A309541 Representable positive integers n that are not the inverse of their inverse in binary64 (double precision) IEEE 754 floating-point arithmetic (Version where 1 and n*(1/n) are unequal).

49, 98, 103, 107, 161, 187, 196, 197, 206, 214, 237, 239, 249, 253, 322, 347, 374, 389, 392, 394, 412, 417, 425, 428, 443, 474, 478, 479, 491, 498, 499, 501, 503, 506, 509, 561, 569, 644, 685, 691, 694, 725, 729, 735, 737, 748, 753, 765, 778, 779, 784, 788, 789, 797, 809, 817, 823, 824, 829, 833, 834, 837, 841, 849, 850

Offset: 1

Author

Adam M. Scherlis, Aug 06 2019

Keywords

• nonn
• fini

Links

• Andrey Zabolotskiy, Table of n, a(n) for n = 1..10000
• Efstratios Gallopoulos, Scientific Computation I, five terms provided (in Greek) [broken link]
• Mark, comment on blog post "Rundungsfehler im Flash Player?", eight terms (in German)
• koDoz, comment in thread "Welche Grenzwerte gelten für Zahlen, Strings usw?", nine terms (in German)

Crossrefs

See A275419 for the n != 1/(1/n) version.

Programs

Fortran

doubleprecision one, r
      integer i
      parameter (one=1.0D0)
      do 10 i = 1, 500
      R = one / dble(i)
      if ( R * dble(i) .ne. one) write (*,1000) i
1000  format (i0)
10    continue
      end
C Hugo Pfoertner, Jan 18 2024