CODEWITHSUNDEEP
ctypesbytesizebit
Andy Sukowski-Bang
Andy Sukowski-Bang

Reputation: 1420

Is the minimum size of datatypes defined in bits or bytes?

I know that the number of bits in a char is defined in CHAR_BIT from <limits.h> and that sizeof(char) is always 1. But what about other basic datatypes, are their sizes defined relative to CHAR_BIT?

For example the minimum size of an int in C is defined as 16 bits on Wikipedia, but on other sites like GeeksforGeeks the minimum size is defined as 2 bytes. Which definition is correct as a byte is not necessarily the same as 8 bits.

Upvotes: 0

Views: 483

Answers (2)

Barmar
Barmar

Reputation: 780688

The minimum required range of values of a type is defined numerically, not in terms of bits or bytes. 5.2.4.2.1 Sizes of integer types <limits.h> says has definitions like:

  • minimum value for an object of type short int
    SHRT_MIN -32767 // -(215 - 1)
  • maximum value for an object of type short int
    SHRT_MAX +32767 // 215 - 1

All the type definitions happen to match up with implementing the types using either two's complement or sign-magnitude representation with a multiple of 8 bit bytes. But since they're just minimums, it doesn't require such representations.

Upvotes: 4

chux
chux

Reputation: 153338

Is the minimum size of datatypes defined in bits or bytes?

Mostly it is defined by a combination of C specification properties and CHAR_BIT. CHAR_BIT >= 8.

int has a minimum range [-32767... 32767] obliging al least 16 bits to encode that.

With the common CHAR_BIT == 8, that is 16-bits or 2 byte".

With the common CHAR_BIT == 16, that is 16-bits or 1 "byte". A byte being 16-bits in this implementation.

With the common CHAR_BIT == 64, that is 64-bits or 1 "byte". 64 bits, as no type is smaller than char.

Exact size integer types like (u)intN_t are effectively defined by bit size. They too have min/max range values, but with the no padding and 2's complement requirement are effectively defined by bits. Note: these are optional types.

Some esoteric thoughts about the minnum floating point bit size.

Upvotes: 3

Related Questions