A Table of Nonnegative Powers of Two

4 comments

1. [quote]
   unsigned long long po2 = 9223372036854775808ULL; //2^63
   [/quote]
   Please don’t ever do that.

   Use the left shift operator (<<) instead, like
   unsigned long long po2 = 1ULL<<63;

2. @lallehat,

   My method assigns a decimal power of two from the table directly to a variable. Your method assigns a power of two indirectly through knowledge of the underlying binary representation (two’s complement) and bit shifting operations. Both methods work: mine uses the “2ⁿ” column, and yours uses the “n” column. I would think it’s just a matter of personal preference as to which method you choose.

   As for performance, there’s no difference. Both should be resolved at compile time. For example, under Visual C++, both resolve to these two assembler statements:

   00419B4E  mov         dword ptr [po2],0  
   00419B55  mov         dword ptr [ebp-8],80000000h
   

3. [quote] Your method assigns a power of two indirectly through knowledge of the underlying binary representation (two’s complement)[/quote]
   No, the << operator (as most things in the C standard) is defined with what it does to the numeric values of its operands, diregarding the actual implementation.

   Of course, as always with arithmetic in C, if you deal with unsigned integers the "overflow" is perfectly defined as modulo "TYPE_MAX", and if you deal with signed integers, you'll get an undefined behavior on overflow. The degree of undefinedness depends on the implementation, of course, but that's another story 🙂

4. @al,

   I see your point, but what I meant was, to write “1ULL<<63”, you have to know that 2⁶³ is a 1 followed by 63 0s.

Comments are closed.