Are multiple sequential array access optimised?

Will C++ compilers automatically optimise the following code to calculate 3*i only once, and increment the result, or must the programmer code this?

void writeToArr开发者_开发百科ay(int i, int a, int b, int c)
{
   array[3*i][1]=a;
   array[3*i][2]=b;
   array[3*i][3]=c;
}

---

Most optimizers will pick up on this, yes. However if you want to encourage the optimizer, here's how:

void writeToArray(int i, int a, int b, int c)
{
   const int offset = 3 * i;
   array[offset][1]=a;
   array[offset][2]=b;
   array[offset][3]=b;
}

---

When you enable optimization, almost all compilers will not only precompute the common index 3*i but the pointer array + (3*i)*sizeof (*array).

---

Common Subexpression Elimination is one of the easiest and most commonly applied optimizations. In general, you can rely on this being done.

---

With modern compilers you can most of the time rely on the compiler doing the clever job and not try anything yourself. See for instance http://dl.fefe.de/optimizer-isec.pdf