Optimizing Area and Power of MAC Arrays in DNN Accelerators via Overflow-Aware Partial Sum Management

On-device fixed-point deep neural network (DNN) accelerators are widely used, but the multiply-accumulate (MAC) units that perform the atomic operations in DNNs have become inefficient due to the overestimation of guard bits. To address this issue, we propose an overflow-aware management mechanism. This mechanism promptly adds the partial sum that is about to overflow to the previous partial sum stored in the output buffer and then writes it back into the buffer, thereby reducing the local guard bit overhead in the processing elements (PEs). We implemented and evaluated a PE array based on this mechanism, which reduced area overhead for adders by 43% and for registers by 35%, along with a 14% decrease in energy consumption.