Critte 0.90 released

[Milos]

With 1300 positions randomness is definitely insignificant!

All positions measure different things so it's not just pure averaging and variance is much larger sigma_single_position/sqrt(n). You can test it yourself. Run full STS 10 times on a single core, write down results and calculate relative sigma. Then run full STS 10 times on quad, write down results and calculate relative sigma. You will see a huge difference.