Should I use floating-point or integer math for sensor values.

time more specific.


my actual code on 2000 lines of object-oriented goodness, i'll give simplified version of i'm doing.

code: [select]


uint64_t accumulator; //the question: should use large integer this, or can use float/double?
uint32_t numsamples;
float actualvalue;

void accumulatevalues()
{
accumulator += readtcvalue()*1000; //this value floating-point, multiply 1000 keep 3 decimal places.
numsamples++
}

void applyvalues()
{
actualvalue = accumulator / numsamples / 1000; //divide 1000 out multiplied earlier
numsamples = 0;
accumulator = 0;
}

void loop()
{
currenttime = millis();

while (millis() - currenttime < 10000) // loops fast can until 10s has passed
{
accumulatevalues(); //accumulates sensor value samples
}

applyvalues(); //grabs average of said samples , resets accumulator , sample counter
serial.println(actualvalue);
}


so right accumulate values integer value. i'm wondering if float/double able work in scenario. way, don't have multiply 1000 , divide out later. concern addition screwy on value exponent keeps changing. like, happens if add 2 floating-point value of 1.00000000000x10^15, add if doesn't register in variable? or data lost?

sorry if bit of roughly-asked question.

no idea, beauty of software development can test

Arduino Forum > Using Arduino > Programming Questions > Should I use floating-point or integer math for sensor values.

arduino