Hi ,
Feeling a bit stupid asking this but could any one tell me an efficient way to modulo decrement. I know how to modulo increment for powers of two i.e -
if i want to modulo 16 increment a char i would do something like -
#define MODULO_SIZE 16
int modulo_increment(int isequence) { //increment isequence++; //map to current size isequence &= (MODULO_SIZE-1); return(isequence); }
but how do I efficiently modulo decrement
Thanks in advance
rate
I've got a feeling that simple:
isequence++; if (MODULO_SIZE == isequence) { isequence = 0; }
will generate the same (or even better) code with most modern compilers. In other words, the compiler will do such tricks for you.
Well, replacing ++ with --, the same idea will work (at least in cases where MODULO_SIZE is a positive integral power of 2, as the original).
Best regards, Spehro Pefhany
It certainly shouldn't generate the same code, because you're asking for something rather different.
Consider the case where you call the function with isequence == MODULO_SIZE. If the ints happen to be 32 bits, it might take quite a few years to get back into sequence.
Best regards, Spehro Pefhany
:-)
It was not my intention to demonstrate the equivalent code. (But well spot on, anyway). Just the word 'efficient' together with a trick of '& (MODULO_SIZE-1)' caught my eye. And my message was: this particular trick is probably no more efficient than other, more straightforward, implementation. That's all.
Indeed. Note that you need to use an unsigned variable for this to be as efficient as the &, and for non-powers of 2 Sphero's sequence is best. The most generic and yet efficient solution allows both variable increments and maximum:
unsigned i, inc, max; for (i %= max, inc %= max; ... ; (i += inc) >= max ? i - max : i) // loop using i
Note i, inc and max can be changed in the loop as long as i < max and inc < max. To answer the OP's question, set inc = max-1.
Wilco
Have something to add? Share your thoughts — no account required.
Ask the community — no account required