Fig. 3
From: Real-time data analysis for medical diagnosis using FPGA-accelerated neural networks
Scalar product module design using both DSP and ALM multipliers. Each scalar product slice is composed of DSP or ALM multipliers and an adder tree with logarithmic complexity to minimize latency. For DSP-based slices, two multipliers and one adder are implemented in DSP blocks while the remaining adder tree is always implemented by using ALMs