C data types

lastmod 2021-06-16created 2021-06-164 edits

In the C programming language, data types are declarations of memory locations or variables that determine the characteristics of the data.

The C language provides basic arithmetic types such as integer and real number types, along with syntax for building array and compound types.

Standard Data Type Limits

Numeric Types

Type	Size (in bits)	Format	Minimum Value	Maximum Value	Literal Suffix	Sig. Digits
BYTE	8	signed integer	-128	+127		2+
UBYTE	8	unsigned integer	0	+255		2+
SHORT	16	signed integer	-32768	+32767		4+
USHORT	16	unsigned integer	0	65535		4+
INTEGER	32	signed integer	-2147483648	+2147483647	%, l	9+
UINTEGER	32	unsigned integer	0	4294967295	ul	9+
LONG	[*]	signed integer	[*]	[*]	&	[*]
ULONG	[*]	unsigned integer	[*]	[*]		[*]
LONGINT	64	signed integer	-9 223 372 036 854 775 808	+9 223 372 036 854 775 807	ll	18+
ULONGINT	64	unsigned integer	0	+18 446 744 073 709 551 615	ull	19+
SINGLE	32	floating point	[**]+/-1.401 298 E-45	[**]+/-3.402 823 E+38	!, f	6+
DOUBLE	64	floating point	[**]+/-4.940 656 458 412 465 E-324	[**]+/-1.797 693 134 862 316 E+308	#	15+

[*] Long and Ulong data types vary with platform. Currently, they are aliases to either Integer or LongInt and their unsigned versions, respectively.
[**] The minimum and maximum values for the floating-point types Single and Double are, respectively, the values closest to zero and the values closest to positive and negative infinity.

C Datatype 32bit 64bit

KLDP Considerations for source compatibility from 32bit to 64bit
64Bit development story
http://redmist.springnote.com/pages/6717419.xhtml
[http://msdn.microsoft.com/ja-jp/library/bb147385(VS.85.aspx 64Bit programming for game developers]

A summary of the variable types that changed in the move from 32bit to 64bit environments. The results below are from my own tests.

CPU	Intel x86	Intel x86	Intel x86_64	Intel x86_64	Intel x86_64	Intel x86_64	Intel x86_64
OS	Win7 32bit	Win7 32bit	Win7 64bit	Win7 64bit	Win7 64bit	Win7 64bit	Fedora 64bit
Compiler	VS9 32bit	MinGW g++ 32bit	VS10 32bit	VS10 64bit	MinGW g++ 32bit	MinGW g++ 64bit	cc 64bit
size_t	4byte	4byte	4byte	8byte	4byte	8byte	8byte
ssize_t	-	4byte	-	-	4byte	8byte	8byte
char	1byte	1byte	1byte	1byte	1byte	1byte	1byte
wchar_t	2byte	2byte	2byte	2byte	2byte	2byte	4byte
short	2byte	2byte	2byte	2byte	2byte	2byte	2byte
int	4byte	4byte	4byte	4byte	4byte	4byte	4byte
long	4byte	4byte	4byte	4byte	4byte	4byte	8byte
long long	8byte	8byte	8byte	8byte	8byte	8byte	8byte
float	4byte	4byte	4byte	4byte	4byte	4byte	4byte
double	8byte	8byte	8byte	8byte	8byte	8byte	8byte
long double	8byte	12byte	8byte	8byte	12byte	16byte	16byte
`` (pointer)*	4byte	4byte	4byte	8byte	4byte	8byte	8byte

Problematic code

Using pointers to variable types of different sizes. (when stack corruption is a concern)
DWORD definitions.
x86-family assembly compatibility.
Mistaken assumptions about Implementation functions. (e.g., not making use of the meaning of size_t at all)
Struct size alignment.
Detecting 64bit via Pre-define.

Cache tuning

If you use gcc, it may be worth looking at the __builtin_expect((long)(m_expression),(long)(m_value)) built-in function for better performance on 64bit. By adjusting the probability that a conditional expression is rarely or likely to be executed, it can lead the compiler to optimize for the best use of the Cache. Since 64bit machine code provides ways to use the Cache effectively, using this well should maximize performance.

Favorite site

Wikipedia (en) C data types

author Zer0 sitemap