Ant Financial's MorseCrypto Wins First Place in the 2024 Financial Cipher Cup Second Stage with High‑Performance SM2/SM3/SM4 Implementations

Recently, Ant Financial's MOS Technology team won the first‑place award in the second‑stage track of the 2024 "Financial Cipher Cup" cryptographic application and technology innovation competition.

The "Financial Cipher Cup" is the most prestigious cryptography contest in China's financial sector, jointly guided by the People's Bank of China and the National Cryptography Administration, featuring innovation, challenge, and second‑stage tracks with over 300 teams from universities and financial institutions worldwide.

Cryptography, as a foundational discipline, underpins modern information technology by ensuring data security through mathematical and computational techniques, balancing algorithm security and efficiency.

The second‑stage competition focused on "High‑speed secure software implementation of national cryptographic algorithms," requiring both server‑side and mobile‑side implementations of SM2, SM3, and SM4 that satisfy functional correctness, security, memory limits, and maximal performance.

Ant Financial's MorseCrypto solution optimized the national algorithms on X86 and ARM platforms from both algorithm design and coding perspectives.

Algorithmically, the team leveraged advanced vector extensions on X86 and A64 instruction set features on ARM to create vectorized parallel implementations, minimizing instruction count and theoretical latency while supporting both single‑task and multi‑task parallelism.

Implementation-wise, they employed assembly language and inline assembly to precisely control registers, reduce memory accesses, and fully exploit superscalar pipelines, arranging independent instructions to keep execution units saturated for peak instruction throughput.

Performance results show that MorseCrypto's SM2 verification runs three times faster on X86 and 4.3 times faster on ARM compared to the fastest existing open‑source libraries.

This technology has been integrated into the MOS privacy computing product, dramatically improving performance for data‑intensive, low‑latency marketing scenarios and meeting financial institutions' requirements for national cryptographic algorithms and domestic CPU support.