OFFSET
1
COMMENTS
The signed binary nonadjacent form is also called "canonical signed digit representation" or the result of a "canonical recoding" algorithm.
It is used to multiply binary numbers by the minimum number of shifts and additions or subtractions.
Multiplication by 1/Pi or, more generally, by 2^k/Pi, can be used in argument reduction in the evaluation of trigonometric functions.
REFERENCES
I. Koren, Computer Arithmetic Algorithms, 2nd edition, page 146.
LINKS
Thomas König, Program in C to calculate the digits using gmp and mpfr
H. Prodinger, On binary representations of integers with digits -1,0,1, Integers 0 (2000), #A08.
Wikipedia, Canonical signed digit
Wikipedia, Non-adjacent form
EXAMPLE
0.01010010T0000T010T000010...
CROSSREFS
KEYWORD
sign,easy,base
AUTHOR
Thomas König, Jul 11 2021
STATUS
approved