1
I came up with the following in GNU C
r=0;m(a,b){x:goto*(int*[]){&&x,&&y}[r+=a&1?0:b,a>>=1,b<<=1,a<1];y:;}
One could argue there is a branch here a&1?0:b
So I also came up with this one
r=0;m(a,b){x:goto*(int*[]){&&x,&&y}[r+=(int[]){0,b}[a&1],a>>=1,b<<=1,a<1];y:;}
I'm depending on implicit int here as well.
Can you make it any smaller?
Explanation of the above code, or at least parts of how I got there can be found here http://codepad.org/L0MvnCe9
Must not use function calls (recursion, standard library functions like fma, exp, etc), loops (while, for, do), or branches (if, ?:, etc). No using the divide or multiply operator either (that's cheating.) goto is allowed for branching.
Good luck!
graphitemaster
Posted 2014-11-10T08:46:41.480
Reputation: 119
14
gotois a branch. – feersum – 2014-11-10T09:03:00.363goto is a form of branch, yes. I was meaning branch in the more high level form where by if/?: though. – graphitemaster – 2014-11-10T09:09:00.110
1It's just an obfuscated form of
if(a<1)goto x;. If an exact list of operators, keywords etc. which can't be used is provided it could be a valid code golf problem though. – feersum – 2014-11-10T09:12:09.037It's literally impossible to do this without ANY sort of branch, you know. – Tal – 2014-11-10T10:10:34.697
2@Tal It's simple but tedious to do it without any branch, as long as the size of
intis known, by calculating each of the n bits from the n^2 interactions between bits. – feersum – 2014-11-10T10:39:27.2203Is C the only allowed language? – ProgramFOX – 2014-11-10T10:46:17.123
1Could you rearrange the information a bit, as it took me a while to see what the rules are. Also, how does the code pad link give explanation? Since it's quite small, you could move it into the question as well. – matsjoyce – 2014-11-10T19:23:22.873