Table of Contents
Fetching ...

Collatz-Weyl Generators: High Quality and High Throughput Parameterized Pseudorandom Number Generators

Tomasz R. Działa

TL;DR

This work introduces Collatz-Weyl Generators (CWGs), a family of high-throughput, uniform PRNGs built from generalized Collatz mappings combined with Weyl sequences to avoid short cycles and guarantee long periods. The authors formalize a general scheme and provide concrete implementations (CWG64, CWG128-64, CWG128) along with initialization and parallel-stream capabilities, and they demonstrate strong statistical performance across standard test suites. They establish theoretical period bounds and practical throughput advantages, including cryptographic considerations and potential applications. Overall, CWGs offer a scalable, low-footprint approach for parallel random number generation with robust uniformity and independence properties, suitable for both standard and cryptographic contexts.

Abstract

We introduce the Collatz-Weyl Generators, a family of uniform pseudorandom number generators (PRNGs) which are based on generalized Collatz mappings, derived from the Collatz conjecture and Weyl sequences. The high-quality statistical properties of our generators is demonstrated by the fact that they pass stringent randomness tests used by the research and standardization community. The proposed Collatz-Weyl Generators have a number of important properties, including solid mathematical foundations, enablement of high throughput and low latency implementation, small code and/or ASIC size, enablement of producing multiple independent streams and potential of support of cryptographic applications.

Collatz-Weyl Generators: High Quality and High Throughput Parameterized Pseudorandom Number Generators

TL;DR

This work introduces Collatz-Weyl Generators (CWGs), a family of high-throughput, uniform PRNGs built from generalized Collatz mappings combined with Weyl sequences to avoid short cycles and guarantee long periods. The authors formalize a general scheme and provide concrete implementations (CWG64, CWG128-64, CWG128) along with initialization and parallel-stream capabilities, and they demonstrate strong statistical performance across standard test suites. They establish theoretical period bounds and practical throughput advantages, including cryptographic considerations and potential applications. Overall, CWGs offer a scalable, low-footprint approach for parallel random number generation with robust uniformity and independence properties, suitable for both standard and cryptographic contexts.

Abstract

We introduce the Collatz-Weyl Generators, a family of uniform pseudorandom number generators (PRNGs) which are based on generalized Collatz mappings, derived from the Collatz conjecture and Weyl sequences. The high-quality statistical properties of our generators is demonstrated by the fact that they pass stringent randomness tests used by the research and standardization community. The proposed Collatz-Weyl Generators have a number of important properties, including solid mathematical foundations, enablement of high throughput and low latency implementation, small code and/or ASIC size, enablement of producing multiple independent streams and potential of support of cryptographic applications.
Paper Structure (16 sections, 1 theorem, 14 equations, 16 figures, 3 tables)

This paper contains 16 sections, 1 theorem, 14 equations, 16 figures, 3 tables.

Key Result

Theorem 1

Let $x$ be a random but not necessarily uniform bit stream, and let $weyl$ be a uniform but not necessarily random bit stream. Then, $x \oplus weyl$ is a uniform bit stream.

Figures (16)

  • Figure 1: Block diagram of n-bit Collatz-Weyl Generator.
  • Figure 2: Implementation of the "generate” operation of the CWG64 generator in C++.
  • Figure 3: Implementation of the “generate” operation of the CWG128-64 generator in C++.
  • Figure 4: Implementation of the “generate” operation of the CWG128 generator in C++.
  • Figure 5: Implementation of the CWG64 initializer in C++.
  • ...and 11 more figures

Theorems & Definitions (4)

  • Definition 1
  • Definition 2
  • Theorem 1
  • proof