Simple float question

turbosupra

I'm trying to get a simple floatmath calculation working that I can baseline off of and build on ... what am I doing wrong here?

CON
_clkmode = xtal1 + pll16x
_xinfreq = 5_000_000

VAR
long numerator
long denominator
long answer

OBJ
debug : "FullDuplexSerial"
f : "FloatMath"

PUB main | freq
debug.Start(31,30,0,115200) 'start cog with serial driver
waitcnt((clkfreq * 3) + cnt) ' makes life easier when starting to debug with a 2 second pause to get the serial viewer active and receiving

debug.Tx(13)
Debug.Tx(13)
Debug.Tx(13)
Debug.Str(string(" */\*/\*/\*/\*/\*/\*/\* START */\*/\*/\*/\*/\*/\*/\* About to enter main pub")) ' for debug purposes
Debug.Tx(13)
Debug.Tx(13)
Debug.Tx(13)
numerator := f.ffloat(22)
denominator := f.ffloat(7)

repeat
f.ffloat(answer)
answer := f.fdiv(numerator,denominator)
debug.str(string("The answer equals "))
debug.dec(f.ffloat(answer))
debug.tx(13)
waitcnt((clkfreq) + cnt)

Thanks for reading! The attachment has the spin file and the necessary math/debug spin files.

kuroneko

Try using

debug.dec(f.[COLOR="Red"]fround[/COLOR](answer))

instead. ffloat converts int to float. As answer is a float you want to convert it to an int (fround) to be displayed by debug.dec().

turbosupra

Hi,

Thanks!

I would like it displayed as a decimal though, is that possible? Are you saying debug.dec is not able to display a decimal?

kuroneko

turbosupra wrote: »

I would like it displayed as a decimal though, is that possible? Are you saying debug.dec is not able to display a decimal?

debug.dec does display a decimal number (an int in fact). If you want this in the form of 3.14 then have a look at the FloatString object (available in the prop tool library). Something along the lines of

debug.str(fstring.FloatToString(answer))

should do the trick.

turbosupra

Thank you, that worked!

Now if I can just get the rest of the calcs to work with this I will be set

kuroneko wrote: »

debug.dec does display a decimal number (an int in fact). If you want this in the form of 3.14 then have a look at the FloatString object (available in the prop tool library). Something along the lines of

debug.str(fstring.FloatToString(answer))

should do the trick.

turbosupra

One more question if I could

numerator/denominator = 16.666667 (numerator is 1000, denominator is 60)
but
(((clkfreq/2)/66656) /36) should also = 1000, but instead equals 1.345247e-42, see the quoted lines below, what would cause the difference?

On serial out I see

The Parallax Serial Terminal Says wrote:

The answer equals 16.66667
1.345247e-42

CON
  _clkmode = xtal1 + pll16x
  _xinfreq = 5_000_000


  
VAR
  long numerator
  long denominator
  long answer
  long answertwo
  long halfperiod
 

OBJ
  debug : "FullDuplexSerial"
  f   : "FloatMath"
  fstring : "FloatString"
  

PUB main  | freq
  debug.Start(31,30,0,115200) 'start cog with serial driver
  waitcnt((clkfreq * 3) + cnt)          ' makes life easier when starting to debug with a 2 second pause to get the serial viewer active and receiving

  debug.Tx(13)
  Debug.Tx(13)     
  Debug.Tx(13)
  Debug.Str(string(" */\*/\*/\*/\*/\*/\*/\* START */\*/\*/\*/\*/\*/\*/\* About to enter main pub"))          ' for debug purposes
  Debug.Tx(13)
  Debug.Tx(13)     
  Debug.Tx(13)
  numerator := 1000
  denominator := 60
  halfperiod := 66656

  
  
  repeat
    'answer := f.ffloat(answer)
    answer := f.fdiv(numerator,denominator)
    debug.str(string("The answer equals "))
    debug.str(fstring.FloatToString(answer))
    'debug.dec(f.fround(answer))
    debug.tx(13)

    answertwo := ((((clkfreq / 2) / halfperiod)/  36) * 60) 'cklfreq/2/halfperiod/36 = 16.666667 .... 16.666667 * 60 = 1000
    debug.str(fstring.FloatToString(answertwo))
    debug.tx(13)
    waitcnt((clkfreq) + cnt)

kuroneko

turbosupra wrote: »

halfperiod := 66656

    answertwo := ((((clkfreq / 2) / halfperiod)/  36) * 60)
    debug.str(fstring.FloatToString(answertwo))

1. answertwo is an integer, so use debug.dec()
2. as you are using integer arithmetic you won't arrive at 1000

• 80_000_000/2 = 40_000_000
• 40_000_000/halfperiod = 600.096015
• 600/36 = 16.666667
• 16*60 = 960

Heater.

turbosupra,

Your calculation:

answertwo := ((((clkfreq / 2) / halsperiod)/ 36) *60)

is being done as normal signed integer. So you have an integer result of 1000 which you should print as

debug.dec(answertwo)

turbosupra

Thank you both,

I found that sometimes I have to use these objects

object fullduplexserial = debug
object floatstring = fstring
object floatmath = f

and then f.div(variable1, variable2)
or f.div(variable1, variable2.0) as it wanted a .0 to properly display and calculate

and then debug.str(fstring.floattostring(variablename))
or debug.dec(variablename)

There are also numbers too big, the float object cannot accept 40,000,000 but can accept 40,000 ? So some scaling might be necessary.