- Add support for using POWER8+ VPSUMD instruction to accelerate GCM
- Use a new algorithm for modular inversions which is both faster and more resistant to side channel attacks
- Optimize NIST prime field reductions, improving ECDSA by 3-9%
- Use the library thread pool instead of a new thread for RSA computations, improving signature performance by up to 20%
- Enable use of raw block ciphers from CommonCrypto
Full changelog available here.