Why use float?

**Sentral** · 07-14-2005

Hello, I have come across a problem. I don't understand why someone would use float nowadays. Float is only accurate up to about 6 numbers, but uses less memory. Wouldn't someone rather use double, it has twice the accuracy, but uses more memory. Would this be a problem if you were holding alot of large nmbers, or will todays processors hold up?

**kuphryn** · 07-14-2005

You said it. Memory usage is key, so save the bits.

Kuphryn

**Zach L.** · 07-14-2005

Well, perhaps you don't need the same precision (or don't have very precise data), but you need a very sizable number... then memory becomes an issue.

Not really an issue for most applications, though.

**Sentral** · 07-14-2005

OK. Thanks.

**Daved** · 07-14-2005

You should always use double unless you specifically need the memory savings and don't need the precision - double is the default floating point type in C++. It is the same reason you should use int instead of short. float would be better named short double.

**Sentral** · 07-14-2005

Would default int be considered long? In other words would I have to do long int for it to be long , or is int automatically long?

**Zach L.** · 07-14-2005

In terms of size, you have the following guarantee:
short <= int <= long

On most common (new) systems, with most compilers I have run across (nowhere near exhaustive), int and long are the same size.

**DougDbug** · 07-14-2005

An int is the same as a short. If you want a long, you have to specify long.

[EDIT]-------------------------
But like Zach says, a long and short can be the same size.

**Sentral** · 07-14-2005

Ok.Cooool.

**Daved** · 07-14-2005

>> An int is the same as a short. If you want a long, you have to specify long.

This is not correct. There are three sizes of integers, short int (i.e. short), int, and long int (i.e. long). Zach L.'s size display is correct.

**sand_man** · 07-14-2005

I thought that the size of an int was implementation specific.
two sizes, short and long. int could be either one of them

Code:

#include <iostream>

int main(int argc, char *argv[])
{
  std::cout << sizeof(short) << std::endl;
  std::cout << sizeof(int) << std::endl;
  std::cout << sizeof(long) << std::endl;

  return 0;
}

/*My output
* 2
* 4
*/4

**Daved** · 07-14-2005

There are three potential sizes, and they are implementation specific. They could all three be the same, or all three be different, or two the same and one different. On a 64 bit machine you might have 2, 4, 8. On another machine you might have 2, 2, 2. The rule is short <= int <= long.

**valis** · 07-14-2005

I've never really liked the C types, I see why they don't use the 'standard' names as those are usually architecturally specific.
If your compiling for IA-32 you can almost be guaranteed an int is a (IA-32) double word, a long is a double word, a short is a word, and a char is a byte and a long long or long int is a quad word. I like microsofts __int8 __int16 __int32 and __int64 types because they are more meaningful to me.

**Ganoosh** · 07-14-2005

Double is more used but float is a lot better for something like a game to save memory.

**major_small** · 07-14-2005

Originally Posted by valis

I've never really liked the C types, I see why they don't use the 'standard' names as those are usually architecturally specific.
If your compiling for IA-32 you can almost be guaranteed an int is a (IA-32) double word, a long is a double word, a short is a word, and a char is a byte and a long long or long int is a quad word. I like microsofts __int8 __int16 __int32 and __int64 types because they are more meaningful to me.

because they're architecturally specific.

Standard C/C++ is designed for use on as many platforms as possible. that's why you can port a standard C program back and forth from linux to windows without a problem. The standards leave all the architecture-specific things to the compiler, and sets guidelines for the compiler to follow. that's why the sizes are defined in relation to eachother.

Thread: Why use float?

Thread Tools

Search Thread

Display

Why use float?

Similar Threads

C in UNIX environment; external bubble sort doesn't work

Okay, giant issue, I think I'm treading into territory outside of cboards maybe...

Could somebody please help me with this C program

Half-life SDK, where are the constants?

How do you search & sort an array?