#include <array_growth_policy.h>

Public Member Functions
	prime_growth_policy (std::size_t &min_bucket_count_in_out)

std::size_t	bucket_for_hash (std::size_t hash) const noexcept

std::size_t	next_bucket_count () const

std::size_t	max_bucket_count () const

void	clear () noexcept

Detailed Description

Grow the hash table by using prime numbers as bucket count. Slower than tsl::ah::power_of_two_growth_policy in general but will probably distribute the values around better in the buckets with a poor hash function.

To allow the compiler to optimize the modulo operation, a lookup table is used with constant primes numbers.

With a switch the code would look like:

switch(iprime) { // iprime is the current prime of the hash table
    case 0: hash % 5ul;
            break;
    case 1: hash % 17ul;
            break;
    case 2: hash % 29ul;
            break;
    ...
}    

Due to the constant variable in the modulo the compiler is able to optimize the operation by a series of multiplications, substractions and shifts.

The 'hash % 5' could become something like 'hash - (hash * 0xCCCCCCCD) >> 34) * 5' in a 64 bits environement.

Constructor & Destructor Documentation

◆ prime_growth_policy()

tsl::ah::prime_growth_policy::prime_growth_policy ( std::size_t & min_bucket_count_in_out )

inlineexplicit

Member Function Documentation

◆ bucket_for_hash()

std::size_t tsl::ah::prime_growth_policy::bucket_for_hash ( std::size_t hash ) const

inlinenoexcept

◆ clear()

void tsl::ah::prime_growth_policy::clear ( )

inlinenoexcept

◆ max_bucket_count()

std::size_t tsl::ah::prime_growth_policy::max_bucket_count ( ) const

inline

◆ next_bucket_count()

std::size_t tsl::ah::prime_growth_policy::next_bucket_count ( ) const