0

I need to implement arbitrary precision floating point routines in OpenCL and I'd like to be able to compare results with mpmath.

For basic operations +,-,*,/ and basic functions exp, log, sqrt etc: does it follow eg IEEE 754 ? And if not is there any other convention/standard/reference that is followed ?

Jim
  • 135
  • 4
  • 1
    Have you consulted the [mpmath](http://mpmath.org/) (1) documentation (2) source code? – njuffa Jan 31 '20 at 02:57
  • I couldn't find something in the documentation. I haven't looked at the source code, it's a number of operations and it would be hard to pin down the answer unless it's explicitly mentioned ... – Jim Jan 31 '20 at 06:28

0 Answers0