High-performance and energy-efficient approximate multiplier for error-tolerant applications.