pow() behaviour different with Float and Double

Question

I'm facing a weird issue with the pow(). Please check the following function

 let r = 8.5/1200
 let n = ((5*12.0)+1)

 let pp = (pow(Float(1 + r), Float(n)))
 debugPrint(pp) // 1.53811562

 let pq = (pow((1 + r), n))
 debugPrint(pq) // 1.5381189771003985

Here pow(Float, Float) function return max. 8 fractional number for Float and pow(Double , Double) returns value with 16 fractional number for Double.

Why it is so could anyone explain it?

And is it possible to get upto 16 fractional number for Float value?

Note: You can copy and paste it in playground to see result.

Thanks in advance.

Answer 1

Float uses an IEEE 754 four byte single precision number. 24 bits are assigned to the significand and 8 bits to the exponent. With 24 bits of precision, it can represent a little over 7 decimal digits of precision

Double - which is the default floating point format for Swift uses an IEEE double precision eight byte number. The significand uses 53 bits and the exponent 11 bits. With 53 bits of precision, it can represent just under 16 decimal digits of precision.

The reason why pq is printed with 16 digits is that by default the compiler has inferred the argument type to be Double.

And is it possible to get upto 16 fractional number for Float value?

No. There aren't enough bits of precision.