A lower bound for integer greatest common divisor computations
- 1 April 1991
- journal article
- Published by Association for Computing Machinery (ACM) in Journal of the ACM
- Vol. 38 (2) , 453-471
- https://doi.org/10.1145/103516.103522
Abstract
It is proved that no finite computation tree with operations { +, -, *, /, mod, < } can decide whether the greatest common divisor (gcd) of a and b is one, for all pairs of integers a and b . This settles a problem posed by Gro¨tschel et al. Moreover, if the constants explicitly involved in any operation performed in the tree are restricted to be “0” and “1” (and any other constant must be computed), then we prove an Ω(log log n ) lower bound on the depth of any computation tree with operations { +, -, *, /, mod, < } that decides whether the gcd of a and b is one, for all pairs of n -bit integers a and b . A novel technique for handling the truncation operation is implicit in the proof of this lower bound. In a companion paper, other lower bounds for a large class of problems are proved using a similar technique.Keywords
This publication has 12 references indexed in Scilit:
- On computations with integer divisionRAIRO - Theoretical Informatics and Applications, 1989
- Lower bounds for computations with the floor operationPublished by Springer Nature ,1989
- On the limits of computations with the floor functionInformation and Computation, 1988
- Applications of Ramsey's theorem to decision tree complexityJournal of the ACM, 1985
- On the control power of integer divisionTheoretical Computer Science, 1983
- Division in idealized unit cost RAMsJournal of Computer and System Sciences, 1981
- Factoring numbers in O(logn) arithmetic stepsInformation Processing Letters, 1979
- A characterization of the power of vector machinesJournal of Computer and System Sciences, 1976
- The Computing Time of the Euclidean AlgorithmSIAM Journal on Computing, 1974
- Berechnung und programm. IActa Informatica, 1972