into lesser dimensional objects for the I/O. No file format or API changes were necessary for this change. The next phase will be to create a "selection driver" for each type of selection, allowing each type of selection to directly call certain methods that only apply to that type of selection, instead of passing through dozens of functions which have switch statements to call the appropriate method for each selection type. This will also reduce the amount of code in the library and speed things up a bit more. Phase 3 will involve generating an MPI datatype for all types of selections, instead of only "regular" hyperslab and 'all' selections. This will allow collective parallel I/O for all I/O operations which don't require type conversions. It will also open up the door for allowing collective I/O on datasets which require type conversion. Phase 4 will involve changing the access pattern to deal with chunked datasets in a more optimal way (in serial). Phase 5 will deal with accessing chunked datasets more optimally for collective parallel I/O operations. Platforms tested: FreeBSD 4.6 (sleipnir) w/ parallel & C++ and IRIX64 6.5 (modi4) w/parallel
2002-07-24 13:56:48 -05:00
} /* H5S_select_fill() */
/*--------------------------------------------------------------------------
NAME
H5S_select_project_intersection
PURPOSE
Projects the intersection of of the selections of src_space and
src_intersect_space within the selection of src_space as a selection
within the selection of dst_space
USAGE
herr_t H5S_select_project_intersection(src_space,dst_space,src_intersect_space,proj_space)
H5S_t *src_space; IN: Selection that is mapped to dst_space, and intersected with src_intersect_space
H5S_t *dst_space; IN: Selection that is mapped to src_space, and which contains the result
H5S_t *src_intersect_space; IN: Selection whose intersection with src_space is projected to dst_space to obtain the result
H5S_t *proj_space; OUT: Will contain the result (intersection of src_intersect_space and src_space projected from src_space to dst_space) after the operation
RETURNS
Non-negative on success/Negative on failure.
DESCRIPTION
Projects the intersection of of the selections of src_space and
src_intersect_space within the selection of src_space as a selection
within the selection of dst_space. The result is placed in the
selection of proj_space.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
herr_t
H5S_select_project_intersection(const H5S_t *src_space, const H5S_t *dst_space,
const H5S_t *src_intersect_space, H5S_t **new_space_ptr)
{
H5S_t *new_space = NULL; /* New dataspace constructed */
herr_t ret_value = SUCCEED; /* Return value */
FUNC_ENTER_NOAPI(FAIL)
/* Sanity checks */
HDassert(src_space);
HDassert(dst_space);
HDassert(src_intersect_space);
HDassert(new_space_ptr);
/* Create new space, using dst extent. Start with "all" selection. */
if(NULL == (new_space = H5S_create(H5S_SIMPLE)))
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTCREATE, FAIL, "unable to create output dataspace")
if(H5S_extent_copy_real(&new_space->extent, &dst_space->extent, TRUE) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTCOPY, FAIL, "unable to copy destination space extent")
/* If the intersecting space is "all", the intersection must be equal to the
* source space and the projection must be equal to the destination space */
if(src_intersect_space->select.type->type == H5S_SEL_ALL) {
/* Copy the destination selection. */
if(H5S_select_copy(new_space, dst_space, FALSE) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTCOPY, FAIL, "can't copy destination space selection")
} /* end if */
/* If any of the spaces are "none", the projection must also be "none" */
else if((src_intersect_space->select.type->type == H5S_SEL_NONE)
|| (src_space->select.type->type == H5S_SEL_NONE)
|| (dst_space->select.type->type == H5S_SEL_NONE)) {
/* Change to "none" selection */
if(H5S_select_none(new_space) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTDELETE, FAIL, "can't change selection")
} /* end if */
/* If any of the spaces use point selection, fall back to general algorithm
*/
else if((src_intersect_space->select.type->type == H5S_SEL_POINTS)
|| (src_space->select.type->type == H5S_SEL_POINTS)
|| (dst_space->select.type->type == H5S_SEL_POINTS))
HGOTO_ERROR(H5E_DATASPACE, H5E_UNSUPPORTED, FAIL, "point selections not currently supported")
else {
HDassert(src_intersect_space->select.type->type == H5S_SEL_HYPERSLABS);
/* Intersecting space is hyperslab selection. Call the hyperslab
* routine to project to another hyperslab selection. */
if(H5S__hyper_project_intersection(src_space, dst_space, src_intersect_space, new_space) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTCLIP, FAIL, "can't project hyperslab ondot destination selection")
} /* end else */
/* load the address of the new space into *new_space_ptr */
*new_space_ptr = new_space;
done:
/* Cleanup on error */
if(ret_value < 0) {
if(new_space && H5S_close(new_space) < 0)
HDONE_ERROR(H5E_DATASPACE, H5E_CANTRELEASE, FAIL, "unable to release dataspace")
} /* end if */
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_project_intersection() */
/*--------------------------------------------------------------------------
NAME
H5S_select_subtract
PURPOSE
Subtract one selection from another
USAGE
herr_t H5S_select_subtract(space,subtract_space)
H5S_t *space; IN/OUT: Selection to be operated on
H5S_t *subtract_space; IN: Selection that will be subtracted from space
RETURNS
Non-negative on success/Negative on failure.
DESCRIPTION
Removes any and all portions of space that are also present in
subtract_space. In essence, performs an A_NOT_B operation with the
two selections.
GLOBAL VARIABLES
COMMENTS, BUGS, ASSUMPTIONS
EXAMPLES
REVISION LOG
--------------------------------------------------------------------------*/
herr_t
H5S_select_subtract(H5S_t *space, H5S_t *subtract_space)
{
herr_t ret_value = SUCCEED; /* Return value */
FUNC_ENTER_NOAPI(FAIL)
/* Sanity checks */
HDassert(space);
HDassert(subtract_space);
/* If either space is using the none selection, then we do not need to do
* anything */
if((space->select.type->type != H5S_SEL_NONE)
&& (subtract_space->select.type->type != H5S_SEL_NONE)) {
/* If subtract_space is using the all selection, set space to none */
if(subtract_space->select.type->type == H5S_SEL_ALL) {
/* Change to "none" selection */
if(H5S_select_none(space) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTDELETE, FAIL, "can't change selection")
} /* end if */
else {
/* Check for point selection in subtract_space, convert to
* hyperslab */
if(subtract_space->select.type->type == H5S_SEL_POINTS)
HGOTO_ERROR(H5E_DATASPACE, H5E_UNSUPPORTED, FAIL, "point selections not currently supported")
/* Check for point or all selection in space, convert to hyperslab
*/
if(space->select.type->type == H5S_SEL_ALL) {
/* Convert current "all" selection to "real" hyperslab selection */
/* Then allow operation to proceed */
hsize_t tmp_start[H5O_LAYOUT_NDIMS]; /* Temporary start information */
hsize_t tmp_stride[H5O_LAYOUT_NDIMS]; /* Temporary stride information */
hsize_t tmp_count[H5O_LAYOUT_NDIMS]; /* Temporary count information */
hsize_t tmp_block[H5O_LAYOUT_NDIMS]; /* Temporary block information */
unsigned i; /* Local index variable */
/* Fill in temporary information for the dimensions */
for(i = 0; i < space->extent.rank; i++) {
tmp_start[i] = 0;
tmp_stride[i] = 1;
tmp_count[i] = 1;
tmp_block[i] = space->extent.size[i];
} /* end for */
/* Convert to hyperslab selection */
if(H5S_select_hyperslab(space, H5S_SELECT_SET, tmp_start, tmp_stride, tmp_count, tmp_block) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTSELECT, FAIL, "can't convert selection")
} /* end if */
else if(space->select.type->type == H5S_SEL_POINTS)
HGOTO_ERROR(H5E_DATASPACE, H5E_UNSUPPORTED, FAIL, "point selections not currently supported")
HDassert(space->select.type->type == H5S_SEL_HYPERSLABS);
HDassert(subtract_space->select.type->type == H5S_SEL_HYPERSLABS);
/* Both spaces are now hyperslabs, perform the operation */
if(H5S__hyper_subtract(space, subtract_space) < 0)
HGOTO_ERROR(H5E_DATASPACE, H5E_CANTCLIP, FAIL, "can't subtract hyperslab")
} /* end else */
} /* end if */
done:
FUNC_LEAVE_NOAPI(ret_value)
} /* H5S_select_subtract() */