(42S) matrix permanent

[Albert Chan]

Hi, Werner

Perhaps you can add some comments to code?

Example, why does it use SQRT ?
Does it have something to do with j = bit.ffs(k) ?

With decimal machine, getting bit.ffs (Find First Set) is expensive.
My guess is it would cost way more than these lines (j = f[1]) for next loop.

f[1] = 1
f[i] = f[i+1]
f[i+1] = i+1