A novel vector/SIMD multiply-accumulate unit based on reconfigurable booth array

This paper presents a 32-bit vector multiply-accumulate (MAC) architecture capable of supporting multiple precisions. The vector MAC can perform one 32÷32, one 32÷16, two 16÷16, four 8÷8 bit signed/unsigned multiply-accumulate using Booth encoding algorithm and Wallace tree compressing. A reconfigurable Booth encoding array is implemented using 8÷8 Booth unit as the basic element, and longer bit modes are obtained by combining these elements selectively. This MAC unit can also perform multiply between scalar and vector operands. 32-bit SIMD (Single Instruction Multiple Date) extended ISA (Instruction Set Architecture) and 3-stage pipeline are implemented for the MAC unit. The design is synthesized in 0.13um SMIC technology under worst case condition, and the critical path of MAC is 2.5ns.

[1]  O. L. Macsorley.  High-Speed Arithmetic in Binary Computers , 1961, Proceedings of the IRE.

[2]  Yuyun Liao,et al.  A high-performance and low-power 32-bit multiply-accumulate unit with single-instruction-multiple-data (SIMD) feature , 2002, IEEE J. Solid State Circuits.

[3]  Lingamneni Avinash,et al.  Novel Architectures for High-Speed and Low-Power 3-2, 4-2 and 5-2 Compressors , 2007, 20th International Conference on VLSI Design held jointly with 6th International Conference on Embedded Systems (VLSID'07).

[4]  Albert Danysh,et al.  Architecture and implementation of a vector/SIMD multiply-accumulate unit , 2005, IEEE Transactions on Computers.

[5]  Michael J. Flynn,et al.  Some Computer Organizations and Their Effectiveness , 1972, IEEE Transactions on Computers.

[6]  D. H. Jacobsohn,et al.  A Suggestion for a Fast Multiplier , 1964, IEEE Trans. Electron. Comput..