Has anyone heard about contaminated digits, and how the equation log n represents the number of contaminated digits. Where n represents the value when a floating point number is in error of n ulps, or if the relative error in a computation is n. Where is the upper bound.
I just don't understand how the equation log n gives the value of the contaminated digits.
Hoping snowtea comes to my rescue again. If more explanation is needed please ask in other to help me with the above problem. I understand what ulps means and how to calculate it, I understand what relative error means and how to calculate it, what I just don't get is how the equation log n gives me the value of contaminated digits.
Thank you.