SimpleIDE ver 0.9.27 float problem

Iamretired

I am using the tutorial to learn C,
I have tried to calc a ratio and get no result. I have a similar problem with mod,

#include "simpletools.h"
#include "math.h"
int main()
{
pause(1000);
float x = 3;
float y = 4;
float ratio;
ratio = y / x;
printf("ratio is %f\n", ratio);
}

On the screen I get ratio is with no calc result.

Johnmb

Dave Hein

simple printf doesn't support floating point. Uncheck the simple printf option and it should work.

SRLM

You need to link with the -lm option. Here is the code that I used (I don't have simpletools):

#include <math.h> 
#include <stdio.h>
int main() 
{
	float x = 3;
	float y = 4;
	float ratio;
	ratio = y / x;
	printf("ratio is %f\n", ratio); 
}

user@desktop:~/Downloads$ propeller-elf-gcc -o float_main.elf float_main.c -lm
user@desktop:~/Downloads$ propeller-load -r -t float_main.elf 
Propeller Version 1 on /dev/ttyUSB0
Loading float_main.elf to hub memory
21976 bytes sent                  
Verifying RAM ... OK
[ Entering terminal mode. Type ESC or Control-C to exit. ]
ratio is 1.333333

This thread has more information: http://forums.parallax.com/showthread.php/136586-did-I-miss-a-note-on-floats-in-printf

Iamretired

Where is the "simple printf option" ?

Johnmb

dgately

Iamretired wrote: »

Where is the "simple printf option" ?

Johnmb

It is a Compiler option, but may actually not be hooked-up at this time and may be ignored. But, yes the Linker option "-lm" should work, no matter.

A Mac OS X view of the Compiler options... Looks almost the same on Windows.

dgately

Iamretired

I tried several ways including the tutorial exercise to calc volume.

This program works.

#include "simpletools.h"

int main() // Main function
{
pause(1000); // Wait 1 s for host computer
float r = 1;
float x = 4;
float y = 3;
float ratio = x / y;
float v = ratio* PI * r * r * r;
printf("ratio = %f \n", ratio);
printf("volume = %f \n", v); // Display volume
}

ratio = 1.333333
volume = 4.188790

The following form doesn’t work and I don’t know why.

#include "simpletools.h" // Include simpletools

int main() // Main function
{
pause(1000); // Wait 1 s for host computer
float r = 1;
float ratio = 4/3;
float v = ratio* PI * r * r * r; // replaced ratio with 4/3 and result same
printf("ratio = %f \n", ratio);
printf("volume = %f \n", v); // Display volume
}

ratio = 1.000000
volume = 3.141593

kuroneko

4/3 is seen as an integer expression and treated as such. Try 4.0/3.0 or 4.0f/3.0f (the former is double IIRC).

Andy Lindsay (Parallax)

Here is a tutorial page on floating point:

__learn.parallax.com/propeller-c-start-simple/floating-point-math

It shows how to check the Math Lib box in Show Project Manager -> Linker Tab -> Math Lib.

Also, to make sure it wasn't something else, I saved a copy of the example program from:

\Documents\SimpleIDE\Learn\Examples\C Intro\Basics\Floating Point Calculations.side

...to \Documents\SimpleIDE\My Projects, then used Run with Terminal to see the results. It worked:




Whenever you use floating point, tell the IDE to link the math library by using the Math Lib check-box shown here. (Show Project Manager button is in bottom-left corner.)



We left it unchecked in the New Project Template so that programs with integer math would not carry the extra weight. (We might want to rethink that.)

Andy

David Betz

#include "math.h"

Shouldn't we be encouraging people to use angle brackets to include system header files like math.h?

jac_goudsmit

David Betz wrote: »

#include "math.h"

Shouldn't we be encouraging people to use angle brackets to include system header files like math.h?

Yes. And we should encourage them to use scaling and fractions instead of float/double, if you want my opinion. I've never encountered a situation where floating point (with all its pitfalls such as non-representable numbers, inaccuracies, comparing for equality etc) couldn't be avoided.

===Jac

Dave Hein

jac_goudsmit wrote: »

Yes. And we should encourage them to use scaling and fractions instead of float/double, if you want my opinion. I've never encountered a situation where floating point (with all its pitfalls such as non-representable numbers, inaccuracies, comparing for equality etc) couldn't be avoided.

I wrote an orbital simulator that uses double. It would be extremely cumbersome to implement with integers. Of course, it currently doesn't run on the Prop, but P2 may be able to run it.

Iamretired

I thank you all for such 'to the point' and helpful replies. I am enjoying the tutorial.

Johnmb