.Net Calculation Speed Test iterates through a large number of calculations and presents the relative speeds
of calculation in a graphical format.
It is a tool for the software developer to use when optimising or designing calculation steps, to be able
to identify, for a specific machine, the relative cost in time when using different calculation methods.
To allow the developer to go further than the basic set of calculations, a version is available with the
full source code, written in C#.
Consider, as an example, the green item marked P in the Math library tests. P is the marker for the calculation
a=Math.Pow(b,2.0); where a and b are double precision numbers. The calculation takes around 200nS (nS=nanoseconds).
The blue item marked x in Double precision math, takes around 3nS, and represents the code a=b*b;
This shows us that using the math library for squares takes approx 60 times as long as a double precision multiply,
yet I often see the code x=Math.Pow(y,2.0) in examples.
Another interesting point is the green item marked isi and the purple item marked ii in Function calls.
Some claim, falsely as it turns out, that a static call has more overhead than a function call within the same
class. isi represents i=Aclass.AFunction(int i); and ii represents i=AFunction(int i); All acronyms
are explained in a separate window which is opened within the application.
The above example shows how it is on my machine.
|