Exact mismatch between Fortran and Python (sin function)

Question

I see a discrepancy between python and Fortran when using the sinus function. Can anyone shed some light on this please?

in python:

      import math
      print(math.sin(6.28318530717959))
      >> 3.3077843189710302e-15

in fortran90:

      print*, sin(6.28318530717959d0)
      >> 3.3077720792452914E-15

EDIT: Since this is a Fortran compiler issue, I used g95 with

       g95 -O3 test.f90 -o test.exe

+1

Christian 25 Feb 17 at 22:29

1 answer

BM · Accepted Answer · 2017-02-25T23:58:05+0000

According to IEEE 754 for floating point representation:

In [7]: bin(3.3077720792452914e-15.view(np.uint64))
Out[7]: '0b11110011101101110010110011010000000000000000000000000000000000'

shows a truncated mantisa when

In [9]: bin(3.3077843189710302e-15.view(np.uint64))
Out[9]: '0b11110011101101110010110011101100111001100111010111010001111111'

shows simple.

Probably a type issue, with float32 in progress, even the origin is mysterious.