Posit floating point numbers: thin triangles and other tricks (2019)

Re:Posit floating point numbers: thin triangles and other tr... by phkahler on Thursday June 19, 2025 @04:02PM (#44322043)

My enthusiasm for posits was never about the precision. It was due to not having signed zero, and not having NaN. Just give me the nearest representable value please. Operations on numbers should never yield results that are not numbers.

[ Parent | on HN ]

Re:Posit floating point numbers: thin triangles and other tr... by thfuran on Thursday June 19, 2025 @04:03PM (#44322052)

What should happen if you divide 1 by 0?

[ Parent | on HN ]

Re:Posit floating point numbers: thin triangles and other tr... by dnautics on Thursday June 19, 2025 @04:56PM (#44322472)

inf

[ Parent | on HN ]

Re:Posit floating point numbers: thin triangles and other tr... by thfuran on Thursday June 19, 2025 @06:20PM (#44323057)

That's not really a number either. But what about 0/0?

[ Parent | on HN ]

Re:Posit floating point numbers: thin triangles and other tr... by kbolino on Thursday June 19, 2025 @08:23PM (#44323755)

It seems that posits have a value called NaR ("not a real") which serves the purpose of all the special values (including both infinities) in IEEE floats. There are some key differences though:

  NaR (posits)                 NaN (floats)
  ------------                 ------------
  one bit pattern              many different bit patterns
  always equal to itself       never equals itself
  less than all other values   falsifies all orderings

[ Parent | on HN ]

Re:Posit floating point numbers: thin triangles and other tr... by dnautics on Monday June 23, 2025 @05:35PM (#44360401)

It's a mode you can turn on that substitutes inf for NaR.

[ Parent | on HN ]

Re:Posit floating point numbers: thin triangles and other tr... by spc476 on Thursday June 19, 2025 @05:40PM (#44322762)

IEEE does define x/0 to be infinity and -x/0 as -infinity. But what about 0/0? Or sqrt(-1)?

[ Parent | on HN ]