Does anyone know of or have a square root procedure (mathematical or assembly). So far I've only found long approximation methods but maybe this is really the best way of doing it? In that case how do they get so accurate with say sqrt function in high level languages (my own thought but feel free to answer if you care to).

I'm messing around trying to calculate something as practice and learning (HP and Torque due to drag) but I need to raise to power of 2.5, which can be done by square root of 5th power, but I need that sqaure root model.

 

All methods of calculating irrational roots are approximate, but can be improved through iteration to any precision desired.

Do you want an implementation in floating point? fixed point? integer? what format?

What's the equation you're trying to use?

 

Rolling Resistance, cell C13.

 

Do you want floating, fixed or integer?

 

Apeiron, glad to see that you have this one. My level of math doesn't mesh that well with assembler.

Another thought would be to use a lookup table.

RBob.

 

I'd also use a lookup table, since speed is the only variable that needs a fractional root.

There are a number of interesting algorithms for calculating square roots to any desired precision though, depending on which format you're using. All of them are slow.

 

Thanks guys, didn't think of the lookup table, thats a good idea. I'll give it a go.

I didn't think about implementing the equation until I was actually writing the code. I think I got too far ahead of myself. BTW, I'm doing it the hard way...with a hex editor, since I don't have a source with the wideband patch, let alone all my values transferred. I guess it's not really that much harder, just have to watch the addressing. Good learning though

Thanks again.