Yarkin Doroz, NVIDIA
NVIDIA announces the release of cuPQC SDK v0.4, which brings significant enhancements and new capabilities to the cuPQC library. This release focuses on primitives that will aid the Post-Quantum Cryptography Alliance (PQCA) in promoting quantum-safe cryptographic solutions and advancing post-quantum cryptography (PQC) research.
Key Features and Enhancements in cuPQC 0.4
- Extended Hash Function Support: cuPQC 0.4 introduces support for a wider range of hash functions, including SHA2, SHA3, SHAKE, and Poseidon2-BabyBear. These robust hash functions are crucial for various cryptographic applications, ensuring secure and efficient processing.
- Comprehensive Merkle Tree Integration: The release includes full support for Merkle tree calculations, including creating tree structures, proof generation, and verification. Merkle trees provide an efficient mechanism for managing data integrity and verification processes, improving time complexity from O(N) to O(log N).
cuPQC is designed to integrate effectively with other open-source projects. Previously, the Open Quantum Safe (OQS) project integrated ML-KEM support from cuPQC. This collaboration allows developers and researchers to benefit from GPU-accelerated cryptographic functions, enhancing both the performance and security of their applications.
cuPQC 0.4 now provides high-performance implementations of Merkle trees and hash functions, which developers can use to construct GPU-accelerated schemes that enhance various PQC algorithms. This enables developers to build and accelerate cryptographic algorithms within open-source projects like the Open Quantum Safe (OQS) framework. Some example schemes that could be implemented and accelerated are hash-based digital signatures such as XMSS, LMS, and SPHINCS+.
Researchers can leverage these new cuPQC functionalities to advance PQC research. The integration of Merkle trees and specialized hash functions facilitates the efficient implementation of data integrity checks, verification processes, zero-knowledge proofs, and membership proofs. These capabilities are vital for developing privacy-preserving applications, including secure voting, confidential transactions, and high-security scenarios such as access control systems and database queries.
Getting started.
Start exploring the new features of cuPQC today. Download cuPQC and see how easily you can integrate it into your projects, with practical examples and usage scenarios provided. Nvidia’s comprehensive documentation includes detailed guides, API references, and troubleshooting tips to help you maximize the benefits of cuPQC. Discover more about the added functionalities in the full blog post.
In order to understand how cuPQC can be used to accelerate LibOQS, start by exploring the previous ML-KEM integration with LibOQS. You can review the sample ML-KEM CUDA code that runs as a backend in LibOQS to understand its structure and functionality. Additionally, you can examine the CMake configuration that manages build-time file selection and analyze the API selection logic used during compilation to determine the appropriate implementation.