Hopefully, we will have posits with hardwaresupport sooner than later. https://spectrum.ieee.org/floating-point-numbers-posits-processor https://spectrum.ieee.

Hopefully, we will have posits with hardwaresupport sooner than later. https://spectrum.ieee.org/floating-point-numbers-posits-processor https://spectrum.ieee.org/floating-point-numbers-posits-proc... https://www.cs.cornell.edu/courses/cs6120/2019fa/blog/posits/ https://www.cs.cornell.edu/courses/cs6120/2019fa/blog/posits... Unfortunately, they are not a solution to the OPs problem, which is fundamentally embedded in the architecture of computers. One has to find an appropriate representation for the needed numbers in bits and bytes. In Python one can use fractions or decimals, if the float format is not good enough. Other options are fixed point arithmetic or arbitrary precision arithmetic. Choose one that combines the needed characteristics with the least amount of work.