Well, that expression is computed a several hundreds of thousands of times / second, with both u, w and V varying. But note that I won't be calculating that product using any for loops, but direct substitution.
In the meantime, got a couple of more question related to the internal functions of the processor.
1) How long does it actually take to allocate memory for a variable type double? This is related to pondering if it is worthwile to pre-compute a simple expression like t_tup2=3*u*u beforehand, if the result only gets used twice. If memory allocation takes those precious CPU cycles, it might be a better idea to simply compute the product twice on the fly rather than substitute the precomputed value.
2) Consider expression n*x, where x is a real number and n is a integer number. Given that summation is usually faster than product, how many times one can sum the same variable until summation becomes more expensive than computing the product directly?
I know I could probably get the answers looking at the processor instruction timing tables, the problem is that those numbers are theoretical, and do not account the operating system running things in the background. And when it comes to operating systems, that is where my computer literacy actually ends.
Mika