ParaEngine::CNPLPool_Char_alloc< UserAllocator, Mutex > Class Template Reference

This class can be used as a custom allocator for stl containers, which constantly create and delete small or medium sized objects. More...

#include <NPLMemPool.h>

Classes
struct	rebind
struct	SingletonMemPool_Type

Public Types
typedef char	value_type
typedef UserAllocator	user_allocator
typedef Mutex	mutex_type
typedef value_type *	pointer
typedef const value_type *	const_pointer
typedef value_type &	reference
typedef const value_type &	const_reference
typedef ParaEngine::PoolThreadSafe< UserAllocator, Mutex >	pool_type
typedef pool_type::size_type	size_type
typedef pool_type::difference_type	difference_type

Public Member Functions
	CNPLPool_Char_alloc (const CNPLPool_Char_alloc< UserAllocator, Mutex > &)
bool	operator== (const CNPLPool_Char_alloc &) const
bool	operator!= (const CNPLPool_Char_alloc &) const

Static Public Member Functions
static pointer	address (reference r)
static const_pointer	address (const_reference s)
static size_type	max_size ()
static pointer	allocate (const size_type n)
static pointer	allocate (const size_type n, const void *const)
static void	deallocate (const pointer ptr, const size_type n)

Static Protected Attributes
static SingletonMemPool_Type	s_mem_pools
	predefined memory pool (free lists)

Detailed Description

template<typename UserAllocator, typename Mutex>
class ParaEngine::CNPLPool_Char_alloc< UserAllocator, Mutex >

This class can be used as a custom allocator for stl containers, which constantly create and delete small or medium sized objects.

it is optimized for frequently used temporary NPL objects, such as NPL related message body, file names, message bodies, etc. it is faster than std::allocator and boost::pool_alloc. But it costs more (double at most) memory than boost::pool_alloc, because we may return more memory than requested.

Internally, it has 7 free lists (boost::pool). Each free list only serves one fixed-sized object at a time. Each pool has their own light weighted locks for thread-safe access. The free list sizes are 32,64,128,256,512,1024,2048; we will use the default allocator for sizes over 2048. One can configure the internal freelists by editing the macro NPL_char_pool_init_size_bytes, NPL_char_pool_init_size, NPL_char_pool_count

[thread safe] Example: typedef std::basic_string<char, std::char_traits<char>, ParaEngine::CNPLPool_Char_alloc<> > NPLString;

This is the standard boost version(DONOT use this) #include <boost/pool/pool_alloc.hpp> typedef std::basic_string<char, std::char_traits<char>, boost::pool_allocator > String;

---

The documentation for this class was generated from the following file:

• Client/trunk/ParaEngineClient/NPL/NPLMemPool.h