Notes

• BDDT optimization delays extension field operations, converting expensive 𝔩𝔩 operations to more base field 𝔰𝔰 operations, improving performance
• SVO uses Lagrange interpolation to replace polynomial expansion, reducing pre-computation cost from exponential to polynomial
• Eq-Poly optimization based on Gruen method, splitting and reducing 𝔩𝔩 multiplications for eq polynomials
• Two-stage strategy: first 3 rounds use SVO, subsequent rounds apply Algorithm 5
• Optimization is targeted at base field such as Baby Bear, with significant higher extension field operation cost

• BDDT优化通过延迟扩展域运算，将昂贵的𝔩𝔩操作转换为更多基础域𝔰𝔰操作，提升性能
• SVO利用拉格朗日插值替代多项式展开，降低预计算成本从指数级到多项式级
• Eq-Poly优化基于Gruen方法，通过拆分和减少𝔩𝔩乘法处理eq多项式
• 实现采用两阶段策略：前3轮使用SVO，后续轮次应用Algorithm 5
• 优化针对基础域如Baby Bear，扩展域运算成本显著高于基础域
• 代码实现强调与理论概念的一对一映射，便于理解核心逻辑