Share Email Print

Proceedings Paper

MultiPADDI-2 board for image processing
Author(s): Vason P. Srini; Nelson Chow; Roy A. Sutton; Jan M. Rabaey
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

We have constructed a prototype image processing board containing 384 processors in 8 VLSI chips. The goal of the prototype is to show how fine grain parallelism present in image processing applications can be exploited by using lots of simple processors interconnected in clever ways. Each processor has a 16-bit data path, a simple instruction set containing 12 instructions, a simple control unit, and a scan chain for loading data and program. Each VLSI chip, called PADDI-2, contains 48 processors. The programing model used for the processors in MIMD. Each processor has 8 words in the instruction memory. There are internal registers and queues in a processor for storing data and partial results. Data is assumed to be entering the system as a stream and processed by the processors. Each VLSI chip is connected to an external memory (64 K by 16). A hardware synchronization mechanism is used for communication between processors, memory, and the external environment. If a sender and receiver is within the same chip, communication can be done in one cycle by the hierarchical interconnect bus structure. Programming the processors and the interconnections are done at compile time. The board is interfaced to a Sun SPARCstation using the SBus. Video input and output is supported by the board and field buffers are used for buffering. Software tools for checking the board, running test programs at the assembly language level, and libraries for application development have been produced. Image processing applications are currently under development. The board is available for experimentation over the Internet. Further details are available from the project web page (

Paper Details

Date Published: 19 September 1997
PDF: 12 pages
Proc. SPIE 3166, Parallel and Distributed Methods for Image Processing, (19 September 1997); doi: 10.1117/12.279605
Show Author Affiliations
Vason P. Srini, Data Flux Systems Inc. (United States)
Nelson Chow, Univ. of California/Berkeley (United States)
Roy A. Sutton, Univ. of California/Berkeley (United States)
Jan M. Rabaey, Univ. of California/Berkeley (United States)

Published in SPIE Proceedings Vol. 3166:
Parallel and Distributed Methods for Image Processing
Hongchi Shi; Patrick C. Coffield, Editor(s)

© SPIE. Terms of Use
Back to Top
Sign in to read the full article
Create a free SPIE account to get access to
premium articles and original research
Forgot your username?