A Novel Superscalar Architecture for Fast DCT Implementation

This paper presents a new superscalar architecture for fast discrete cosine transform (DCT). Comparing with the general SIMD architecture, it speeds up the DCT computation by a factor of two at the cost of small additional hardware overheads.

• Type: Conference
• Year: 2000
• Where: IPPS
• Authors: Zhang Yong, Min Zhang
• Keywords: DCT Computation, Discrete Cosine Transform, Distributed And Parallel Computing, General Simd Architecture, IPPS 2000