What is responsible for the space taken to store variables?

[yahaa]

the space taken to store variables of type integer, float etc depends on which factor?????
operating system???or compiler?????or computer architechture??????
if possible give an example

 

[sinu_gowde]

space for variables

Hi Yahaa,
I think its dependent on the compiler design. i am saying this because, compare the Turbo C & VC++ compilers... if a define a integer variable in a C program & compile it with both compilers seperately. you can see the deference. in Turbo C it allocates 2 words but in VC++ it allocates 4 words.

does anybody have any comments on this...

cheers...

 

[echo47]

space for variables

yahaa, are you asking about C, or some other language?

Yes, C integer sizes depend on the compiler, although the ANSI C standard specifies minimum number of integer bits, and floating point resolution (but not number of floating point bits). Refer to your header files limits.h, float.h, stdint.h.

char is at least 8 bits.
short int is at least 16 bits.
int is at least 16 bits.
long int is at least 32 bits.
long long int is at least 64 bits.
charbits <= shortintbits <= intbits <= longintbits <= longlongintbits.

The compiler authors usually choose integer and floating-point sizes that suit the CPU architecture and comply with the ANSI C standard. Turbo C dates back to the days of 16-bit CPUs such as 8086 and 80286, so it uses 16-bit int. Most hardware floating point units comply with IEEE Std 754, so we see a lot of 32-bit floats and 64-bit doubles.

 

[fdarkangel]

space for variables

you can usually safely assume that a char is 1 byte, a short in 2 bytes, a long is 4 bytes, and long long (c99) is 8 bytes.
however, the size of int depends on the target processor. it's the same size of processor registers.
floats are 32-bits. however, depending on architecture, you can see 64-bits, 80-bits, 128-bits extended versions of floating points, under names double and long double.