[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

1 files changed, 201 insertions, 7 deletions

diff --git a/src/H5Omtime.c b/src/H5Omtime.c
index e31dece..ff3eb5a 100644
--- a/src/H5Omtime.c
+++ b/src/H5Omtime.c
@@ -20,11 +20,16 @@
 
 #define PABLO_MASK	H5O_mtime_mask
 
+static void *H5O_mtime_new_decode(H5F_t *f, const uint8_t *p, H5O_shared_t *sh);
+static herr_t H5O_mtime_new_encode(H5F_t *f, uint8_t *p, const void *_mesg);
+static size_t H5O_mtime_new_size(H5F_t *f, const void *_mesg);
+
 static void *H5O_mtime_decode(H5F_t *f, const uint8_t *p, H5O_shared_t *sh);
 static herr_t H5O_mtime_encode(H5F_t *f, uint8_t *p, const void *_mesg);
 static void *H5O_mtime_copy(const void *_mesg, void *_dest);
 static size_t H5O_mtime_size(H5F_t *f, const void *_mesg);
-static herr_t H5O_mtime_free (void *_mesg);
+static herr_t H5O_mtime_reset(void *_mesg);
+static herr_t H5O_mtime_free(void *_mesg);
 static herr_t H5O_mtime_debug(H5F_t *f, const void *_mesg, FILE *stream,
 			     int indent, int fwidth);
 
@@ -37,8 +42,25 @@ const H5O_class_t H5O_MTIME[1] = {{
     H5O_mtime_encode,		/*encode message		*/
     H5O_mtime_copy,		/*copy the native value		*/
     H5O_mtime_size,		/*raw message size		*/
-    NULL,			/*free internal memory		*/
-    H5O_mtime_free,		            /* free method			*/
+    H5O_mtime_reset,		/* reset method			*/
+    H5O_mtime_free,		/* free method			*/
+    NULL,			/*get share method		*/
+    NULL,			/*set share method		*/
+    H5O_mtime_debug,		/*debug the message		*/
+}};
+
+/* This message derives from H5O */
+/* (Only encode, decode & size routines are different from old mtime routines) */
+const H5O_class_t H5O_MTIME_NEW[1] = {{
+    H5O_MTIME_NEW_ID,		/*message id number		*/
+    "mtime_new",		/*message name for debugging	*/
+    sizeof(time_t),		/*native message size		*/
+    H5O_mtime_new_decode,	/*decode message		*/
+    H5O_mtime_new_encode,	/*encode message		*/
+    H5O_mtime_copy,		/*copy the native value		*/
+    H5O_mtime_new_size,		/*raw message size		*/
+    H5O_mtime_reset,		/* reset method			*/
+    H5O_mtime_free,		/* free method			*/
     NULL,			/*get share method		*/
     NULL,			/*set share method		*/
     H5O_mtime_debug,		/*debug the message		*/
@@ -48,11 +70,74 @@ const H5O_class_t H5O_MTIME[1] = {{
 static int interface_initialize_g = 0;
 #define INTERFACE_INIT	NULL
 
+/* Current version of new mtime information */
+#define H5O_MTIME_VERSION 	1	
+ 
+/* Track whether tzset routine was called */
+static int	ntzset=0;
+
 /* Declare a free list to manage the time_t struct */
 H5FL_DEFINE(time_t);
 
 
 /*-------------------------------------------------------------------------
+ * Function:	H5O_mtime_new_decode
+ *
+ * Purpose:	Decode a new modification time message and return a pointer to a
+ *		new time_t value.
+ *
+ * Return:	Success:	Ptr to new message in native struct.
+ *
+ *		Failure:	NULL
+ *
+ * Programmer:	Quincey Koziol
+ *		koziol@ncsa.uiuc.edu
+ *		Jan  3 2002
+ *
+ * Modifications:
+ *
+ *-------------------------------------------------------------------------
+ */
+static void *
+H5O_mtime_new_decode(H5F_t UNUSED *f, const uint8_t *p,
+		 H5O_shared_t UNUSED *sh)
+{
+    time_t	*mesg, the_time;
+    int		version;        /* Version of mtime information */
+    void        *ret_value;     /* Return value */
+
+    FUNC_ENTER_NOAPI(H5O_mtime_new_decode, NULL);
+
+    /* check args */
+    assert(f);
+    assert(p);
+    assert (!sh);
+
+    /* decode */
+    version = *p++;
+    if(version!=H5O_MTIME_VERSION)
+        HGOTO_ERROR(H5E_OHDR, H5E_CANTLOAD, NULL, "bad version number for mtime message");
+
+    /* Skip reserved bytes */
+    p+=3;
+
+    /* Get the time_t from the file */
+    UINT32DECODE(p, the_time);
+
+    /* The return value */
+    if (NULL==(mesg = H5FL_MALLOC(time_t)))
+	HGOTO_ERROR(H5E_RESOURCE, H5E_NOSPACE, NULL, "memory allocation failed");
+    *mesg = the_time;
+
+    /* Set return value */
+    ret_value=mesg;
+
+done:
+    FUNC_LEAVE(ret_value);
+} /* end H5O_mtime_new_decode() */
+
+
+/*-------------------------------------------------------------------------
  * Function:	H5O_mtime_decode
  *
  * Purpose:	Decode a modification time message and return a pointer to a
@@ -77,7 +162,6 @@ H5O_mtime_decode(H5F_t UNUSED *f, const uint8_t *p,
     time_t	*mesg, the_time;
     int	i;
     struct tm	tm;
-    static int	ncalls=0;
     void        *ret_value;     /* Return value */
 
     FUNC_ENTER_NOAPI(H5O_mtime_decode, NULL);
@@ -88,7 +172,10 @@ H5O_mtime_decode(H5F_t UNUSED *f, const uint8_t *p,
     assert (!sh);
 
     /* Initialize time zone information */
-    if (0==ncalls++) HDtzset();
+    if (!ntzset) {
+        HDtzset();
+        ntzset=1;
+    } /* end if */
 
     /* decode */
     for (i=0; i<14; i++) {
@@ -166,7 +253,7 @@ H5O_mtime_decode(H5F_t UNUSED *f, const uint8_t *p,
 #endif
     
     /* The return value */
-    if (NULL==(mesg = H5FL_ALLOC(time_t,1)))
+    if (NULL==(mesg = H5FL_MALLOC(time_t)))
 	HGOTO_ERROR (H5E_RESOURCE, H5E_NOSPACE, NULL, "memory allocation failed");
     *mesg = the_time;
 
@@ -179,6 +266,50 @@ done:
 
 
 /*-------------------------------------------------------------------------
+ * Function:	H5O_mtime_new_encode
+ *
+ * Purpose:	Encodes a new modification time message.
+ *
+ * Return:	Non-negative on success/Negative on failure
+ *
+ * Programmer:	Quincey Koziol
+ *		koziol@ncsa.uiuc.edu
+ *		Jan  3 2002
+ *
+ * Modifications:
+ *
+ *-------------------------------------------------------------------------
+ */
+static herr_t
+H5O_mtime_new_encode(H5F_t UNUSED *f, uint8_t *p, const void *_mesg)
+{
+    const time_t	*mesg = (const time_t *) _mesg;
+    herr_t ret_value=SUCCEED;   /* Return value */
+    
+    FUNC_ENTER_NOAPI(H5O_mtime_new_encode, FAIL);
+
+    /* check args */
+    assert(f);
+    assert(p);
+    assert(mesg);
+
+    /* Version */
+    *p++ = H5O_MTIME_VERSION;
+
+    /* Reserved bytes */
+    *p++ = 0;
+    *p++ = 0;
+    *p++ = 0;
+
+    /* Encode time */
+    UINT32ENCODE(p, *mesg);
+
+done:
+    FUNC_LEAVE(ret_value);
+} /* end H5O_mtime_new_encode() */
+
+
+/*-------------------------------------------------------------------------
  * Function:	H5O_mtime_encode
  *
  * Purpose:	Encodes a modification time message.
@@ -247,7 +378,7 @@ H5O_mtime_copy(const void *_mesg, void *_dest)
 
     /* check args */
     assert(mesg);
-    if (!dest && NULL==(dest = H5FL_ALLOC(time_t,0)))
+    if (!dest && NULL==(dest = H5FL_MALLOC(time_t)))
         HGOTO_ERROR (H5E_RESOURCE, H5E_NOSPACE, NULL, "memory allocation failed");
     
     /* copy */
@@ -262,6 +393,42 @@ done:
 
 
 /*-------------------------------------------------------------------------
+ * Function:	H5O_mtime_new_size
+ *
+ * Purpose:	Returns the size of the raw message in bytes not
+ *		counting the message type or size fields, but only the data
+ *		fields.	 This function doesn't take into account
+ *		alignment.
+ *
+ * Return:	Success:	Message data size in bytes w/o alignment.
+ *
+ *		Failure:	0
+ *
+ * Programmer:	Quincey Koziol
+ *		koziol@ncsa.uiuc.edu
+ *		Jan  3 2002
+ *
+ * Modifications:
+ *
+ *-------------------------------------------------------------------------
+ */
+static size_t
+H5O_mtime_new_size(H5F_t * UNUSED f, const void * UNUSED mesg)
+{
+    size_t ret_value=8;        /* Return value */
+
+    FUNC_ENTER_NOAPI(H5O_mtime_new_size, 0);
+
+    /* check args */
+    assert(f);
+    assert(mesg);
+
+done:
+    FUNC_LEAVE(ret_value);
+} /* end H5O_mtime_new_size() */
+
+
+/*-------------------------------------------------------------------------
  * Function:	H5O_mtime_size
  *
  * Purpose:	Returns the size of the raw message in bytes not
@@ -298,6 +465,33 @@ done:
 
 
 /*-------------------------------------------------------------------------
+ * Function:	H5O_mtime_reset
+ *
+ * Purpose:	Frees resources within a modification time message, but doesn't free
+ *		the message itself.
+ *
+ * Return:	Non-negative on success/Negative on failure
+ *
+ * Programmer:	Quincey Koziol
+ *		Mondey, December 23, 2002
+ *
+ * Modifications:
+ *
+ *-------------------------------------------------------------------------
+ */
+static herr_t
+H5O_mtime_reset(void UNUSED *_mesg)
+{
+    herr_t      ret_value=SUCCEED;       /* Return value */
+
+    FUNC_ENTER_NOAPI(H5O_mtime_reset, FAIL);
+
+done:
+    FUNC_LEAVE(ret_value);
+}
+
+
+/*-------------------------------------------------------------------------
  * Function:	H5O_mtime_free
  *
  * Purpose:	Free's the message