问题
Would any compiler experts be able to comment on the efficient use of boolean values? Specifically, is the compiler able to optimize a std::vector<boolean> to use minimal memory? Is there an equivalent data structure that would?
Back in the day, there were languages that had compilers that could compress an array of booleans to a representation of just one bit per boolean value. Perhaps the best that could be done for C++ is to use std::vector<char> to store the boolean values for minimal memory usage?
The use case here would be storing hundreds of millions of boolean values, where a single byte would save lots of space over 4 or more bytes per value and a single bit, even more.
回答1:
std::vector for bool is a template specialization that does what you are asking for.
You can read more here.
You may also want to explore the standard bitset.
回答2:
See std::vector
Specializations
The standard library provides a specialization of std::vector for the type bool, which is optimized for space efficiency.
vector<bool> space-efficient dynamic bitset (class template specialization)
and from "Working Draft C++, 2012-11-02"
23.3.7 Class vector [vector.bool]
1 To optimize space allocation, a specialization of vector for bool elements is provided:
template <class Allocator> class vector<bool, Allocator> {
...
}3 There is no requirement that the data be stored as a contiguous allocation of bool values. A space-optimized representation of bits is recommended instead.
So there is no requirement, but only a recommendation, to store the bool values as bits.
回答3:
Note, that vector<bool> is not a container, however it pretends to be one and provides iterators.
One day that may cause confusion and errors if you treat it like a normal container, e.g. trying to get an address of elements.
You may consider std::bitset or boost::dynamic_bitset if you need to store 1 bit per Boolean value. These data structures do not pretend to be containers, so it is unlikely you make any errors when using any of them, especially in template code.
回答4:
In what is widely considered to be a flaw in the standard, std::vector is specialised to use a single bit to represent each bool value.
If that happens to be what you are looking for, then just use it.
回答5:
As a standard-agnostic way of guaranteeing efficient storage, you could create your own Bitvector class. Essentially for every 8 bool values you only need to allocate a single char and then you can store each bool in a single bit. You can then use bit shifting techniques in the accessors/mutators to store/retrieve your individual bits.
One such example is outlined in Ron Penton and André LaMothe's Data Structures for Game Programmers (which I'd also recommend as a general data structure reference). It's not too difficult to write your own though, and, though I haven't searched at great length, there are probably some further examples on the Internet.
来源：https://stackoverflow.com/questions/15126881/efficient-use-of-boolean-true-and-false-in-c