Skip to content

M
MariaDB
Commit a8456e68 authored by sergefp@mysql.com's avatar sergefp@mysql.com
Browse files
Options

Numerous small fixes to index_merge read time estimates code

parent 50f29b0e
Show whitespace changes
Inline Side-by-side
Showing with 188 additions and 108 deletions
include/my_global.h
View file @ a8456e68
......@@ -666,6 +666,17 @@ extern double my_atof(const char*);
#define FLT_MAX ((float)3.40282346638528860e+38)
#endif
/* Define missing math constants. */
#ifndef M_PI
#define M_PI 3.14159265358979323846
#endif
#ifndef M_E
#define M_E 2.7182818284590452354
#endif
#ifndef M_LN2
#define M_LN2 0.69314718055994530942
#endif
/*
Max size that must be added to a so that we know Size to make
adressable obj.
......
mysql-test/r/index_merge.result
View file @ a8456e68
......@@ -109,7 +109,7 @@ key1 key2 key3 key4 key5 key6 key7 key8
explain select * from t0 where
(key1 < 3 or key2 < 3) and (key3 < 4 or key4 < 4) and (key5 < 2 or key6 < 2);
id select_type table type possible_keys key key_len ref rows Extra
1 SIMPLE t0 index_merge i1,i2,i3,i4,i5,i6 i5,i6 4,4 NULL 4 Using where
1 SIMPLE t0 index_merge i1,i2,i3,i4,i5,i6 i1,i2 4,4 NULL 6 Using where
explain select * from t0 where
(key1 < 3 or key2 < 3) and (key3 < 100);
id select_type table type possible_keys key key_len ref rows Extra
......
sql/item_create.cc
View file @ a8456e68
......@@ -18,10 +18,6 @@
#include "mysql_priv.h"
#ifndef M_PI
#define M_PI 3.14159265358979323846
#endif
Item *create_func_abs(Item* a)
{
return new Item_func_abs(a);
......
sql/item_func.cc
View file @ a8456e68
......@@ -750,7 +750,7 @@ double Item_func_log2::val()
double value=args[0]->val();
if ((null_value=(args[0]->null_value || value <= 0.0)))
return 0.0;
return log(value) / log(2.0);
return log(value) / M_LN2;
}
double Item_func_log10::val()
......
sql/opt_range.cc
View file @ a8456e68
......@@ -296,6 +296,9 @@ typedef struct st_qsel_param {
char min_key[MAX_KEY_LENGTH+MAX_FIELD_WIDTH],
max_key[MAX_KEY_LENGTH+MAX_FIELD_WIDTH];
bool quick; // Don't calulate possible keys
uint *imerge_cost_buff; /* buffer for index_merge cost estimates */
uint imerge_cost_buff_size; /* size of the buffer */
} PARAM;
static SEL_TREE * get_mm_parts(PARAM *param,Field *field,
......@@ -953,6 +956,7 @@ int SQL_SELECT::test_quick_select(THD *thd, key_map keys_to_use,
param.table=head;
param.keys=0;
param.mem_root= &alloc;
param.imerge_cost_buff_size= 0;
thd->no_errors=1; // Don't warn about NULL
init_sql_alloc(&alloc, thd->variables.range_alloc_block_size, 0);
......@@ -1011,7 +1015,7 @@ int SQL_SELECT::test_quick_select(THD *thd, key_map keys_to_use,
ha_rows found_records;
double found_read_time= read_time;
if (!get_quick_select_params(tree, &param, needed_reg, true,
if (!get_quick_select_params(tree, &param, needed_reg, false,
&found_read_time, &found_records,
&best_key))
{
......@@ -1254,31 +1258,30 @@ static int get_index_merge_params(PARAM *param, key_map& needed_reg,
*/
/*
It may be possible to use different keys for index_merge scans,
e.g. for query like
It may be possible to use different keys for index_merge scans, e.g. for
query like
...WHERE (key1 < c2 AND key2 < c2) OR (key3 < c3 AND key4 < c4)
we have to make choice between key1 and key2 for one scan and
between key3,key4 for another.
We assume we'll get the best way if we choose the best key read
inside each of the conjuncts. Comparison is done without 'using index'.
we have to make choice between key1 and key2 for one scan and between
key3, key4 for another.
We assume we'll get the best if we choose the best key read inside each
of the conjuncts.
*/
for (SEL_TREE **ptree= imerge->trees;
ptree != imerge->trees_next;
ptree++)
{
SEL_ARG **tree_best_key;
uint keynr;
tree_read_time= *read_time;
if (get_quick_select_params(*ptree, param, needed_reg, false,
if (get_quick_select_params(*ptree, param, needed_reg, true,
&tree_read_time, &tree_records,
&tree_best_key))
{
/*
Non-'index only' range scan on a one in index_merge key is more
expensive than other available option. The entire index_merge will be
more expensive then, too. We continue here only to update SQL_SELECT
members.
One of index scans in this index_merge is more expensive than entire
table read for another available option. The entire index_merge will
be more expensive then, too. We continue here only to update
SQL_SELECT members.
*/
imerge_too_expensive= true;
}
......@@ -1286,23 +1289,18 @@ static int get_index_merge_params(PARAM *param, key_map& needed_reg,
if (imerge_too_expensive)
continue;
uint keynr= param->real_keynr[(tree_best_key-(*ptree)->keys)];
imerge->best_keys[ptree - imerge->trees]= tree_best_key;
keynr= param->real_keynr[(tree_best_key-(*ptree)->keys)];
imerge_cost += tree_read_time;
if (pk_is_clustered && keynr == param->table->primary_key)
{
/* This is a Clustered PK scan, it will be done without 'index only' */
imerge_cost += tree_read_time;
have_cpk_scan= true;
cpk_records= tree_records;
}
else
{
/* Non-CPK scan, calculate time to do it using 'index only' */
imerge_cost += get_index_only_read_time(param, tree_records,keynr);
records_for_unique += tree_records;
}
}
DBUG_PRINT("info",("index_merge cost of index reads: %g", imerge_cost));
if (imerge_too_expensive)
......@@ -1359,14 +1357,23 @@ static int get_index_merge_params(PARAM *param, key_map& needed_reg,
DBUG_PRINT("info",("index_merge cost with rowid-to-row scan: %g", imerge_cost));
/* PHASE 3: Add Unique operations cost */
double unique_cost=
Unique::get_use_cost(param->mem_root, records_for_unique,
register uint unique_calc_buff_size=
Unique::get_cost_calc_buff_size(records_for_unique,
param->table->file->ref_length,
param->thd->variables.sortbuff_size);
if (unique_cost < 0.0)
if (param->imerge_cost_buff_size < unique_calc_buff_size)
{
if (!(param->imerge_cost_buff= (uint*)alloc_root(param->mem_root,
unique_calc_buff_size)))
DBUG_RETURN(1);
param->imerge_cost_buff_size= unique_calc_buff_size;
}
imerge_cost +=
Unique::get_use_cost(param->imerge_cost_buff, records_for_unique,
param->table->file->ref_length,
param->thd->variables.sortbuff_size);
imerge_cost += unique_cost;
DBUG_PRINT("info",("index_merge total cost: %g", imerge_cost));
if (imerge_cost < *read_time)
{
......@@ -1415,8 +1422,8 @@ inline double get_index_only_read_time(PARAM* param, ha_rows records,
tree in make range select for this SEL_TREE
param in parameters from test_quick_select
needed_reg in/out other table data needed by this quick_select
index_read_can_be_used if false, assume that 'index only' option is not
available.
index_read_must_be_used if true, assume 'index only' option will be set
(except for clustered PK indexes)
read_time out read time estimate
records out # of records estimate
key_to_read out SEL_ARG to be used for creating quick select
......@@ -1424,16 +1431,17 @@ inline double get_index_only_read_time(PARAM* param, ha_rows records,
static int get_quick_select_params(SEL_TREE *tree, PARAM *param,
key_map& needed_reg,
bool index_read_can_be_used,
bool index_read_must_be_used,
double *read_time, ha_rows *records,
SEL_ARG ***key_to_read)
{
int idx;
int result = 1;
bool pk_is_clustered= param->table->file->primary_key_is_clustered();
/*
Note that there may be trees that have type SEL_TREE::KEY but contain
no key reads at all. For example, tree for expression "key1 is not null"
where key1 is defined as "not null".
Note that there may be trees that have type SEL_TREE::KEY but contain no
key reads at all, e.g. tree for expression "key1 is not null" where key1
is defined as "not null".
*/
SEL_ARG **key,**end;
......@@ -1450,22 +1458,29 @@ static int get_quick_select_params(SEL_TREE *tree, PARAM *param,
(*key)->maybe_flag)
needed_reg.set_bit(keynr);
bool read_index_only= index_read_can_be_used?
param->table->used_keys.is_set(keynr): false;
bool read_index_only= index_read_must_be_used? true :
(bool)param->table->used_keys.is_set(keynr);
found_records=check_quick_select(param, idx, *key);
if (found_records != HA_POS_ERROR && found_records > 2 &&
read_index_only &&
(param->table->file->index_flags(keynr) & HA_KEY_READ_ONLY))
(param->table->file->index_flags(keynr) & HA_KEY_READ_ONLY) &&
!(pk_is_clustered && keynr == param->table->primary_key))
{
/* We can resolve this by only reading through this key. */
found_read_time=get_index_only_read_time(param, found_records, keynr);
}
else
{
/*
cost(read_through_index) = cost(disk_io) + cost(row_in_range_checks)
The row_in_range check is in QUICK_RANGE_SELECT::cmp_next function.
*/
found_read_time= (param->table->file->read_time(keynr,
param->range_count,
found_records)+
(double) found_records / TIME_FOR_COMPARE);
}
if (*read_time > found_read_time && found_records != HA_POS_ERROR)
{
*read_time= found_read_time;
......
sql/sql_class.h
View file @ a8456e68
......@@ -1233,8 +1233,16 @@ class Unique :public Sql_alloc
}
bool get(TABLE *table);
static double get_use_cost(MEM_ROOT *alloc, uint nkeys, uint key_size,
static double get_use_cost(uint *buffer, uint nkeys, uint key_size,
ulong max_in_memory_size);
inline static int get_cost_calc_buff_size(ulong nkeys, uint key_size,
ulong max_in_memory_size)
{
register ulong max_elems_in_tree=
(1 + max_in_memory_size / ALIGN_SIZE(sizeof(TREE_ELEMENT)+key_size));
return sizeof(uint)*(1 + nkeys/max_elems_in_tree);
}
friend int unique_write_to_file(gptr key, element_count count, Unique *unique);
friend int unique_write_to_ptrs(gptr key, element_count count, Unique *unique);
};
......
sql/uniques.cc
View file @ a8456e68
......@@ -72,112 +72,161 @@ Unique::Unique(qsort_cmp2 comp_func, void * comp_func_fixed_arg,
}
#ifndef M_PI
#define M_PI 3.14159265358979323846
#endif
/*
Calculate log2(n!)
NOTES
Stirling's approximate formula is used:
n! ~= sqrt(2*M_PI*n) * (n/M_E)^n
#ifndef M_E
#define M_E (exp((double)1.0))
#endif
Derivation of formula used for calculations is as follows:
log2(n!) = log(n!)/log(2) = log(sqrt(2*M_PI*n)*(n/M_E)^n) / log(2) =
= (log(2*M_PI*n)/2 + n*log(n/M_E)) / log(2).
*/
inline double log2_n_fact(double x)
{
return (2 * (((x)+1)*log(((x)+1)/M_E) + log(2*M_PI*((x)+1))/2 ) / log(2));
return (log(2*M_PI*x)/2 + x*log(x/M_E)) / M_LN2;
}
/*
Calculate cost of merge_buffers call.
Calculate cost of merge_buffers function call for given sequence of
input stream lengths and store the number of rows in result stream in *last.
SYNOPSIS
get_merge_buffers_cost()
buff_elems Array of #s of elements in buffers
elem_size Size of element stored in buffer
output_buff Pointer to storage for result buffer size
first Pointer to first merged element size
last Pointer to last merged element size
RETURN
Cost of merge_buffers operation in disk seeks.
NOTES
It is assumed that no rows are eliminated during merge.
The cost is calculated as
cost(read_and_write) + cost(merge_comparisons).
All bytes in the sequences is read and written back during merge so cost
of disk io is 2*elem_size*total_buf_elems/IO_SIZE (2 is for read + write)
NOTE
See comment near Unique::get_use_cost for cost formula derivation.
For comparisons cost calculations we assume that all merged sequences have
the same length, so each of total_buf_size elements will be added to a sort
heap with (n_buffers-1) elements. This gives the comparison cost:
total_buf_elems* log2(n_buffers) / TIME_FOR_COMPARE_ROWID;
*/
static double get_merge_buffers_cost(uint* buff_sizes, uint elem_size,
int last, int f,int t)
static double get_merge_buffers_cost(uint *buff_elems, uint elem_size,
uint *output_buff, uint *first,
uint *last)
{
uint sum= 0;
for (int i=f; i <= t; i++)
sum+= buff_sizes[i];
buff_sizes[last]= sum;
uint total_buf_elems= 0;
for (uint *pbuf= first; pbuf <= last; pbuf++)
total_buf_elems+= *pbuf;
*last= total_buf_elems;
int n_buffers= t - f + 1;
double buf_length= sum*elem_size;
int n_buffers= last - first + 1;
return (((double)buf_length/(n_buffers+1)) / IO_SIZE) * 2 * n_buffers +
buf_length * log(n_buffers) / (TIME_FOR_COMPARE_ROWID * log(2.0));
/* Using log2(n)=log(n)/log(2) formula */
return 2*((double)total_buf_elems*elem_size) / IO_SIZE +
total_buf_elems*log(n_buffers) / (TIME_FOR_COMPARE_ROWID * M_LN2);
}
/*
Calculate cost of merging buffers into one in Unique::get, i.e. calculate
how long (in terms of disk seeks) the two call
how long (in terms of disk seeks) the two calls
merge_many_buffs(...);
merge_buffers(...);
will take.
SYNOPSIS
get_merge_many_buffs_cost()
alloc memory pool to use
maxbuffer # of full buffers.
max_n_elems # of elements in first maxbuffer buffers.
last_n_elems # of elements in last buffer.
elem_size size of buffer element.
buffer buffer space for temporary data, at least
Unique::get_cost_calc_buff_size bytes
maxbuffer # of full buffers
max_n_elems # of elements in first maxbuffer buffers
last_n_elems # of elements in last buffer
elem_size size of buffer element
NOTES
It is assumed that maxbuffer+1 buffers are merged, first maxbuffer buffers
contain max_n_elems each, last buffer contains last_n_elems elements.
maxbuffer+1 buffers are merged, where first maxbuffer buffers contain
max_n_elems elements each and last buffer contains last_n_elems elements.
The current implementation does a dumb simulation of merge_many_buffs
actions.
function actions.
RETURN
>=0 Cost of merge in disk seeks.
<0 Out of memory.
Cost of merge in disk seeks.
*/
static double get_merge_many_buffs_cost(MEM_ROOT *alloc,
static double get_merge_many_buffs_cost(uint *buffer,
uint maxbuffer, uint max_n_elems,
uint last_n_elems, int elem_size)
{
register int i;
double total_cost= 0.0;
int lastbuff;
uint* buff_sizes;
uint *buff_elems= buffer; /* #s of elements in each of merged sequences */
uint *lastbuff;
if (!(buff_sizes= (uint*)alloc_root(alloc, sizeof(uint) * (maxbuffer + 1))))
return -1.0;
/*
Set initial state: first maxbuffer sequences contain max_n_elems elements
each, last sequence contains last_n_elems elements.
*/
for(i = 0; i < (int)maxbuffer; i++)
buff_sizes[i]= max_n_elems;
buff_sizes[maxbuffer]= last_n_elems;
buff_elems[i]= max_n_elems;
buff_elems[maxbuffer]= last_n_elems;
/*
Do it exactly as merge_many_buff function does, calling
get_merge_buffers_cost to get cost of merge_buffers.
*/
if (maxbuffer >= MERGEBUFF2)
{
/* Simulate merge_many_buff */
while (maxbuffer >= MERGEBUFF2)
{
lastbuff=0;
for (i = 0; i <= (int) maxbuffer - MERGEBUFF*3/2; i += MERGEBUFF)
total_cost += get_merge_buffers_cost(buff_sizes, elem_size,
lastbuff++, i, i+MERGEBUFF-1);
total_cost+=get_merge_buffers_cost(buff_elems, elem_size, lastbuff++,
buff_elems + i,
buff_elems + i + MERGEBUFF-1);
total_cost += get_merge_buffers_cost(buff_sizes, elem_size,
lastbuff++, i, maxbuffer);
total_cost+=get_merge_buffers_cost(buff_elems, elem_size, lastbuff++,
buff_elems + i,
buff_elems + maxbuffer);
maxbuffer= (uint)lastbuff-1;
}
}
/* Simulate final merge_buff call. */
total_cost += get_merge_buffers_cost(buff_sizes, elem_size, 0, 0,
maxbuffer);
total_cost += get_merge_buffers_cost(buff_elems, elem_size, buff_elems,
buff_elems, buff_elems + maxbuffer);
return total_cost;
}
/*
Calclulate cost of using Unique for processing nkeys elements of size
Calculate cost of using Unique for processing nkeys elements of size
key_size using max_in_memory_size memory.
SYNOPSIS
Unique::get_use_cost()
buffer space for temporary data, use Unique::get_cost_calc_buff_size
to get # bytes needed.
nkeys #of elements in Unique
key_size size of each elements in bytes
max_in_memory_size amount of memory Unique will be allowed to use
RETURN
>=0 Cost in disk seeks.
<0 Out of memory.
Cost in disk seeks.
NOTES
cost(using_unqiue) =
......@@ -190,17 +239,15 @@ static double get_merge_many_buffs_cost(MEM_ROOT *alloc,
comparisons, where n runs from 1 tree_size (we assume that all added
elements are different). Together this gives:
n_compares = 2*(log2(2) + log2(3) + ... + log2(N+1)) = 2*log2((N+1)!) =
= 2*ln((N+1)!) / ln(2) = {using Stirling formula} =
= 2*( (N+1)*ln((N+1)/e) + (1/2)*ln(2*pi*(N+1)) / ln(2).
n_compares = 2*(log2(2) + log2(3) + ... + log2(N+1)) = 2*log2((N+1)!)
then cost(tree_creation) = n_compares*ROWID_COMPARE_COST;
Total cost of creating trees:
(n_trees - 1)*max_size_tree_cost + non_max_size_tree_cost.
Approximate value of log2(N!) is calculated by log2_n_fact function.
2. Cost of merging.
If only one tree is created by Unique no merging will be necessary.
Otherwise, we model execution of merge_many_buff function and count
......@@ -213,7 +260,7 @@ static double get_merge_many_buffs_cost(MEM_ROOT *alloc,
these will be random seeks.
*/
double Unique::get_use_cost(MEM_ROOT *alloc, uint nkeys, uint key_size,
double Unique::get_use_cost(uint *buffer, uint nkeys, uint key_size,
ulong max_in_memory_size)
{
ulong max_elements_in_tree;
......@@ -221,15 +268,16 @@ double Unique::get_use_cost(MEM_ROOT *alloc, uint nkeys, uint key_size,
int n_full_trees; /* number of trees in unique - 1 */
double result;
max_elements_in_tree= max_in_memory_size /
ALIGN_SIZE(sizeof(TREE_ELEMENT)+key_size);
max_elements_in_tree=
max_in_memory_size / ALIGN_SIZE(sizeof(TREE_ELEMENT)+key_size);
n_full_trees= nkeys / max_elements_in_tree;
last_tree_elems= nkeys % max_elements_in_tree;
/* Calculate cost of creating trees */
result= log2_n_fact(last_tree_elems);
result= 2*log2_n_fact(last_tree_elems + 1.0);
if (n_full_trees)
result+= n_full_trees * log2_n_fact(max_elements_in_tree);
result+= n_full_trees * log2_n_fact(max_elements_in_tree + 1.0);
result /= TIME_FOR_COMPARE_ROWID;
DBUG_PRINT("info",("unique trees sizes: %u=%u*%lu + %lu", nkeys,
......@@ -241,13 +289,15 @@ double Unique::get_use_cost(MEM_ROOT *alloc, uint nkeys, uint key_size,
/*
There is more then one tree and merging is necessary.
First, add cost of writing all trees to disk.
First, add cost of writing all trees to disk, assuming that all disk
writes are sequential.
*/
result += n_full_trees * ceil(key_size*max_elements_in_tree / IO_SIZE);
result += ceil(key_size*last_tree_elems / IO_SIZE);
result += DISK_SEEK_BASE_COST * n_full_trees *
ceil(key_size*max_elements_in_tree / IO_SIZE);
result += DISK_SEEK_BASE_COST * ceil(key_size*last_tree_elems / IO_SIZE);
/* Cost of merge */
double merge_cost= get_merge_many_buffs_cost(alloc, n_full_trees,
double merge_cost= get_merge_many_buffs_cost(buffer, n_full_trees,
max_elements_in_tree,
last_tree_elems, key_size);
if (merge_cost < 0.0)
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7