Take the 2-minute tour ×
Programmers Stack Exchange is a question and answer site for professional programmers interested in conceptual questions about software development. It's 100% free, no registration required.

i'm trying to understand a constant 0x9e3779b9 what kind of data is this? it's not binary, not decimal, what is this?

it's a constant used on the TEA algorithm it says it's derived from the golden number but the golden number is 1.618?

share|improve this question
    
Do you mean the number format or how did they come up with the value? –  Matthew Whited Mar 30 '11 at 13:17
    
what about the format? i've never seen it... i read that the value is from (1 + √5)/(2^31) but after i try it the result is 2654435769.4972302964775847707926? –  pandr01d Mar 30 '11 at 13:19
11  
In most C style languages 0x is a prefix for a hexadecimal number. I'm wondering if that is the hex format for the floating point representation of the golden number. Of course that would beg the question of why store an FP value as an integer? Perhaps they optimized the implementation of the TEA algorithm to work with integers instead of floating point values? –  Berin Loritsch Mar 30 '11 at 13:21
    
It is en.wikipedia.org/wiki/Hexadecimal format. –  TZHX Mar 30 '11 at 13:22
    
i think they store it as integer because later it would be multiplied by some other number –  pandr01d Mar 30 '11 at 13:28

2 Answers 2

up vote 7 down vote accepted

As others have said, the constant is an integer in hexadecimal form. Specifically, it is a 32-bit integer in hexadecimal form. If the constant is a signed integer, then 0x9e3779b9 is negative 1640531527 decimal in two's complement form; therefore, it may be a scaled integer fraction that has been adjusted to deal with non-integral of 2 related problems.

Two's complement negative to positive conversion in hex

0x9e3779b9 ⊕ 0xffffffff + 0x00000001 = 0x61C88647 = 1640531527 in decimal

or using the 1's complement operator ~ in the C family of languages

~0x9e3779b9 + 0x00000001 = 0x61C88647 = 1640531527 in decimal 

Two's complement negative to positive conversion in binary

10011110001101110111100110111001 ⊕ 11111111111111111111111111111111 + 00000000000000000000000000000001 =  01100001110010001000011001000111 = 1640531527 in decimal
share|improve this answer

I think this StackOverflow question answers it:

http://stackoverflow.com/questions/4948780/magic-numbers-in-boosthash-combine

Essentially, it is a "golden number" for hash functions, and it being an integer is quicker to calculate.

The notation 0x is for a hexadecimal number, or base 16. The benefit of a base 16 number is that each pair of digits represents one byte exactly. Given a little practice, you can almost see the bit pattern in your mind, assuming you've ever worked with binary numbers.

share|improve this answer
    
I had just found something similar +1 –  Matt Ellen Mar 30 '11 at 13:31
    
okay i guess i have to understand the hash function first –  pandr01d Mar 30 '11 at 13:45
    
@Berin - is there a reason for this particular number? You presumably just need an irregular set of 0 and 1. Is the golden ratio link just arbitrary or is there some deeper mathematical logic? –  Martin Beckett Mar 30 '11 at 20:08
    
People with better understanding of hash algorithms would be better suited for that answer. My understanding of this "magic number" is that it is mathematically random, but that's where it ends. –  Berin Loritsch Mar 30 '11 at 22:53
    
In the link I provided, the selected answer says that the number is derived from the golden ratio. The answer also has the math that created the answer. –  Berin Loritsch Mar 30 '11 at 23:39

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.