Unsigned Long Long - weird output

Question

OK, I'm trying to print the decimal/hexadecimal version of a relatively big unsigned long long and the results I'm getting are quite weird...

The code :

unsigned long long a = 1llu<<63;

printf("decimal = %llu\n",a);
printf("hexadecimal = %llx\n",a);

The output :

decimal = 9223372036854775808
hexadecimal = 8000000000000000

Now, here's what :

Why's that happening? What am I doing wrong???

Possible duplicate? Interesting though: http://stackoverflow.com/questions/2844/how-do-you-printf-an-unsigned-long-long-int — abelito, Dec 13 '12 at 04:14
@chris Really??? I don't think it even fits a 64-bit integer, does it? And if so, why? — Dr.Kameleon, Dec 13 '12 at 04:14
@Dr.Kameleon, An unsigned 64-bit integer can hold up to 2^64 - 1. This is 2^63. — chris, Dec 13 '12 at 04:15

score 8 · Accepted Answer · answered Dec 13 '12 at 04:14

8

Since five is not a factor of any power of two, no power of two ends in zero. Your other source which gave 9223372036854780000 is incorrect.

answered Dec 13 '12 at 04:14

Potatoswatter

1

My other source is Mac OS X's Programmer's calculator. But, yep, your mathematical point of view definitely makes sense. – Dr.Kameleon Dec 13 '12 at 04:15
1

@Dr.Kameleon Punching "1 << 63" gives a number ending in 808. Using the scientific calculator's 2^x button for 2^63 then switching to programmer mode results in 9,223,372,036,854,776,000. Not sure exactly what you did. – Potatoswatter Dec 13 '12 at 04:17
I'm really confused. Just switched to `Programmer` view. `1` `X< – Dr.Kameleon Dec 13 '12 at 04:23
4

Congratulations, you have found a bug in the Mac calculator. – brian beuning Dec 13 '12 at 04:25
1

Mac calculator is probably computing with double precision floating point math, and using a low-quality print-to-decimal algorithm that cannot produce the exact value (note that the exact value **is** representable in IEEE double). – R.. GitHub STOP HELPING ICE Dec 13 '12 at 04:59

1 Answers1