Scholarlink

This paper presents a comprehensive software module for testing pseudorandom number generators, incorporating 31 statistical tests from NIST, Diehard, and the author’s methods. The system is implemented on the.NET 7 platform, supporting parallel computing, SIMD instructions, and optimized algorithms, which provide a processing speed of up to 148 million numbers per second. The paper presents three generators, Mersenne Twister (speed), CryptMT (balancing speed and security), and Rant (cryptographic strength), each validated with a p-value deviation of less than 0.5%. The results demonstrate the efficiency of the architecture, including scalability up to 16 cores, a 68% reduction in memory load, and support for extensibility through the IRandomGenerator and IRandomnessTest interfaces. The practical relevance of the system spans cryptography, scientific modeling, and educational applications.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

High-Performance Modular Testing System for Pseudorandom Number Generators on the .NET

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • I. I. Minnullin,
  • R. F. Gibadullin

摘要

This paper presents a comprehensive software module for testing pseudorandom number generators, incorporating 31 statistical tests from NIST, Diehard, and the author’s methods. The system is implemented on the.NET 7 platform, supporting parallel computing, SIMD instructions, and optimized algorithms, which provide a processing speed of up to 148 million numbers per second. The paper presents three generators, Mersenne Twister (speed), CryptMT (balancing speed and security), and Rant (cryptographic strength), each validated with a p-value deviation of less than 0.5%. The results demonstrate the efficiency of the architecture, including scalability up to 16 cores, a 68% reduction in memory load, and support for extensibility through the IRandomGenerator and IRandomnessTest interfaces. The practical relevance of the system spans cryptography, scientific modeling, and educational applications.