TFHE-rs v1.0: Stable CPU Backend
Blog post from Zama
Zama has released TFHE-rs v1.0, marking a significant milestone for the TFHE-rs library by stabilizing its high-level API and refining key cryptographic parameters to enhance security while maintaining performance. This update reduces the probability of computational errors to less than 2^-128, aligning with modern cryptographic standards, and introduces a TUniform noise distribution for better support in distributed protocols. Alongside the software release, Zama published a comprehensive handbook detailing the library's implementations, algorithms, and benchmarks, with a focus on Gaussian noise distribution. The release also simplifies the contribution process and encourages community engagement through the Zama Bounty Program, incentivizing developers to advance the library's features. Looking forward, TFHE-rs aims to enhance user accessibility and expand hardware support, particularly focusing on optimizing the GPU backend for improved homomorphic computations.