"Flexible multiplier blocks for accelerated processing in a 3D softchip" by Oliver Pfaender, Stefan Lachowicz et al.
 

Flexible multiplier blocks for accelerated processing in a 3D softchip adaptive computing system

Document Type

Conference Proceeding

Faculty

Faculty of Computing, Health and Science

School

School of Engineering and Mathematics

RAS ID

3569

Comments

Pfaender, O., Lachowicz, S., & Pfleiderer, H. J. (2005). Flexible multiplier blocks for accelerated processing in a 3D softchip adaptive computing system. IFIP VLSI-SoC 2005 : IFIP WG 10.5 International Conference on Very Large Scale Integration System-on-Chip.

Abstract

The 3D-SoftChip Adaptive Computing System utilizes advanced wafer bonding techniques to vertically integrate two 2D chips. It manipulates configurable hardware arithmetic primitives to form a powerful soft-configurable and highly efficient processing system, providing high computation power and data throughput for today's telecommunication, photonic networking and content-rich multimedia applications. This paper first reviews the 3D-SoftChip architecture and then concentrates on the accelerating processing elements which feature a run-time reconfigurable multiplier with data exchange interfaces and a concatenation option to enable adaptable word-length computations. A flexible and low-overhead multiplier architecture suitable for enhancing the array of processing elements is presented. Finally, the interconnect requirements for full-flexibility multiplication is estimated, and a simple way of transferring a considerable amount of control circuitry from the intelligent configurable switch into the processing elements to reduce the switch complexity is proposed.

This document is currently not available here.

Share

 
COinS
 
 
 
BESbswy