Accelerating Quantized DNN Inference through Precision-Scalable Multipliers Integrated in Hardware Accelerators and RISC-V Processors