Question

Possible Duplicate:
int_least64_t vs int_fast64_t vs int64_t

For example, what is the difference between following types and why/when use them?

• int64_t
• int_least64_t
• int_fast64_t

Answer 1

int64_t is a type of exactly 64 bit width, two’s complement representation, and no padding bits. Such is useful e.g. for work in binary interfaces; its presence is optional, defined only on architectures that have an integer type of exactly that width.

int_least64_t is the smallest possible type of at least 64 bit width, with implementation-defined representation and padding. It might be wider, if exact 64 bit width is unavailable, e.g. for work in algorithmics where you need a specific minimum value range.

int_fast64_t is a type of at least 64 bit width. It might be wider if the wider type can be handled more efficiently by the CPU, trading memory footprint for speed.

The first one is very specific, but optional. The other two give the compiler some wiggle room if the underlying hardware is tricky, i.e. you tell the compiler under which conditions you would accept a wider type than requested.