In a significant development for cryptographic engineering, researchers Diego F. Aranha of Aarhus University and Paulo L. Barreto of the University of Washington Tacoma have released a comprehensive study detailing a breakthrough in scalar multiplication for prime curves. Their research, documented in Paper 2026/2667 on the Cryptology ePrint Archive, revisits the mathematical foundations of prime curves governed by the equation $y^2 = x^3 + b$ where $p equiv 1 pmod 3$. This specific family of elliptic curves underpins some of the most widely used cryptographic standards and pairings in modern digital security, including the heavily scrutinized secp256k1 curve utilized in Bitcoin, as well as the Barreto–Naehrig (BN) and Barreto–Lynn–Scott (BLS) families essential for pairing-based cryptography and zero-knowledge proofs.
For a decade, the cryptographic community has relied on established efficiency limits for complete addition formulas on these curves. However, by cleverly breathing new life into a fifteen-year-old technique, Aranha and Barreto have managed to break this long-standing barrier. They successfully reduced the cost of the best complete addition formula from twelve multiplications down to eleven, or an even more impressive ten multiplications when utilizing mixed coordinates. The secret to this optimization lies in the final stage of the computation, which the researchers structured as a three-dimensional cross product requiring a mere five multiplications. This subtle yet profound architectural shift marks the first time in ten years that the baseline efficiency for complete point addition on these curves has been improved.
To demonstrate the practical viability of their theoretical advancements, the authors implemented their novel formulas directly into the widely adopted libsecp256k1 library. The integration yielded a tangible, measurable performance boost, achieving a 5.2 percent improvement in constant-time fixed-base scalar multiplication. In the world of high-security cryptographic software, where every clock cycle and optimization matters for mitigating side-channel attacks and reducing latency, a performance gain of over five percent in a foundational library is a substantial achievement.
Building upon their breakthrough in point addition, Aranha and Barreto extended their research to address point tripling, a crucial operation in advanced elliptic curve routines. They proposed a complete tripling formula that achieves the exact same computational cost as incomplete tripling in traditional Jacobian coordinates. By merging this new complete tripling formula with their optimized complete addition formula, the researchers constructed an exception-free width-3 $tau$-Non-Adjacent Form ($tau$-NAF) utilizing $tau = 1 – omega$ alongside a compact six-point precomputed table.
When evaluated through a standard operation-count model, this integrated approach demonstrates remarkable efficiency gains. It proves to be approximately 14 percent faster than exception-free Gallant–Lambert–Vanstone (GLV) based scalar multiplication operating at a 256-bit security level. Furthermore, it outperforms even the fastest incomplete GLV methods by three to seven percent, proving that practitioners no longer need to sacrifice safety for speed.
To validate their operational models outside of theoretical benchmarks, the researchers conducted extensive constant-time implementations within the RELIC cryptographic meta-toolkit across various BLS curve families. The empirical results strongly corroborated their mathematical models. Within the $mathbbG_1$ group of BLS12-381, BLS24-509, and BLS48-575 curves, the $tau$-NAF method demonstrated speeds nine to thirteen percent faster than traditional GLV utilizing the same complete formulas. Moreover, the regular variant of their approach achieved speeds ten to twelve percent faster than regular GLV paired with complete formulas, and five to nine percent faster than regular GLV running on legacy, incomplete Jacobian formulas.
The performance advantages extend naturally to the $mathbbG_2$ groups of these same pairing-friendly curves, where the Gallant–Lambert–Stallings (GLS) method traditionally serves as the primary performance baseline. In this complex operational environment, the researchers’ $tau$-NAF implementation maintains a ten to eleven percent speed advantage over GLS using complete formulas, and remains seven to nine percent faster when compared against Jacobian formulas, while the regular variant captures gains of up to six percent.
Beyond standard elliptic curve group operations, Aranha and Barreto also turned their attention to extension field arithmetic, which plays a critical role in the pairing operations powering modern advanced cryptographic protocols. They developed an "imaginary" version of the classic Toom-Cook multiplication algorithm tailored specifically for arithmetic within the $mathbbG_T$ extension field. This specialized mathematical adaptation successfully yields a four to six percent performance improvement in $mathbbG_T$ exponentiation benchmarks for both the BLS24-509 and BLS48-575 curve families.
To cap off their extensive optimization suite, the researchers integrated a cyclotomic cube requiring a mere 0.77 multiplications, effectively bridging their $tau$-adic method directly over to the $mathbbG_T$ extension field. This final architectural bridge makes $mathbbG_T$ exponentiation three percent faster on BLS24-509 curves and an impressive eleven percent faster on BLS48-575 curves. Together, these theoretical refinements and practical software implementations provide cryptographers and blockchain engineers with a faster, safer, and highly efficient toolkit for securing next-generation cryptographic protocols.
