On 19 June 2015 at 16:20, Scott Jones <[email protected]> wrote:

> Not being an astronomer, I don't know how this is supposed to be, but you
> have calculations based on the bod array inside the loop, but you are never
> updating bod...  Are you just depending on the compiler to move those out
> of the loop?
>


Oh, the program is not complete. In a real program the "bod" array would be
updated elsewhere. Updating the bod requires O(n) additions,
multiplications and memory look-ups. Computing the forces on each particle
requires O(n^2) additions, multiplications, divisions and memory look-ups.
So all attempts at optimization focus on the O(n^2) part, so that's the
only part that I posted.

Cheers,
Daniel

Reply via email to