Why does the compiler hate this statement?

accel = sqrt(pow(Ax,2)+pow(Ay,2)+pow(Az,2))/1333.0;

In particular the 'pow'...  Am I missing somethign in the manual? The compiler won't do it..  I ran into this once before, about a week ago..  I just looked at http://www.keil.com/support/docs/2518.htm

and http://www.cplusplus.com/reference/clibrary/cmath/pow/

I mean... I know it's almost 4am, but..  can someone enlighten me?  I'm mostly curious, and probably totally missing something obvious....

I copied and paste your line in one of my program and it compiled with no error.

Thanks for the reply!  Interesting...  Perhaps I will revisit after some sleep then... ;)

BTW, using pow() for small integer powers is really inefficient... Ax*Ax will likely be much faster.

Igor Skochinsky wrote:

BTW, using pow() for small integer powers is really inefficient... Ax*Ax will likely be much faster.

Thanks for the input!  That was actually a grab from some code I found on the web that I'm trying to get to work on the mbed for th SCA3000-D01 accelerometer...  Not with much luck as of yet I might add... =/ 

I do appreciate the help though!  I will certainly keep that in mind.. =) 

why doesnt this work:

float temp;

float ppmm;

temp = (5 - voltage)/voltage;

ppmm = 103*(pow(temp, -1.495));

I thought pow works with floating point....

Errors

"More than one instance of overloaded function "pow" matches the argument list: (E308)" in file "/main.cpp"

"            function "std::pow(float, int)" (E0)" in file "/main.cpp"

Hi Gowtham,

Try -1.495f (add an f); that'll tell the compiler to use the pow(float, float).

Currently, it doesn't know whether is should use pow(float, float) or pow(int, int).

Simon

A note on the pow() function:

It's really, really terrible for small integer bases. Even if the exponent is a variable, it's probably better just to multiply it out. I just found a case where I was (stupidly) using it to create a bitmask - it took 31 uSec to execute (and that was with the {float, float} version - the {long double, int} version took 61 uSec!). By comparison, generating the mask manually with a for loop took 450ns.

Sure, this is a simple example, since I'm only doing some shift operations. But I figured someone else might also benefit from realizing just how incredibly inefficient pow() can be, since this is of special concern in the world of embedded development.

Also, as far a I know the cmath function pow, will not work with int, int arguments.