UltrafastSecp256k1 v3.3

This update introduces significant speed improvements in batch operations, achieving speeds 17 to 67 times faster than the previous version (v3.21.1). These enhancements are primarily due to the transition from Jacobian coordinates to an all-affine accumulation strategy for batch scalar multiplications, which significantly reduces computational workload by employing mixed Jacobian+Affine additions. Moreover, a 10% increase in OpenCL generator multiplication speed was achieved through the introduction of a precomputed affine table, thus eliminating the need for per-thread table construction. The Schnorr batch verification process has been optimized with eight distinct enhancements, such as cached x-only pubkeys and reused scratch buffers, improving efficiency especially in systems like the Lightning Network that verify numerous Schnorr signatures in bulk.

The GPU backends have also received attention, with CUDA being enhanced with precomputed tweak tables for BIP352 silent payment scanning, and Metal updates ensuring feature parity with CUDA and OpenCL through several key features. From a security standpoint, UltrafastSecp256k1 v3.3 addresses over 463 code-scanning alerts and completes all planned audit infrastructure items, reinforcing its defense against potential vulnerabilities. This comprehensive update not only boosts performance across all supported platforms but also significantly strengthens the library's security posture.

For developers and users interested in integrating or upgrading to this version, detailed instructions are provided here, facilitating access across multiple languages and platforms. The project’s commitment to advancing cryptocurrency infrastructure through open-source collaboration is evident from its ongoing developments aimed at further GPU acceleration and support for new hardware architectures, as detailed in its GitHub repository.

In addition to these updates, the UltrafastSecp256k1 team has conducted a thorough audit covering various aspects of the library. This includes mathematical correctness, constant-time layer equivalence, standard test vectors, randomized differential testing, and more, ensuring comprehensive validation of its functionalities and security measures. The audit spanned numerous platforms and configurations, employing advanced tools like ASan, UBSan, TSan, Valgrind, and libFuzzer harnesses to scrutinize the library under different scenarios. The successful completion of these audits underscores the reliability and robustness of UltrafastSecp256k1 v3.3, making it a highly secure and efficient implementation of the secp256k1 library for developers and users alike.