tubGEMM: Energy-Efficient and Sparsity-Effective Temporal-Unary-Binary Based Matrix Multiply Unit
tubGEMM combines temporal-unary and binary encoding for exact matrix multiplication while exploiting value sparsity. A 128×128 INT8 implementation in TSMC N5 is evaluated for area, power, and energy, with workload sparsity reducing energy by more than 3×.