dot-product

提出一种"add-before-multiply"（scatter-accumulate）点积重构方法，通过将k位量化权重直接作为累加器寻址偏移，几乎完全移除逐元素乘法，仅在最后进行一次融合解码-乘法，从而降低低比特权重神经网络推理的硬件开销，并给出了SA-PE硬件设计与FPGA实现评估。

This article examines the equivalence between the component and geometric definitions of the vector dot product in Euclidean space, providing a geometric proof using the law of cosines and a projection proof using orthonormal basis vectors.