How to Optimize the Cryptographic Symmetric Primitives on Loongson-2E Microprocessor

The Loongson project is the first attempt to design high performance general purpose microprocessors in China. It implements a four-issue general purpose RISC microprocessor based on 64-bit MIPS instruction set. Loongson-2E is the latest version of the project. High performance is an important design criterion for Cryptographic symmetric primitives (include block ciphers and hash functions). This is the first paper which discusses how to optimize AES-128, SHA-256, SHA-512 and Whirlpool algorithms on Loongson-2E. Our programs are, to our best knowledge, the fastest implementations on Loongson-2E. This paper also proposed several new instructions for speedup symmetric primitives and a parallel hash function mode.