Guys, I have the following code in C:

http://paste.ubuntu.com/6904734/

I am trying to raise the number 2.55 to two different exponents (2 in one case, and 3 in another).

I haven't gotten to the part of the book where it tells me to import <math.h> so I won't be using that.

Since C doesn't support exponents directly, I thought I could turn the decimal number into a hexadecimal number, which does support exponents.

2.55 in hexadecimal is 2.8CCC; but I am not getting the right answer.

x = 0X2.8CCCp3

When I print that out, I should be getting 2.55 ^ 3 = 16.5814. Instead it gives me 20.3999.

Where am I going wrong?

If you don't want to use the pow function defined in math.h, then you could always write your own. Write a for loop that multiplies a number with itself 'x *= x' for the number of time you need.

The p in 0X2.8CCCp3 doesn't mean exactly exponent. It's like scientific notation where 1e3 means 1 * 10^3 but in base 2 instead:


Though for this particular case, I would just write out the multiplies inline (all three of them).

"C" certainly does "support exponents," in its intrinsic float type.

Any and all assumptions as to the binary format of the underlying value is entirely implementation-dependent and, I would say, "to be discouraged." If you do want to know how, say, Intel does it, simply look at their microprocessor documentation.

Thanks this is very helpful.

I guess I will have to go with the for loop suggestion.

Thanks sundial but that seems to be an entirely different sphere of difficulty. What I was looking for was the C equivalent of Python's 2 ** 3, which evaluates to 8.

That'w what pow(2,3) does. (You might not heave read of math.h yet, but you still can use it.)

NevemTeve, I am hoping that the author of the book was thoughtful enough to pair the concepts he's teaching in a chapter to the exercises he is having me solve, so I am trying to solve it without resorting to tools I "shouldn't" have yet.

pow works on doubles, therefore it cannot be replaced with a simple for cycle. Actually it works with the taylor series of natural logarithm and exponential function (or something similar). That will allow to calculate it with the basic mathematical operations in a few steps.

@OP:
if your book wants you to calculate x^2 and x^3 via successive multiplication, then just do that without any non-standard compiler-dependent trick.

Oh yes it can be replaced by a for loop. It doesn't matter what you multiply, the math still works. Taking something to a power is just multiplying it by itself that many times. The type doesn't matter, except for complex numbers.

On the original post. Great that you're exploring some code, I do salute that. I wonder the purpose of the code, if it's for numerical study or for code learning. If for code learning, then it is rather odd and a bit confusing.

The code compiles, gives a result. I have to admit that I've never represented any powers ever. That makes me wonder why this is important? ntubski's answer explains why that horrific representation turns out to be 20.3999023438 ... I for one would never use that, I'd use 20.3999023438!

Maybe this is a precision thing. For instance even if the preprocessor accepted such a variable, would it then truncate it or something? I printed x out using %2.20g for my format and got 20.39990234375. Next I created a constant named "z = 20.39990234375" and printed that out, same result.

Personally I'd rather use a constant number which is readable to me versus learning a notation, especially now that I have an example that doesn't prove it help in any way.

As far as that programming example, if it is for programming, then would it also not be better in the printf() statement to use the variables a, b, x, and y and illustrate how to format the output for the prints versus the way that printf() statement is now?

A loop will only work for an integer exponent, pow does not have that restriction.

Just because it doesn't make sense in THAT example doesn't mean it never makes sense. Try doing some physics-based programming or unit conversion without using exponential notation, it's a typo nightmare.

I find
much more difficult to type or read than
Counting zeros is a royal PITA

1 members found this post helpful.

Agreed and I appreciate the follow-up. I also feel that what you wrote makes sense. You write and I get that it means .

I had to go to Calculator to determine thatwhich means as the original coder at some point, I had to use the number 2.549987792969. Not typing it a ton is why we have #define statements.

Apples vs. Oranges I guess. I'm a proponent of clear versus unclear; and I guess my "clear" filter only goes as far as representing fractional numbers in their decimal form. Unless for some reason that number represents bit fields, mapped to something. However if that were the case, then I don't believe it would be a fraction.