The paper "MOSFHet: Optimized Software for FHE over the Torus" by Antonio Guimarães, Edson Borin, and Diego F. Aranha focuses on improving the performance of the TFHE (Torus-based Fully Homomorphic Encryption) scheme, which is currently the state-of-the-art for evaluating arbitrary functions. The authors divide their work into two parts: reviewing and implementing existing techniques to enhance TFHE's performance and error behavior, and introducing novel improvements and new approaches. They provide a single library, MOSFET, that includes all reviewed techniques and their original contributions. Key contributions include a new method for multi-value bootstrapping based on blind rotation unfolding and a faster-than-memory seed expansion (FTMSE) technique, which can speed up basic arithmetic operations by up to 2 times. The library is designed to be highly optimized and portable, supporting Intel AVX2, FMA, and AVX-512 Instruction Set Extensions. The paper also discusses the challenges and opportunities in optimizing TFHE, emphasizing the importance of combining different techniques for better performance.The paper "MOSFHet: Optimized Software for FHE over the Torus" by Antonio Guimarães, Edson Borin, and Diego F. Aranha focuses on improving the performance of the TFHE (Torus-based Fully Homomorphic Encryption) scheme, which is currently the state-of-the-art for evaluating arbitrary functions. The authors divide their work into two parts: reviewing and implementing existing techniques to enhance TFHE's performance and error behavior, and introducing novel improvements and new approaches. They provide a single library, MOSFET, that includes all reviewed techniques and their original contributions. Key contributions include a new method for multi-value bootstrapping based on blind rotation unfolding and a faster-than-memory seed expansion (FTMSE) technique, which can speed up basic arithmetic operations by up to 2 times. The library is designed to be highly optimized and portable, supporting Intel AVX2, FMA, and AVX-512 Instruction Set Extensions. The paper also discusses the challenges and opportunities in optimizing TFHE, emphasizing the importance of combining different techniques for better performance.