Multiplier architectures for media processing

Media processing applications typically involve large amounts of data-level parallelism and operate on low-precision operands. This paper presents multiplier architectures for multimedia processing and compares them to conventional multiplier architectures for general-purpose processing in terms of area and delay. The proposed multiplier architectures support subword parallelism and additional features, which enhance their performance in multimedia applications, yet require only slightly more area and delay than conventional multipliers for general-purpose processing.