Lower bounds for the depth of modular squaring

Benjamin Wesolowski; Ryan Williams

Paper 2020/1461

Lower bounds for the depth of modular squaring

Benjamin Wesolowski and Ryan Williams

Abstract

The modular squaring operation has attracted significant attention due to its potential in constructing cryptographic time-lock puzzles and verifiable delay functions. In such applications, it is important to understand precisely how quickly a modular squaring operation can be computed, even in parallel on dedicated hardware. We use tools from circuit complexity and number theory to prove concrete numerical lower bounds for squaring on a parallel machine, yielding nontrivial results for practical input bitlengths. For example, for , we prove that every logic circuit (over AND, OR, NAND, NOR gates of fan-in two) computing modular squaring on all -bit inputs (and any modulus that is at least ) requires depth (critical path length) at least . By a careful analysis of certain exponential Gauss sums related to the low-order bit of modular squaring, we also extend our results to the average case. For example, our results imply that every logic circuit (over any fan-in two basis) computing modular squaring on at least of all -bit inputs (for any RSA modulus that is at least ) requires depth at least .

Metadata

Available format(s): PDF
Category: Applications
Publication info: Preprint. MINOR revision.
Keywords: Verifiable delay function circuit modular squaring RSA
Contact author(s): benjamin wesolowski @ math u-bordeaux fr
History: 2020-11-19: received
Short URL: https://ia.cr/2020/1461
License: CC BY

BibTeX

@misc{cryptoeprint:2020/1461,
      author = {Benjamin Wesolowski and Ryan Williams},
      title = {Lower bounds for the depth of modular squaring},
      howpublished = {Cryptology {ePrint} Archive, Paper 2020/1461},
      year = {2020},
      url = {https://eprint.iacr.org/2020/1461}
}