|
@@ -341,11 +341,17 @@ $ kamcmd htable.dump htable
|
|
</listitem>
|
|
</listitem>
|
|
<listitem>
|
|
<listitem>
|
|
<para>
|
|
<para>
|
|
- <emphasis>size</emphasis> - number specifying the size of hash
|
|
|
|
- table. Larger value means less collisions. The number of entries
|
|
|
|
- (aka slots or buckets) in the table is 2^size. The possible range
|
|
|
|
|
|
+ <emphasis>size</emphasis> - number to control how many slots
|
|
|
|
+ (buckets) to create for the hash table. Larger value means more
|
|
|
|
+ slots with higher probability for less collisions. The actual number
|
|
|
|
+ slots (or buckets) created for the table is 2^size. The possible range
|
|
for this value is from 2 to 31, smaller or larger values will be
|
|
for this value is from 2 to 31, smaller or larger values will be
|
|
- increased to 3 (8 slots) or decreased to 14 (16384 slots).
|
|
|
|
|
|
+ increased to 3 (8 slots) or decreased to 14 (16384 slots). Note
|
|
|
|
+ that each slot can store more than one item, when there are
|
|
|
|
+ collisions of hash ids computed for keys. The items in the same
|
|
|
|
+ slot are stored in a linked list. In other words, the size is not
|
|
|
|
+ setting a limit of how many items can be stored in a hash table, as
|
|
|
|
+ long as there is enough free shared memory, new items can be added.
|
|
</para>
|
|
</para>
|
|
</listitem>
|
|
</listitem>
|
|
<listitem>
|
|
<listitem>
|