Printable View
I'm very new to C++ and was curious why long and int are the same size (both are 4 bytes/32 bits) on my computer. But according to a data type table an int should be 16 bits and a long 32 bits. Why is this so?
In this case, why would I even consider using a long if they are both the same?
Thank you.
Data types size vary depending on platform and compiler. That is to say, on some systems, long is bigger than int, but on x86 platforms, they are typically the same size.
Although, if you are looking for a specific size of a type, then look no further than stdint.h and its (u)intxx_t typedefs.
That table probably gives the minimum sizes (though these are formally given as numeric limits, not number of bits). Other than the minimum limits, the other requirement is that sizeof(int) <= sizeof(long), and having int and long both 32 bits satisfies the requirements.Quote:
But according to a data type table an int should be 16 bits and a long 32 bits. Why is this so?
long may be 64 bit on some platforms, and int may be 16 bit on other platforms.Quote:
In this case, why would I even consider using a long if they are both the same?
Oh, I see. Thank you for clearing that up. So on x86 platfoms there is no point in using the long data type?
EDIT: Damn, that's some fast replies...
It may or may not be.
For pure size, there's no real reason, but try to match types. If a function returns long, then use a long data type, for example.
This makes it more portable.
Thanks very much. I always like to make sure I get something. Again, thanks for the fast and informative replies laserlight and Elysia.
Formally, long is 32 bits, int is the native register size, which on intel processors is 32 bits. Technicalyl on a 64 bit machine an int is 64 bits, although it is usually treated as 32 bits there as well.
That is not true. The requirement is just that long have a minimum range of [-2147483647, 2147483647], and that sizeof(int) <= sizeof(long). Consequently, long is at least 32 bits.Quote:
Formally, long is 32 bits
Yes, the C++ Standard does state that "plain ints have the natural size suggested by the architecture of the execution environment39".Quote:
int is the native register size, which on intel processors is 32 bits. Technicalyl on a 64 bit machine an int is 64 bits, although it is usually treated as 32 bits there as well.
In fact, Windows compiler for x64 doesn't even make long 64-bits, but rather uses their own type (__int64) for 64-bit integers.Quote:
Originally Posted by abachler
I don't think it's correct to say that "int is the native register size", but rather that it's a size that is generally efficient on the machin and big enough to hold a few tens of thousands or more. Which so happens to be 32 bit on x86/x64. To extend int to 64-bit would for 99% of all applications just cause a waste of space because the integers are now taking up 8 bytes instead of 4. Making "long" 64-bit would actually make sense, but MS has been using LONG for so many of their internal types that it didn't make a good choice for the API to change all over the place. And of course, redefining the API would break a number of existing applications. So they took the choice of making the API stay the same, and make long 32 bits.
--
Mats
Well, teh ansi standard states that long is specificlly larger than short, and that int is the native reigister size.
No, it states that long is specifically not shorter than short, and it does not quite say that int is the native register size, though it does hint along that direction:Quote:
Well, teh ansi standard states that long is specificlly larger than short, and that int is the native reigister size.
The above references a footnote concerning plain ints:Quote:
Originally Posted by Paragraph 2, Section 3.9.1, C++ Standard (2003)
Quote:
that is, large enough to contain any value in the range of INT_MIN and INT_MAX, as defined in the header <climits>.