RU2131617C1 - Optical digital paging fixed-point multiplier - Google Patents

Abstract

FIELD: electronics. SUBSTANCE: device uses original algorithm for algebraic multiplication of pages of fixed point numbers and usage of original methods and connection s and transmission of operand pages, original operand shift circuits in digital adder, and circuits of unit which calculates signs of product. Device has input optical register for multiplicand, which has sign bits and main bits, optical connection unit, optical unit for calculation of sign of product, optical connection units, unit of optical gates, optical connection unit, optical digital adder, which has main and sign bits and additional bit, optical unit for additional single bit, optical connection unit, controlled shift register for factor, which has sign and main bits, bit optical gates, optical connection units, output optical register, which has sign and main bits respectively. EFFECT: increased reliability, simplified design, increased functional capabilities due to execution of operations in algebraic representation. 2 cl, 2 dwg

Description

 The invention relates to electronics and can be used in various computing devices when processing information in a variety of radar, radio navigation and computer systems, both ground and airborne.

 Known optical page converter for optoelectronic storage device [1], containing light combiners, collimating lenses, optically controlled transparency, focusing blocks, a shifting unit, a beam shaper, a polarizing beam splitter, a zero code shaper, a prism reflector, a controlled light switch, optical communication units, a forming unit unit code, controlled light combiners, optical adder, memory unit, beam forming unit, controlled polycubic mu tiplikator feedback unit correcting lens, a light receiving unit and control unit. The main disadvantages of this device are the relatively low reliability, complexity and cumbersomeness, as well as the inability to perform operations taking into account the signs of the operands.

 Closest to the proposed one is an optical multiplier [2], containing an optical register of the multiplier, the input of which is the input of the page of the multipliers in the fixed-point optical digital page multiplier and an optical digital adder. The main disadvantages of this device are the relatively low reliability and complexity, as well as the inability to perform operations taking into account the signs of the operands.

 The technical result is to increase the reliability and simplification of the device, as well as expanding its functionality by performing operations in algebraic form.

 This is achieved by the fact that an optical shift multiplier register, an optical gate unit, an additional optical unit are introduced into the optical fixed-point digital page multiplier containing the optical register of the multiplier, the input of which is the input of the multiplier page in the fixed-point optical digital page multiplier and the optical digital adder. a single discharge, a controlled optical switch, a discharge optical valve, an optical unit for determining the sign of a product, output opt optical register, optical communication nodes from the first to the ninth, wherein the output of the optical register of the multiplicable through the first optical communication node is optically connected to the optical input of the optical gate unit, the optical output of which through the second optical communication node is optically connected to the input of the main bits of the first input of the optical digital adder, the input of the additional discharge of the first input of which is optically connected with the output of the optical unit of the additional unit discharge, the output of the optical digital adder through t this optical communication node is optically connected to the optical input of the controlled optical switch, the first output of which through the fourth optical communication node is optically connected to the second input of the optical digital adder, the second output of the controlled optical switch through the fifth optical communication node is optically connected to the input of the main bits of the output optical register, the input of the sign bits through the sixth optical communication unit is optically connected to the output of the optical unit for determining the sign the first input of which through the seventh optical communication node is optically connected to the output of the significant bits of the optical register of the multiplier, the second input of the optical sign-determining unit through the eighth optical communication node is optically connected to the input of the significant bits of the optical shift register of the multiplier, the input of which is the input of the multiplier page fixed-point optical digital page multiplier, low-order output of the optical shift register of the multiplier through a serial false and optically coupled with an optical isolator bit and the ninth optical communication node optically coupled to a control input unit of optical gates, the output of the output register is the output of the optical pages optical digital works paging fixed-point multiplier.

 As well as the fact that the optical shift register of the multiplier contains the optical register, the optical combiner, the tenth and eleventh nodes of optical communication, the first input of the optical combiner being the input of the optical shift register of the multiplier, the output of the optical combiner through the tenth optical communication node is optically connected to the input of the optical register whose output through the eleventh optical communication unit is optically coupled to a second input of the optical combiner.

 This set of essential features and the relationships between them allows you to get a device with high reliability, and to expand its functionality by identifying the signs of the product with any combination of signs of the factors.

 The essence of the invention lies in the fact that on the basis of the original optical algorithm for calculating the algebraic product of pages of numbers with a fixed point and the use in the optical multiplier of the original methods and schemes of communication and transmission of pages of operands, the original optical schemes: shift register, shift scheme of operands in a digital adder and circuit unit for determining the signs of works, the structure of an optical digital page multiplier with a fixed point, with the above nstvami.

 Thus, the proposed optical digital page multiplier with a fixed point has properties that are not inherent in known devices. This is due to a new set of essential features and new relationships outlined above.

 Comparison of the proposed device with the known indicates that it meets the criterion of "novelty", and the absence of analogues of the distinctive features of the proposed device is about the criterion of "inventive step".

 In FIG. 1 is a diagram of a fixed-point optical digital page multiplier; FIG. 2 is a diagram of an optical shift register of a factor.

 An optical fixed-point digital page multiplier comprises an input optical register of a multiplicable 1 having sign digits 1-1 and main digits 1-2; optical communication unit 2; optical unit for determining the sign of the product 3; optical communication nodes 4, 5, 6; block of optical gates 7; optical communication unit 8; an optical digital adder 9 having basic and significant digits 9-1 and an additional digit 9-2; optical unit additional unit discharge 10; optical communication unit 11; controlled optical switch 12; optical communication unit 13; an optical shift register of the multiplier 14 having the signed 14-1 and main 14-2 bits; discharge optical valve 15; optical communication nodes 16, 17; output optical register 18 having the significant and main bits 18-1 and 18-2, respectively.

 The input optical register of the multiplier 1 is intended for input into the computer pages of multiplicable in the form of optical paraphase signals, including the conversion from electrical ones. With the electrical input of the pages of the operands, the register can consist of interconnected electronic memory, for example, on the registers and the matrix of laser diodes, and with optical input at the input of the electronic memory, an additional array of photodetectors is located.

 The optical communication unit 2 is designed to transmit optical signals from the significant bits 1-1 of the optical register 1 to the first input of the optical unit for determining the sign of the product 3 and can be performed, for example, in the form of a light guide (fiber or integral) matrix.

 The optical unit for determining the sign of the product 3 is designed to determine the signs of the works to form the output page for any combination of signs of the factors. As such a block 3 can be used, for example, an optical adder modulo 2, which can be performed on the basis of both optical fiber and classical optics.

 The optical communication unit is designed to transmit optical signals from the sign bits 14-1 of the shift register 14 to the second input of the optical unit for determining the sign of the product 3 and can be performed similarly to the optical communication unit 2.

 The optical communication unit 5 is designed to transmit optical signals from the output of the optical unit for determining the sign of the product 3 to the sign bits 18-1 of the output optical register 18 and can be performed similarly to the optical communication unit 2.

 The optical communication unit 6 is designed to transmit optical signals from the output of the optical register of the multiplicable 1 to the first input of the optical gate unit 7 and can be performed similarly to the optical communication unit 2.

 The block of optical gates 7 serves to pass or block optical signals of the multiplier page arriving at its optical input from the optical register 1, depending on the value of the control signals arriving at its controlled input from the lowest n-th digit of the optical shift register of the factor 14 and can be made on the basis of optical fiber or classical optics.

 The optical communication unit 8 is designed to transmit optical signals from the output of the optical inverter unit 7 to the inputs of the sign and main bits 9-1 of the optical digital adder 9 and can be performed similarly to the optical communication unit 2.

 An optical digital adder 9 serves to obtain both intermediate results and a page of complete works and can be performed as described in US Pat. N 2079872 (Russia).

 The optical unit of the additional unit discharge 10 serves to round off the result of the work and can be made of sequentially located and optically coupled paraphase arrays of laser diodes, a focon and a mask (for example, photographic stencil) of an arrays (column) of paraphase units.

 The optical communication unit 11 is designed to transmit optical signals from the output of the optical digital adder 9 to the input of the controlled optical switch 12 and can be performed similarly to the optical communication unit 2.

 The controlled optical switch 12 serves to switch the operation of the optical digital adder 9 from the shift mode of its information content to the mode of transmitting the result of the work to the output of the optical digital page multiplier with a fixed point.

 The optical communication unit 13 is designed to transmit optical signals from the first output of the controlled optical switch 12 to the second input (9-1 and 9-2) of the optical digital adder 9.

 The optical shift register of the multiplier 14 is designed to store the page of the factors and shift them to the right in each cycle of the operation by one bit. The optical shift register of the multiplier 14 may comprise (FIG. 2): an optical register 19 having signed 19-1 and main 19-2 bits, respectively; optical communication unit 20; optical combiner 21; the optical communication node 22. The optical register 19 can be performed in the same way as, for example, the optical register of the multiplicable 1; the implementation of the nodes of the optical communication is similar to the implementation of the node of the optical communication 2; optical combiner 21 can be made on the basis of optical fiber or classical optics.

 The discharge optical gate 15 is designed to allow or prohibit the supply of optical signals from the lower n-th digit of the optical shift register of the multiplier used as control to the controlled inputs of the gates of the block of optical gates 7 and can be made on the basis of optical fiber or classical optics.

 The optical communication nodes 16, 17 are used to transmit optical signals from the output of the optical control unit 15 to the controlled input of the optical gate unit 7 and from the second output of the controlled optical switch 12 to the input of the main bits 19-2 of the output optical register 19, respectively, and can be performed similarly to the node optical communication 2.

 The output optical register 19 is intended for the formation, storage and delivery of the result and can be performed, for example, from a series of photodetector arrays and electronic registers connected in series with an electrical output or an array of laser diodes with an optical output can be located.

Optical digital multiplication algorithm. The source data for the multiplication — the multiplier page A = [a _ij ] and the multiplier page B = [b _ij ] —are entered on the optical registers of the multiplier and multiplier. The product page C = [c _ij ], calculated by a series of additions of intermediate results, is formed on the optical adder.

где b_(i(-1), b_i(-2), ..., b_i(-j), ..., b_i(-(n-1)),b_i(-n) - цифры парафазного двоичного кода множителя, равные 01 или 10,
тогда произведение C можно записать в виде
C_i = A_ib_i(-1)2^-1 + A_ib_i(-2)2^-2 + ... + A_ib_i(-j)2^-j + ... A_ib_i(-(n-1))2^-(n-1) + A_ib_i(-n)2^-n (1)
где A_i - i-я строка страницы A.If you expand the factor in the form

where b _{(i (-1)} , b _{i (-2)} , ..., b _{i (-j)} , ..., b _{i (- (n-1))} , b _{i (-n)} are numbers paraphase binary multiplier code equal to 01 or 10,
then the product C can be written as
C _i = A _i b _{i (-1)} 2 ^-1 + A _i b _{i (-2)} 2 ^-2 + ... + A _i b _{i (-j)} 2 ^-j + ... A _i b _{i (- (n-1))} 2 ^{- (n-1)} + A _i b _{i (-n)} 2 ^-n (1)
where A _i is the i-th line of page A.

Multiplication can be implemented by performing in a arithmetic device some cyclic process. In accordance with this process, formula (1) can be represented in a cyclic form. So, if we write formula (1) in the form
C _i = ((... ((0 + A _i b _{i (-n)} ) 2 ^-1 + A _i b _{i (- (n-1))} ) 2 ^-1 + ... A _i b _{i ( -i)} ) 2 ^-1 + ... + A _i b _{i (-2)} ) 2 ^-1 + A _i b _{i (-1)} ) 2 ^-1 , (2)
then the process of finding the resulting i-th row of the page of works by the formula (2) can be reduced to n-fold loop execution:
C _{i (j + 1)} = (C _ij + A _i b _{i (- (nj))} ) 2 ^-1 (3)
at initial values: j = 0, C ₀ = 0 (4)
According to expressions (3) and (4), in each (i.e., in the jth) cycle, the ith multiplicand from the page of multipliers A is either transmitted to the optical adder if b _{- (nj)} = 1, or is not transmitted if b _(-nj) = 0, after which the contents of the optical adder are multiplied by 2 ^-1 , i.e. shifts to the right by one digit. After the end of the nth cycle, the product page C is formed in the optical adder, i.e.

C _n = C = A • B.

The next digit of the factor b _{i (- (nj))} , which determines whether or not it is necessary to transmit the multiplier to the adder, is easiest to remove from the least significant bit of the optical register of the factor, which is shifted one bit to the right in each cycle.

 Implementation of multiplication by this method requires a shift to the right in the optical register of the multiplier and in the optical adder. To round off a work, it is enough to extend the optical adder by only one additional digit so that after calculating the work, add one to the additional discharge and remove the page of works from the optical adder without an additional discharge. If the additional rank of the ith product of the page of works of C was 0, then this ith product from the main digits is removed with a drawback. If the additional rank of the ith product was equal to 1, then when 1 is added to it, the transfer of 1 to the lower main rank occurs and this product is displayed in excess. Thus, the maximum value of the product error in absolute value is reduced to half the least significant bit.

 The principle of operation of an optical digital multiplier. Before starting the operation of multiplication, the pages of the multiplier and the multiplier are input 1 and 2, respectively, to the optical register of the multiplicative 1 and the optical shift register of the multiplier 14.

 Sign discharges from these registers through optical communication units 2 and 4 are transmitted to the optical unit for determining the sign of the product 3, in which they are added, and the result, through the optical communication unit 5, enters the sign discharges 18-1 of the output optical register 18 and represent the product sign in it.

 In the first step of calculating the product, if there is one in the lowest nth place of the ith row of the optical shift register of the multiplier 14, then the corresponding gate in the block of optical gates 7 is open and the bits of the corresponding multiplicable from the cells of the optical register of the multiplicative 1 optical communication unit 6, the optical unit gates 7 and the optical communication unit 8 are transmitted to the bits 9-1 of the optical digital adder 9.

 In the second cycle, the information content of the optical digital adder 9 and the optical shift register of the multiplier 14 are shifted to the right by one bit. To this end, the optical signals displaying the intermediate results page are received from the output of bits 9-1 of the optical adder 9 through the optical communication unit 11, the controlled optical switch 12 and the optical communication unit 13 to the second input of the optical digital adder 9 so that the least significant bit of the intermediate page The results entered the additional discharge 9-2 of the optical adder 9.

 The shift of the multiplier page to the right in the optical shift register of the multiplier 14 is carried out by transmitting its information content from the output of the optical register 19 through the optical communication unit 20, combiner 21, the optical communication unit 22 to the input of the optical register 19 so that (n-1) - The ith digit of the multiplier page entered its nth digit. The information content of the nth discharge of the optical shift register of the multiplier 14 is passed through the discharge optical gate 15, the optical communication unit 16 and fed to the block of optical gates 7 as control signals. Single control signals open the corresponding line of optical gates of the block of optical gates 7 and pass the corresponding multiplicates from the optical register of multiplicative 1 to the optical digital adder 9, which adds them with its information content, forming intermediate products. Zero control signals support the corresponding line of optical gates of the block of optical gates 7 in the closed state, so they block the optical signals of the corresponding multiplicative optical register of multiplicative 1.

 In the nth cycle, the complete digital product is formed on the optical digital adder 9, which is rounded in the (n + 1) -th cycle due to the transfer of a single discharge from the optical unit of the additional single discharge 10 to the additional discharge 9-2 of the optical digital adder 9 and their addition . The final result of the operation is transmitted from the output of the optical digital adder through the optical communication unit 11, the controlled optical switch 12, the optical communication unit 17 to the output optical register 18 (its main cells 18-2). If necessary, the result of the operation — the page of the work — is read from the optical register 18.

 Using the invention will allow to realize, by optical methods, the operation of algebraic multiplication of pages of factors in various computing devices, while increasing 10-100 times the reliability, accuracy, performance of calculations, and will also expand their functionality and increase the compactness of devices.

 Such devices can be widely used in a variety of radar, radio navigation and computing systems, both ground and airborne.

LITERATURE
1. USSR author's certificate N 1169022, MKI G 11 C 11/42, 1985.

 2. Patent N 2076548 (Russia), MKI G 06 E 1/02, 1997.

Claims (2)

 1. An optical fixed-point digital page multiplier, comprising an optical multiplier register, the input of which is an input of a multiplier page in a fixed-point optical digital page multiplier, and an optical digital adder, characterized in that an optical shift register of the multiplier, an optical gate unit, are inserted , an optical unit for an additional single discharge, a controlled optical switch, a discharge optical valve, an optical unit for determining the sign of the product, the output optical register, optical communication nodes from the first to the ninth, and the output of the optical register multiplied through the first optical communication node is optically connected to the optical input of the optical gate unit, the optical output of which through the second optical communication node is optically connected to the input of the main bits of the first input of the optical digital adder , the input of the additional discharge of the first input of which is optically connected with the output of the optical unit of the additional unit discharge, the output of the optical digital adder Without a third optical communication unit, it is optically connected to the optical input of the controlled optical switch, the first output of which through the fourth optical communication unit is optically connected to the second input of the optical digital adder, the second output of the controlled optical switch through the fifth optical communication unit is optically connected to the input of the main bits of the output optical register whose sign discharges input through the sixth optical communication unit is optically coupled to the output of the optical sign detection unit reference, the first input of which through the seventh optical communication node is optically connected to the output of the significant bits of the optical register of the multiplier, the second input of the optical sign-determining unit through the eighth optical communication node is optically connected to the input of the significant bits of the optical shift register of the multiplier, the input of which is the input of the multiplier page fixed-point optical digital page multiplier, low-order output of the optical shift register of the multiplier through sequentially located and optically connected, the discharge optical gate and the ninth optical communication unit are optically connected to the controlled input of the optical gate unit, the output of the output optical register is the output of the product page of the fixed-point optical digital page multiplier.  2. The device according to claim 1, characterized in that the optical shift register of the multiplier comprises an optical register, an optical combiner, the tenth and eleventh optical communication nodes, the first input of the optical combiner being an input of the optical shift register of the multiplier, the output of the optical combiner through the tenth optical communication node optically connected to the input of the optical register, the output of which through the eleventh optical communication node is optically connected to the second input of the optical combiner.

Images