login
a(n) is the largest power of n such that all numbers n^k <= a(n), k=1,..,A336774(n)-1 can be exactly represented as single precision 32-bit floating point numbers according to the IEEE 754 standard.
8

%I #11 Aug 06 2020 03:57:15

%S 170141183460469231731687303715884105728,14348907,

%T 85070591730234615865843651857942052864,9765625,470184984576,5764801,

%U 85070591730234615865843651857942052864,4782969,10000000000,1771561,15407021574586368,4826809,1475789056,11390625,21267647932558653966460912964485513216

%N a(n) is the largest power of n such that all numbers n^k <= a(n), k=1,..,A336774(n)-1 can be exactly represented as single precision 32-bit floating point numbers according to the IEEE 754 standard.

%H Hugo Pfoertner, <a href="/A336775/b336775.txt">Table of n, a(n) for n = 2..16384</a>

%F a(n) = n^(A336774(n)-1).

%e a(3) = 14348907 = 3^15, because the next power 3^16 = 43046721 cannot be exactly represented as a binary32 floating point number, but only rounded to 43046720.

%Y Cf. A336774, A336778, A336779.

%K nonn,fini

%O 2,1

%A _Hugo Pfoertner_, Aug 04 2020