Browse Prior Art Database

Extended Precision in Signal Processor

IP.com Disclosure Number: IPCOM000048359D
Original Publication Date: 1982-Jan-01
Included in the Prior Art Database: 2005-Feb-08
Document File: 2 page(s) / 44K

Publishing Venue

IBM

Related People

Belloc, J: AUTHOR

Abstract

The signal processor represented in the figure includes conventional elements such as two input registers RX, RY, two multiplexing gates M1, M2, an ALU (arithmetic and logic unit) and two output registers RA, RB. The output registers RA, RB store the most significant bits and the least significant bits, respectively, of the ALU output. Connected to the outputs of input registers RX, RY is a multiplier (5) performing unconditional multiplications of the RX and RY contents over two operating steps. Such a multiplier operates on 12-bit words and has no extended precision capability. Currently known signal processing application may generally be satisfied with 12-bit precision; others (e.g., accumulation of a great number of terms or multiplication results involved in correlations or filtering) may require larger results.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 64% of the total text.

Page 1 of 2

Extended Precision in Signal Processor

The signal processor represented in the figure includes conventional elements such as two input registers RX, RY, two multiplexing gates M1, M2, an ALU (arithmetic and logic unit) and two output registers RA, RB. The output registers RA, RB store the most significant bits and the least significant bits, respectively, of the ALU output. Connected to the outputs of input registers RX, RY is a multiplier (5) performing unconditional multiplications of the RX and RY contents over two operating steps. Such a multiplier operates on 12-bit words and has no extended precision capability. Currently known signal processing application may generally be satisfied with 12-bit precision; others (e.g., accumulation of a great number of terms or multiplication results involved in correlations or filtering) may require larger results. For these cases, extended precision is required.

This is the reason why this signal processor has been provided with two output registers RA, RB and additional logic circuitry (10).

The operating principle of the signal processor is based on a two-step operation including the storing and delay of controls and data during the first step and then performing the second step under control of the results of the operations performed during the first step. In other words, the output indication of the DECODE unit used to decode the type of ALU operation to be performed is not only used to control the ALU during a first...