Abstract: Emerging applications with dynamic DSP and ML workloads demand high-performance, energy efficient, and low-latency hardware reconfiguration. However, existing reconfigurable architectures [1 ...