unordered/doc/unordered/rationale.adoc

[#rationale]

:idprefix: rationale_

= Implementation Rationale

The intent of this library is to implement the unordered
containers in the standard, so the interface was fixed. But there are
still some implementation decisions to make. The priorities are
conformance to the standard and portability.

The http://en.wikipedia.org/wiki/Hash_table[Wikipedia article on hash tables^]
has a good summary of the implementation issues for hash tables in general.

== Data Structure

By specifying an interface for accessing the buckets of the container the
standard pretty much requires that the hash table uses chained addressing.

It would be conceivable to write a hash table that uses another method. For
example, it could use open addressing, and use the lookup chain to act as a
bucket but there are some serious problems with this:

* The standard requires that pointers to elements aren't invalidated, so
  the elements can't be stored in one array, but will need a layer of
  indirection instead - losing the efficiency and most of the memory gain,
  the main advantages of open addressing.
* Local iterators would be very inefficient and may not be able to
  meet the complexity requirements.
* There are also the restrictions on when iterators can be invalidated. Since
  open addressing degrades badly when there are a high number of collisions the
  restrictions could prevent a rehash when it's really needed. The maximum load
  factor could be set to a fairly low value to work around this - but the
  standard requires that it is initially set to 1.0.
* And since the standard is written with a eye towards chained
  addressing, users will be surprised if the performance doesn't reflect that.

So chained addressing is used.

== Number of Buckets

There are two popular methods for choosing the number of buckets in a hash
table. One is to have a prime number of buckets, another is to use a power
of 2.

Using a prime number of buckets, and choosing a bucket by using the modulus
of the hash function's result will usually give a good result. The downside
is that the required modulus operation is fairly expensive. This is what the
containers used to do in most cases.

Using a power of 2 allows for much quicker selection of the bucket to use,
but at the expense of losing the upper bits of the hash value. For some
specially designed hash functions it is possible to do this and still get a
good result but as the containers can take arbitrary hash functions this can't
be relied on.

To avoid this a transformation could be applied to the hash function, for an
example see
http://web.archive.org/web/20121102023700/http://www.concentric.net/~Ttwang/tech/inthash.htm[Thomas Wang's article on integer hash functions^].
Unfortunately, a transformation like Wang's requires knowledge of the number
of bits in the hash value, so it was only used when `size_t` was 64 bit.

Since release 1.79.0, https://en.wikipedia.org/wiki/Hash_function#Fibonacci_hashing[Fibonacci hashing]
is used instead. With this implementation, the bucket number is determined
by using `(h * m) >> (w - k)`, where `h` is the hash value, `m` is the golden
ratio multiplied by `2^w`, `w` is the word size (32 or 64), and `2^k` is the
number of buckets. This provides a good compromise between speed and
distribution.