JPRS ID: 10064 USSR REPORT CYBERNETICS, COMPUTERS AND AUTOMATION TECHNOLOGY

Document Type: 
Collection: 
Document Number (FOIA) /ESDN (CREST): 
CIA-RDP82-00850R000400060041-6
Release Decision: 
RIF
Original Classification: 
U
Document Page Count: 
105
Document Creation Date: 
November 1, 2016
Sequence Number: 
41
Case Number: 
Content Type: 
REPORTS
File: 
AttachmentSize
PDF icon CIA-RDP82-00850R000400060041-6.pdf6.11 MB
Body: 
APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040400060041-6 FOR OFFiCIAL USE ONLY JPRS L/ 10064 22 Octoaer 1981 USSR Re ort p CYBERNETICS, COMPUTERS AND AUTOMATION TECHNOIOGY (FOUO 23/S 1) ~ - FBIS FOREIGN BR0.4DCAST INFORMATION SERVICE FOR OFF'ICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R400440060041-6 NOTE ~ JPRS publications contain information primarily from foreign - newspapers, periodicals and books, but also from news agency transmissions and broadcasts. Materials from foreign-language sources are translated; those from Englisl~-language sources are transcribed or reprinted, with the original phrasing and other characteristics retained. Headlines, editorial reports, and material enclosed in brackets - are supplied by JPRS. Processing indicators such as [Text) or [Excerpt] in the first line of each itea~, or following the last line of a brief, indicate how the original information was processed. Where no processing indicator is given, the infor- mation was summarized or extracted. Unfamiliar names rendered phonetically or transliterated are - enclosed in parentheses. Words or names preceded by a ques- tion mark and enclosed in parentheses were not clear in the original but have been supplied as appropriate in context. ~ Other unattributed parenthetical notes within the body of an ~ item originate with the source. Times within items are as given by source. - The contents of this publication in no way represent the poli- cies, views or at.titudes of the U.S. Government. COPYRIGHT LAWS AND REGULATIONS GOVERNING OWNERSHIP OF MATERIALS REPRODUCED HEREIN R~QUIRE THAT DISSEMINATION OF THIS PUBLICATION BE RESTRICTED FOR OFFICIAL USE OD1LY. APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2047/02/09: CIA-RDP82-00850R000404060041-6 FOR OFFICIAL USE ONLY JPRS L/10064 22 October 1981 USSR REPORT CYBERNETICS, COMPUTERS AND AUTOMATION TECHNOLOGY (FOUO 23/81) CONTENTS HARDWAftE Digital Correlator 1 Principles of Organization of Para11e1 Computations. Structures and Realization of Computer Systems: The M-10 1~ SOFTWARE Operation of Unified System Computers in SVS Mode 14 Excprpts From 'ALGORITHMS AND PROGR,AMS~, February 1981 16 Additional Excerpts From ~AI,GORITI~I5 AND PROGRAMS~, February 1981 19 Excerpts ~om tALGORITHI~IS Al~?D PROGRAM5~, March 1981 2? ~ Excerpts P`rom ~ ALGORITHMS t~ND F'ROGRAMS April 1981 . . . . . . . . . . . . . . . 30 Abstracts From ~APPLIED CqMPU1'ER SCIENCEt, No l, 1981 35 Parma, Network Data Base Managerrent System Based on GODASYL Recorrmiendations 38 DSP Data Px'ocessing System 39 Virtual Storage Access Method in Ye5 OS Operating System ~.l Implementation of APL on Unified System of Computers 50 _ APPLICATIONS System for Monitoring Metal Product Deliveries to Major National Economic Construction Frojects 65 - a- [III - USSR - 21.C 5&T FOUC] - G'!lA f1F~~''~ A i i f.QF (1NT Y APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY Processing of Online Data on ~,iel Reserves at Major USSR Enterprises and Power Stations 67 _ PUBLICATIONS Abstracts From Journal ~AUTOMATION AND C(~IPUTER TECHNOLOGY~, July-August 1981 69 Collection of Problems on Machine Frocessing of Economic Information 74 Multifunctional Automata and the Element Base of Digital Computers 78 - Parallel Processors for Control Systems 83 -b- FOR OFFICIAL lJSE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 NOR OFF'ICIAL USE ONLY HARDWARE DIGITAL CORRELATOR Moscow OPISANIYE IZOBRETENIYA in Russian No 739544, 1980 [Author certificate by S. N. Britin, V. P. Ipatov, Yu. A. Kolomenskiy and B. I. KorniyevskiyJ [Text] The invention relates to digital computer technology. It can be used for determining the correlation functions of various process~s. It can also Ue used in systems and devices based on correlation methods of signal process- - ing, in particular, radar and radio navigation systems. A previously known digital correlator contains four dynamic memory units, two _ multiplication units, two accumulators, and a register (1). This single-channel device is relatively complex and has insufficient speed. The presence of mu_ltipliers in the correlators is the usual reason for their relatively slow speed. The use of high-speed me~ory devices as multipliers is not always possible, because the number of quantized signal levels can reach a significant size and because this then requires an excessively large memory. Also already known is a digital correlator containin~ analog-to-digital con- verters for t:he input and reference sigrcals. Their outputs are attached via the multiplier to the adder, which sarves as integrator. The signal on adder output determines the size of the correlation of the reference and input sig- nal.s (2). "'he correlator's low speed is a drawback. This is caused by the following. It is necessary to increase the number of quantization l~vels of the input signal in order to increase the correlator's accuracy and to eliminate the suppression of weak signals. In this case, the multiplication time, which is proportional to the number of: digits in tlie multig'liers, increases. This in turn increases the minimally permitted time quantization sample of the input processes; that is, it lowers the cor.relator's speed. The goal of this invention is an increase in digital correlator speed. This goal is achieved as follows. The correlator contains first and second analog-to-digital converters (their j.nputs are the inputs of the correlator), ~nultiplication units, accumulators and equalizer. Included in the correlator ~ ~ 1 FOR OFF[CIAL. USi; ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00854R000440060041-6 FOR OFFICIAL US~ ONLY is a switch. The number of the switch's outputs equals the number of quanti- zation levels of the reference signal. The switch's inputs are connected to the outputs of the corresponding analog-to-digital converters, and the outputs are connected tu the inputs of the corresponding accumulators. The output df each accumulator is connected to the input of the multiplication unit. The output of each multiplication unit is attached t~ the corresponding equalizer input. This allows during input signal operation the execution of solely the sucnning of its quantized selected values. Each value corresponds to a definite quantization level of the reference signal. The slower multiplication opera- tion, which lim9.ts the operating speed, is executed following the completion of the input signal operation. ~ ~ 4 ~ ~ ? f f _ t i l The diagram shows the structural organization of the digital correlator. The correlator consists of analog-to-digital converters (1 and 2) of the in- put and reference signals, whose outputs are connected to the switch (3). The number of switch outputs equals the number of quantization levels of the . reference signal. Each output is attached to the corresponding adder (4), ~ which is linked to the corresponding multiplication unit (5). All outputs from number 5 units are attached to the equalizer`s inputs (6). The digital correlator operates in the following manner. The input signal arrives at the analog-to-digital converter (1), where sampling and level quantization occur. From the output of the analog-to-digital con- verter (1) the signal, which is an n-digit number, goes to the switch (3). The switch is controlled by signals arriving from the analog-to-digital con- verter (2) of the referenck signal. At each moment of the sampling, the selected quantized value of the inr~ut signal (going through switch (3)) arrives at one of the adders (4), which corre~ponds to a definite quantized level of the refer- ence signal. Following completion of the input signal, amounts stored in the adders (4) are multiplied in the multiplication units (5) by coefficients which 2 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OEF[CIAL USE ONLY equal the carrzsonding values of the quantizatian levels of the reference sig- nals. The resulting products undergo final addition at the equalizer (6). The result determines the desired correlation value. Example: Signals fro:n the outputs of the analog-to-digital convertexs (1 and 2) are represented by 10-digit numbers. Let the execution time for one addition operation equal t~; then, for ordinary multiplicat:ton devices, the execution time for one multiplication operation is approximately tm - 20t~. The quantized signal sample in a known correlator (prototype) must be selected according to the condition tk > t~ + tm = 20t~, while in the described correlator the condition is tk Z t~. The speed of the digital correlator deQ~ribed here is approximately 20 times greater than the speed of the prototype. In practice there are often cases when a digital multidigit (for example, 10 - digit) correlator reference signal accepts only 2 values. Then an increase in equipment, in comparison with the prototype, is not large. Specifically, a switch, an adder, a multiplier and an output adder are added. When the cor- relator's elements are implemented by integrated circuits, the increase in equipment volume is entirely acceptable (if one considers that correlator speed will then increase 20 times). Invention Formula The digital correlator conLains first and second analog--to-digital converters (whose inputs are the inputs of the correlator), multiplication units, accumu- ~ lators and an equalizer. The digital correlator is unique in that, in order to increase correlator speed, it contains ~a switch such that: the number of the switch's outputs equals the number of reference signal quantization levels; the correlator inputs are connected with the outputs of the corresponding analog-to-digital converters; and the outputs are connected to the inputs of . the corresponding accumulators. The output of each accumulator is connected to the multiplication unit input, each of whose output is attached to the corr.esponding equalizer input. Information sources considered during evaluation 1. USSR Author Certificate No 468247, class G Oo F 15/34, 1974. 2. Gribanov, Yu. I. et al. "Automatic Digital Correlators," Moscow, Izdatel' stvo "ENERGIYA," 1971, 150 pp (prototype) CSO: 8144/1846-P 3 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY unc ~.9.w7 PRINCIP'LES OF ORGANIZATION OF PARALLE~ COMP[JTATIQNS. STRUCTURE AND REALIZATION - OF CONff'UTER SYSi'EMS: THE M-10 Kiev KIBERNETIl{A in Russian No 2, Ma,r-Apr 81 (manuscript received 9 Dec 79) pp 68-74 [Article by Professor Mikha3.1 Aleksa.ndrovich Kartsev, doctor of technica.l sciences, Moscow~ [Text~ We will ayply the term "computer systems" to computer ha.rdware intended for the performance uf parallel computations [1]. By now at least two situations, differen~ in principle, in which computer systems are used, have been formulated. The requirements for the structure of the computer systems and the parameters on the ba,sis of which their possibilities must b~ evalu- ated are correspondingly different. The first typieal situa.tion is the use of a computer system in a shaxed multicom- puter center or au+.oma.ted. system for the control of production, of the educational process in a WZ, ete. Cha,racteristic of this situa.tion is the passage through the system of a laxge flow of tasks independent or almost independent of one another; each of them could be solved. with an ordinaxy computer. The sense of using a com- puter system instead of an aggregate of sepaxate machines in a shaxed multicomputer center consists in the possibility of more effective use of collectivized resources (professorial time, internal memory~ peripherals~ etc~, in an economy of the volume of informa.tion stored in external memories (for example, general purpose programs ca,n be stored in external memories in a single copy and copied as needed. into the internal memory several times in the interests of different ta.sks), and also in increased reliability of the sha,red. multicomputer center. When a computer center is used to solve ASU tasks an additional and, possibly, a deciding argument in favor of use of the computer center ia tr.e need to refer different ASU tasks to a common da.ta bank . The second typical situation is the use of a computer system to control a complex - technologica,l process in real time or to solve laxge-sca.le scientific or engineer- ing problems. It.is cha.racteristic her~s tha.t all the resourcea of the computer system are ma.de available if possible to a single priority task (if it ean use those resources); tha.t task is so labor-int~nsive that no combination of separate computers of the ord3.nary type can assure the obtainin~ of a solution in an accept- able time. Also possible are situations in which the purpose of the creation of the computer � system is the achievement of greater reliability tha,n with an ordinary ma,chine or . ~ FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FUR OFFICIAL USE ONLY simplification of d.ata transmission channels and the obtaining of a gain in rela- tion to some other factors, but these situations will not be discussed. The main - attention ha,s been concentrated on a typical situation in which it is necessary to obtain from a computer system a high rate of solution of a single separately taken � task. If the system is n~t strictly specialized, then in another time segment it will be used to solve another large-scale task, but in tha,t case a high rate of solution of the given task will again be required, and not a high rate of passage of a flow of ma: relatively small-scale tasks. WY~en it is a ma.tter of some specific computer system, izsually reported among its parameters is the nominal system speed, obtained as the sum of the speeds of its components. However, the nomina,l speed cha.racterizes not so much the possibilities of the system as the expenditures on its creation. To chara.cterize the possibili- ties of a system in the first typical situation it is necessary to know its real - system speed, tha.t is, the speed of t~e system during the passage through it of a large flow of relatively small-scale independent or almost independent tasks, and in the second, the real user speed~ that is, the speed of the system during the solution of a single separately taken large-scal~ task. The ratio of the real speed of a computer system (system or user) to the nominal is called the system or user efficiency of the computer system. Obtaining high user productivity in the presence of adequate efficiency of the com- puter system is a very complex problem. Achievement of ihat doal requires a defin- - ite correspondence between the structure of the computer system and the properties of the task to be solved. A given computer system cannot be identically efficient in solving tasks of different classes. Three characteristics can be distinguished which are to a different degree intrin- sic to different tasks and ca.n be used for the organization of parallel computa- tions during their solution (three types of parallelism~. 1. Parallelism of independent branches of a task. Separate large sections of a task can be solved independently of one another~ only interchanging results with one another from time to time. ~ 2. Parallelism of related operations. The result of performance of some operation is not used without fail as the input operand in the operation following it, and perha,ps is not used also in several subsequent operations~ as a result of which those related operat3.ons can be performed simultaneously. 3. Parallelism of a set of ob jects and natural parallelism. It is cha.ra.cteristic of tasks in processing informa.tion on a large number of identica.l or almost identi- ' cal objects, in other words~ of tasks dealing with da.ta fi1Ps. Th:.s is a particular case of a broader understanding of natural parallelism, when a task has been formu- lated from the very start in terms of the processing of multidimensional vectors or ma,trices, lattice fltnctions or other objects similar to them [2~ 3~� Different structures of computer systems correspond to the use of different types of pa-rallelism. Parallelism of independent branches is realized ei�her by a.inultima.chine complex or a multiprocessor system with separate control, where each processor coiztains 5 - FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFF'IC1AL USN: QNLY an autonomous control device and works in accordance with its own independent pro- gram (Fig 1) or a system of the conveyor (main-line~ type containing just as many control devices in the conveyor~ connected alternately (on a ring) to different stages of the conveyor, as there are stages in the conveyor. After a certain com- puter in the mul~ima,chine complex (or a certain processor of the multiprocessor _ system~ or a certain control device in the ma,in-line system) ha,s completed the filling of its branch, there is a possibility of veri~ring somehow whether the other parallel branches ha,ve been completed, of organizing, if necessary, the ex- pectation and exchange of informa.tion and of triggering the autonomous filli~g of following branches. Figure 1. Exampl~ af construction of 1 y y y a multiprocessor system with sepaxate control 2 A A A ~ ~ i i n~ ~ ~ i 1. GOritrO~. 3 ~ ~ ~ ~ ~ ~ 2-- arithmetic-?ogica,l unit ~ ' 3-- individual storage device of adwaA nas+l+m.b processors (can be absent) 4 common memory f 5 channel ~ ~ ~ - 5 Unfortunately, the parameters proposed. in scientific works for numerical estimation of the parallelism of independent branches intrinsic to a certain task relate not so much to the specific task being solved as to a specific program for solving it. But even detailed knowledge of those paxameters far from always makes it possible to estima.te the real user speed and efficiency of a computer system oriented toward parallelism of independent branches, in solving a given task. Therefore interest is presented by works in which those cha.racteristics are determined by simulation of com utational processes in solving real tasks. In particular, in one of those works ~4~~ tas~cs of x-ray structure analysis, weather foreca,sting~ etc~ were inves- tigated. The results of all those works are more or less similax and are presented in aver- aged form on Fig 2, which shows the nominal speed.sH~ the real.user.efficiency sP(n) ~.nd the user efficiency of the computer system Fh are shown as""a f~.inction of _ n, the number of machines in a multima.chine complex, of processors in a multipro- cessor system with separate control or of ind~pendent control devices in a conveyor ~ system. It is evident from Fig 2 that~ if the purpose of the creation of a computer system of the type under consideration is the achievement of high user speed in solving a broad circle of tasks, it is harc.ry advisable to specify n larger than 4-6. Some- times systems of that kind are constructed with a larger number of ma,chines (or ~ 6 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY SH.SPv+J ss / SN 15 / 6s . / 55 ' ~ SP~N 4s / 3s 2S ~ Figure 2. Speed and user ~s efficiency of computer systems t 2~ y s 6~ e 9 ip q~ using pa.~allelism of independent branches. En t,0 ~ 0,6 Q4 0,2 , t 2 J ti4, s, 6 7 8 9 10 n. processors or control devices). In tha.t ca.se the purpose is ma.inly to achieve high system productivity. But if in the stream of tasks a relatively large-scale task appear~, then a subsystem can be sepaxated for it, one containing as many machines (processo;-s or control devices) as tha,t ta.sk can efficiently load. A~ a result, the s;~stem efficiency of the computer system obtai.ned is high enough, but the user _ productivity in the solution of a single separately taken task, except special ca.ses, cannot exceed by more than 3--4- times the productivity of an ord.inary single- processor machine. Parallel~sm of rela.ted operations is used diff~rently in different develo~oments. Its classica,l use is a multiprocessor system with general control and vectorial instruction, containing separate operationa,l codes for different pr~cessors (Fig 3). _ It is possible to create conveyor (main-line~ systems oriented toward this type of parallelism. A numerica,l chaxacteris~ic of the paxallelism af related operations, intrinsic to a given task, is the indica.tor of connectedness of related opearations c( the probability that the result of some operation in a program will be used in the operation following it. The value of a depends not so much on the properties of the task as on the quality of the ~.ocal optimization of the program, done in the . process of translation into ma,chine language or~ in some ca,ses, by appa,ratus directly in the process of interpreting the program. For many tasks of a 7 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2047/02/09: CIA-RDP82-00850R000404060041-6 ~ FOR OFFICIAL USE ONLY ' I , 1 y � 9p Sy~ 4s ~s� - z i ~ 3 ~ n ~ ~ ~ , . . . ~ ~ ~ ~s ~ -~-~-a-~1~ L J L J ~ ~ ~ _ / � � � 23 S !rt) ~a'43 ~ Odwc~ nawAms ~z~ Is 5 ~ ~ o i z .f e s 6 n Figur.e 3. Exemple of construction of a Figure 4. Speed of computer systems multiprocessor system with common con- using paxallelism of related trol, oriented toward parallelism of opera,tions. related operations. 1-- for oc. = 0.1 2-- for oC = 0�5 1 to 5-- as for Figure 1 compu~`~,tional character the value of a obtained is more or less identica,l--from 0.1 when the program is optimized very caxeflzlly to about 0.5. It is evident from _ Fig 4~ that, yn,a system oriented taward the use of parallelism of related operations~ it is difficult to expect gains in user productivity as compaxed with an ordinaxy computer of more tha,n 2.5-3 times; to obtain tha.t, n(the number of processors in a multiprocessor system or stages of a conveyor) must be of the order of magnitude of 3-5 � Parallelism of a set of objects and natural parallelism are used in the structure of a multiprocessor system with general co~trol (Fig 5), when all the procsssors simultaneously perform identica,l operations, but on different da,ta (relating to different objects or different components of multidimensional vectors, etc). Ana.logous constrizction of a conveyor system also is possible. It is desirable to supplement that structure with a ma,sk mechanism, a file sorting processor~ and general arithmetic circuits for performing such operations as computation of the scalar derivative of vectors~ finding the minima,l or m?.xi,-nal element in a file, etc, for Axample. The principal qua.nti�~,tive characteristic of parallelism of a set of objects or na,tural parallelism is the rank af the task r--the number of objects of the sa.me kind in the set of objects being processed, or the dimensiona,lity of multidimen- sional vectors, the number of points at which the values of f~anctions are given, etc. It can readily be shown tha.t the user efficiency of a multiprocessor system oriented toward parallelism of a set of objects or na.tural parallelism depends only ~ on the ratio r~n, and the user efficiency of an ana.logous conveyor system--ma,inly on the ratio r~n and to a small degree on n. Here the value of n~ as previously, 8 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00854R004400060041-6 FGR OFFIC(AL USE ONLY 1 y En,n..L ~ 2 A A q 1,0 - ~ 3 ~n~ ~n~ ~n~ 0,8 fi- - L J L J L J ~~s 1 - - I - -I- - ~ - ~ _ ~ ( ~ I ~ ~ ~.4 -T- - - r- - - f- - -t - - t- - -i- L~ 06WnA naMAme O,z - -t- - - - - -F - - - ~ a o ~ z 3 a ~s e.~ 5 ~ Figure 5. Example of construction of a Figure 6. Efficiency of a multipro- multicompressor system with general con- cessor system using natural parallel- trol, oriented toward na,tural pa,rallelism. ism. 1 to 5-- as for Figure 1 is the number of processes in a multiprocessor system or the number of stages in a conveyox system. In a somewhat idea.lized form those dependences are presented for a multiprocessor system on Fig 6. It is evident f`rom the figure that an r~n ratio larger than or equal to 4~, for example, the user efficiencey will be at least 0.8 in any ca.se. This means tha.t at a rank of the tasks~ for example, of not less thaxi 100, 25 pro~essors under general control can be efficiently used in the sys- tem, and in tha,t case the gain in real user spead obtained is at least 20 times as compared with an ordinaxy computer, and at a rank of the tasks of at least 100, 250 rocessors ca.n be used correspondingly (if other circumstances do not prevent that~ and a gain in real user speed of at least 200 times can be obtained. - Similar resu].ts ca,nnot be obtained during orientation toward parallelism of inde- pendent branches or of related operations. Of course, such a gain in rea.l user speed is possible only for a certain class of tasks which have a high rank of n~,~;ural parallelism or, in particular~ parallelism of a set of objects. However, among large-sca.le tasks, for the solution of which computer systems with high user pioducti_vity are required, such tasks occupy a ver,y important place. Attention must also be turned toward a difference :'~n the cha,racteristics of pro- gramming when various types of parallelism and correspondingly different struc- tures of computer systems are used for the organiza.tion of parallel computations. When parallel independent tra.~~ches and natural parailelism (parallelism of a set - of objects) are used~ new programming languages or supplements to existing ones are nesded. Some means for description of parallel processes are available only in thE _ PL~1 langua,ge . It is necessary, however, to take into consideration the presence of a large ~nd of programs written in ALGOL-60, FORTRAN and other widespread languages~ and the presence of a large number of programmers who are well acquainted with those languages and ha,ve experience in working with them. 9 ~ FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 , FOR OFFICIAL U~E ONLY - Let a program be written in ALGOL-60 or FORTRAN without any ~ort of supplementary instructions; it is proposed to use paxallelism of independent branches. In tha.t" case the translator must conduct an investigation on the object of the possibility - of parallel execution of all sections of the program on a single level {blocks of AIGOL-60 or FORTRAN subroutines) . It is not so complica.ted a ma,tter to formulate the rules according to which the tra,nslator would determine the possibility of - executing a pair of sections of the prqgram in parallel. However, the work which the translator must do in tha.t ca.se is enormous for a laxge program~ as an exha.us- tive search is required for all pairs of blocks of the AI~GOL program or subrou- tines of the program in FORTRAN. Therefore when parallel independent branches a~ used it is necessary to require tha,t the programmers ma,rk the program in the source language by means of additional ~ If it is proposed to use natural parallelism or parallelism of a set of objects, - and the program is written in AIG OL-60 or FORTRAN~ for example, then the task of the translator consists in investigating the programming langua.ges on the subject of the possibility of their parallel execution. The formal rules according to which the translator must act prove in that case to be of ap. �oxima.tely the sa,ma complexity as the rules according to which the translator de1 :rmines the possibil- ity of parallel execution of two sections of the program dur3zg use of parallel independent branches. However, the work of the translator on the whole in tha,t case is far simpler~ since each program cycle is investigated sepa,rately (perha,ps only together with the cycles built into it) and joint investigation of several cycles is not required. The additional language resources which simplify this task could consist in the introduction into the language, together with such objects as files (which already exist in AI~OL-60 and FORTRAN) , also of vectors, ?~~a,trices and lattice functions - (real, purpose and Boolean) and in the determination of opera.tions on those ob- jects. Let us note tha,t from the point of view of the programmer there is a substantial difference between those supplementary resources which are suitable for describing parallelism of independent branches and the means of description of natural paral- lelism. The introduction of supplemen+,,ary instructions regarding the possibility of separating independent branches represents "excessive" work for the programmer, and the efficiency of the computational process which will be obtained during execution of the program will depend to a considerable degree on how well that work - has been done (in other words, on the skill of the programmer). On the other hand~ the introduction of such objects of langua,ge as files~ vectors~ ma,trices, f~nctions _ and the determina.tion of operations on them greatly simplifies the work of the programmer. These tasks in the processing of large flows of information~ which have to be solved in practice by means of software with high prod.uctivity, in ma,ny ca,ses have to some degree all types of parallelism. To obtain high efficiency it is advan- tageous to construct combined systems [5]. For example, tne M-10 computer (Fig 7) is oriented toward use of natural parallelism and parallelism of related operations [6~. The ma.in processor part of the ma.chine consists of two ~crogram-tuned lines of pro- cessors. Each line, depending on the operationa,l code~ represents either a 10 ~ FOR OFFICIAL USE ONLY ~ APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2047/02/09: CIA-RDP82-00850R000404060041-6 FOR OFFICIAL USE ONLY 1 ~b ~y~i.s Pfg~; , , , L . : . . rH^ i~ ~ . ~'t~.. q., { ~4:t^,: 4. jw ~ . . N ~d , '.'ti~. ~ ~ Mf, .4:..::. ~.,Y ~ 4 + Figure 7. The M-10 computer. 16-digit figure of eight~ a 32-digit tetrad or a pair of 6~b-digit processors which perform one and the same operation (on different data, of course). The other line of p~ocessors for a separate operational code contained in a vector-instruction - can simultaneously perform the same or a different operation. Definite operational codes gather all the processors of the line in a single vectorial processor. For exam~le~ during performance of the operation "sca.lar prodact", in the line of pro- cessors in the course of a single cycle eight pairs of 16-digit or four pairs of 32-digit numbers are multiplied in pairs and the obtained products: are summed with one�.another and with the sum accumulated in the preceding cycle. Simultaneausly with the performance of operations on numbers~ in the ma.in processor lines up to five lines of Boolean variables are produced in which each binary digit _ corresponds to a definite sign relating to operands participating in the operation or tQ the result of the operation. For example, during addition (of eight pairs of 16-digit qr four pairs of 32-digit, or two pairs of 64-digit numbers), lines of $ aolean var~ables are produced which contain 8, 4 or 2 binary digits respectivelys --the sign o~ overf~llingi e--the sign of equality of components with one an- other~ m--th,e sign that the first component'~ in the given pair proved to be larg-� er than the secor~d; z--the sign of equality of the result to zero; x--the sign of - obtai~ning a nega~ive result. These signs cart ~e transmitted directly or through the memory into a special pro- cessor for processing lines of Boolean variables which acts simultaneously with the lines of main processors and is capa.ble of executing the complete set of logical operations on Boolean variables. The lines of Boolean variables obtained directly during execution of the main arithmetic or logica,l operations on the numbers or as a result of the work of the indicated special ~rocessor, or read from the memory, can be used ~.irther to organize provisional transmissions of control or as "ma,sks" to perform the basic arithmetic-logica,l operations. If in the operational code for tl~e ba.sic line of processors it is pointed out tha.t the operation must be car- ried out under a ma.sk, then in the line those processors are blocked for which the corresponding places of the mask contain zeros (and when the operation is per- ~ formed under an"inverse" ma.sk the processors are blocked for whi.ch the correspond- ing places of the ma,sk contain units~. 11 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFtCIAL USE ONLY The mechanism of ma.sks is an important tool in the efficient organiza,tion of parallel computations. Still another specialized processor in the.central part of the ma,chine is intended for the execution of index operations. For tasks whl.ch ha,ve sufficient natural parallelism the mean speed of the M-10 is more than 5 million operations per second. The ca.pacity of the internal memory of the ma.chine is 5 Mbytes (1,310,720 words). In tha,t case, to provide all the processors with operands, each reference to the memory is performed in a forma,t of up to 64~ bytes (up to 512 informa.tion positions) and the machine instructions axe two-address. The operating system of the machine assures users working in an interactive mode - with time sha.ring access to the transla~,ors and means of debugging programs in algorithmic languages, reference on the logica.l level to peripherals and standard procedures, the use as finished modules for the cha.rging of programs from the li- brary of standard programs (lineax algebra, approxima.tion of ~znctions, squaring~ the integration of ordinary differential equations and equations in partial deriva~~ tives~ etc). External exchange of information is done through a multiplex cha.nnel with a tatal throughput ca.pacity of about 7 Mbytes~second. The multiplex channel ha,s 24 duplex subchannels~ to each of which in turn up to six peripherals of the same kind can be connected. The basic complex of the ma,chine includes as peripherals terminals constructed on the basis of a typewriter with punched-tape equipment~ alphanumeric line-printers, punched card equipment and an engineering panel typewriter by means of which the apnaratus journal of the ma.chine is kept. Through additional coupl~ng devices, terminals based on hatched displays with keyboaxds and a light pen can b~ - connected to the channel, as well as ma,gnetic tapes and disks and other YeS com- puter peripherals. The cha.nnel ha,s its own buffer memory with a volume of 6~b Kbytes, in which channel command words also axe stored, but it can be addxessed directly also to any cell of the internal memory of the ma.chine. In the machine provision ha.s been ma,de for circuits which permit combining up to seven M-10 computers in a single synchronous complex working from a common clock pulse generator. In each cycle the machine, working in the complex can issue to its output lines a data file of 6~1~ bytes and receive a file of the same size ~om any other ma,chine of the complex. In that ca,se the addressing of the ma.chines is - virtual Special apparatus of the descriptors of connections efficiently re- places the virtual number of the machine~ ~om which information is received~ by a physica,l number. When necessary this must pe.rmit forming several subcomplexes within the synchronous complex. At the same time, as was supposed, together with natural parallelism and parallelism of related operations, parallelism of indepen- dent branches could be used in the complex, and increase of the system productivity could also be achieved. However, the possibility of organizing synchronous com- plexes of M-10 computers has not yet been realized anywhere in practice. In the ].ong term, obviously, it makes sense to become oriented toward simultaneous use of parallelism of independent branches (2 �0 4~ branches) and parallelism of a set of objects. 12 , APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2047/02/09: CIA-RDP82-00850R000404060041-6 FOR OFFiCIAL USE ONLY In tha.t case the equipment of the system can be constructed so tha.t programs for the solution of various tasks~ even in a ma.chine language, do not depend at all on the number of processors performing para.llel processing of sets of objects, and look exactly the same as programs of an ordinary (one-processor) ma.chine. In this way cne achieves both complete program compatibility of systems different in their completeness of sets (number of processors) and high viability of the system (the possibility of continuing to work when some of the processors fail). BIBLIOGRAPHY 1. Kartse~~, M. A. "Tre Problem of Organiz~.ng Parallel Computations an3 the Structures of Computer Systems." Materia.ly Vsesoyuznoy konferentsii "Vychis- - litel'nyye sistemy, seti i tsentry kollektivnogo pol'zovaniya" (Ma,terials of the All-Union ConfPrence on "Computer Systems, Networks and Sha,red Multicom- puter Centers") ~.1. Nov^gibirsk, 1980~ pp 64~-77. 2. Glushkov, V. M., Kapitonova, Yu. V.~ and Letichevskiy, A. A. 'fiheory of Data = Structures and Synchronous Pa:rallel Computations." KIBERNETIl{A~ 1976~ No 6, PP 2-15� 3. Glivenko, Ye. V. "On One Principle of Pa,rallel Computations." PRfiGRAMI'~RO- v~~, ~977~ xo 1~ pp 3-9� 4~. Ma,rtin, D. F., and Estrig, G. "Experiments on Models of Computation and 5ystem." IEEE TRANSACTIONS ON ELECTRONIC COMP'UTERS~ EC-16, 1967, No 1, pp 67-79. 5. Kartsev~ M. A. "Questions of the Construction of Multiprocessor Systems." VOPRO5Y RADIOELII{TRONIKI, 1970, No 5-6~ pp 3-19. 6. Kartsev~ M. A. "The M-10 Computer." DCd{LADY AKADEMII NAUK SSSR, 1979~ 2~5~ No 2~ PP 309-312. COPYRIGHr: IZDATEL'STVO "NAUKOVA DUMKA"~ "KIBERNE"rIKA", 1981 2174~ cso: 1863/258 13 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY SOFTWAAE UDC 681.3.06:[658.012.011.56:658.7/.8] - OPERATION OF UNIFIED SYSTEM COMPUTERS IN SVS MODE Moscow MATERIAL'NO-TEKHNICHESKOYE SNABZHENIYE, SERIYA 4: PRIMENENIKE - MATEMATICHESKIKH METODOV, VYCHISLITEL'NOY TEKHNIKI I ORGTEIQ3NIKI V MATERZAL'NO- TEKHNICHESKOM SNABZHENII in Russian No 5, May 80 (manuscript received 17 Jan 80) pp 10-11 [Abstract prepared by G. S. Gradov, deputy department chief, USSR Gossnab Main Com- ~puter Center, and V. K. Burmistrov, senior scientific associate of the NIIMS] [Text] Operation of YeS OS in the Mode of Multiprograuuning with a Variable Number of Tasks that Jointly Use Virtual Storage (SVS) Associates of the USSR Gossnab Main Computer Center and the NIIMS [Scientific Research Institute of Economic and Organization of the Supply of Materials and Equipment] have generated a specific version of YeS OS in the SVS mode for the YeS-1055 computer. The logic of the operation of YeS OS in this mode affords effective support in building complex software systems for large-scale information problems for the USSR Gossnab, such as: interoutput balances, planning of warehouse facilities, bookkeeping reporting and other tasks of the USSR Gossnah ASU. Operation of YeS Computer in SVS Mode The YeS OS operating system of the SVS mode is a practical implementation of the concept of virtual storage in YeS computers. It can be used on "Series-2" compu- ters as well as on all computers with similar principles of operation. The segment-page method of organizstion of virtual storage is used in YeS OS of the ~ SVS mode. Segment size is 64K bytes and page size is 2K bytes. The beginning _ address of a segment is an aliquot of its size. The beginning address of a page is an aliquot of its size. The maximal size of virtual storage may reach 16M bytes. Actual storage holds the pages of virtual storage in use at a given time. Unused pages are stored in external storage in the system data set SYSL PAGE. The software part ehat defines the page replacement algorithm is called the replacement algorithm. 14 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 ~ao replacement algorithms are provided in the SVS mode: the standard (LRU--least recently used) and the modified. Sometimes the guaranteed availability o� a page of virtual storage may be required in actual storage in the course of a certain _ time. For this, pages are fixed that are not replaced. Pages are fixed for a shart or a long time. Virtual addresses are converted into real by both hardware and software: the hardware feature is the dynamic address translation feature (DAT); the software features are address translation tables: segment tables and page tables. COPYRIGHT: Tsentxal'nyy nauchno-issledovatel'skiy institut informatsii i tekhniko- ekonomicheskikh issledovaziy po material'no-tekhnicheskomu snabzheniyu Gossnaba SSSR, 1980. 8545 CSO: 1863/245 15 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R400440060041-6 FOR OFFICIAL USE ONLY ~ EXCERPTS FROM 'ALGORITHMS AND PROGRAMS', FEBRUARY 1981 Moscow ALGORITMY I PROGRAMMY in Russian No 2, Feb 81 pp 1-132 [ExcerptsJ ~ 475. Simurzin, V. N., "Selection of Rectangular Network by Specified Irregular Networks," in the book "Osnovnyye metady matematicheskoy geologii i resul'taty issledovaniy" [Basic Methods of Mathematical Geology and Results of Research], Yakutsk, 1980, pp 112-118, bibl. of 7 titles. Gives an example of selecting a rectangular network according to the irregular net- work af seismic stations of the Sredne-Botuobinskoye gas deposits on the M-222 com- puter. 584. Bezruk, I. A. and Safonov, A. S., "Principles of Constructing an Automated System to Process Data Obtained by Using Digital Electric Geophysical Exploration Stations, PRIKL. GEOFIZIKA, All-Union Scientific Research Institute of Geophysical ~ Methods of Prospecting, 1980, issue 98, pp 93-102, bibl. of i" titles. The basic problem is solved in three stages: preprocessing, the processing proper and interpretation. Basic properties of the EPAK system have been formulated: adaptation of the system on all basic types of computers; simplicity of including new programs in the system, simplicity of operation and introduction of the system. - 585. Zabelin, V. A. and Pyatkin, V. P., "Identification of Checkpoints in Aerospace Images," in the book "Matematicheskiye i tekhnicheskiye problemy obrabotki izobrazheniy" [Mathematical and Technical Problems of Image Processing], Novosibirsk, 1980, pp 50-59, bibl. of 4 titles. Identification of checkpoints is reduced to determining local coincidence of frag- ments of images using a program in the FORTRAN-TsERN language with a size of about 100 statements. Average operating time for a reference of 20 x 20 elements and a search field of 50 x 50 elements is 3 minutes taking image rotation into account. 586. Lukasheva, 0. E.; Pyatkin, V. P. and Shevelev, S. L., "Systems Software for an Image Processing Complex," in the book "Matematicheskiye i tekhnicheskiye _ problemy obrabotki izobrazheniy", Novosibirsk, 1980, pp 84-94, bibl. of 4 titles. 16 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00854R000440060041-6 The KOIZ image processing complex includes three BESM-6 computers operating with a common field of disk storage. The FORTRAN, MADLEN and PASCAL languages were used to implement the image processing complex software. 589,. Sidorova, V. S., "Textural Analysis of Aerospace Images hy Computer," in the book "Matematicheskiye i tekhnicheskiye problemy obrabotki izobrazheniy", Novosihirsk, 1980, pp 30-36, bibl. of 2 titles. Ilescribes FORTRAN;programs for classification of ohjects hy t;zo systems of textural features based on the image processing complex of the Center for Geoinformation Processing. 590. Torogushina, Ye. F., "Raising the Efficiency of Primary Processing of Tiltmeter Ohservat~.ons," in the book "Seysmichnost' i sovremennyye dvizheniya zemnoy kory vostochnoy chasti Baltiyskogo shchita" [Seismicity and Modern Movements of Crust of Eastern Part of Baltic Shield], Apatity, 1980, pp 108-111, bibl. of 1 title. . Described are the FORTRAN programs BULL-Z and BULL-3 that speed up the process of preparing materials of tiltmeter observations for harmonic analysis. 591. Voronin, Yu. A. ar~.i Gradova, T. A., "REGION-kompleks programm dlya postanovki i resheniya zadach rayonirovaniya" [REGION Complex of Programs for Statement and Solution of Prob~lems of RegionalizingJ, Novosibirsk, 1980, 46 pages, Preprint/ Sitaerian Branch of the USSR Academy of Sciences Computer Center, No 258, 1~ib1. of 3 titles. REGION includes two subsystems: HOMItEG and DIAREG (homogeneous and diagnostic regionaliaing) and the complex of auxiliary programs PREDAT--preliminary data pro-- cessing in FORTRAN. 706. Pokatayev, A. Yu., "Optimal Organization of Computation of Full Normed Gradients," PRIKL. GEOFIZIK.A, All-Union Scientific Research Institute of Geophysical Methods of prospecting, 1980, issue 98, pp 133-136, Uital. of 3 titles. Auth~r presents original construction of comptitational process using direct access external storage to store intermediate result~. The algorithm has been implemented on the BESM-6 computer. 789. Mikhaylov, N. N. and Yanitskiy, P. A., "Numerical Analysis of Dynamics of Water Saturation in Forming Zone of Penetration in Productive Strata," PRIKL. GEOFIZIKA, All-Union Scientific Research Institute of Geophysical Methods of Prospecting, 1980, issue 98, pp 168-178, bibl. of 13 titles. Authors describe a physicomathematical model of penetration of filtrate of drilling mud in petroleum and gas hearing strata. Computations were implemented on a YeS- 1040 c~mputer. - 795. Gubin, P. I., "Software for Processing Frame Video Information (in DOS on YeS Computers)," in the book "Trudy MFTI. Seriya Aerofizika i prikladnaya matematika" [Transactions of the Moscow Physicotechnical Institute: Series on Aerophysics and Applied MathematicsJ, Moscow, 1980, pp 125-127. 17 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2047/02/09: CIA-RDP82-00850R000404060041-6 FOR OFFICIAL USE ONLY The SMOOKI program package for processing aerospace video information supports putting information into a computer from ML [magnetic tape], storage of it, statistical processing and ref?ection of processing results in various forms. 812. Kondrat'yev, 0. K. and Orlov, V. P., "Diffraction Transformation of Seismo- grams in the Method of Transmitted Composite Waves," PRIKL. GEOFIZIKA, All-Union Scientific Research Institute of Geophysical Methods of Prospecting, 1980, issue 98, pp 42-53, bibl. of 13 titles. Algorithm for construction of dynamic deep sections by seismograms of inethod of transmitted composite waves (MPOV) t_as been implemented on the BESM-4 computer. Application of D-transformation fn the MPOV eliminates correlation and determination of travel time curves of waves from the procedure for construction of a deep section. To obtain one path of a section, two-three minutes are required. 813. "Method of Calculation of Composition and Properties of Transfer of Hetero- geneous Systems, Ye. V. Samuylov, I. B. Rozhdestvenskiy, N. N. Tsitelauri and N. A. Budko, in the book "Osnovnyye metody matematicheskoy geologii i resul'taty issledovaniy", Yakutsk, 1980, pp 78-82, bibl. of 13 titles. 814. "Computer Modeling of Process of Processing of Frequency Modulated Signals in Devices on Surface Acoustic Waves," Ye. Ya. Bershadskiy, V. A. Vinogradova, V. K. Sirotko and A. P. Fedorov, ELEKTRON. TEKHNIKA. SER. 10. MIKROELEKTRON. USTROYSTVA, - 1980, issue 5(23), pp 61-66, bibl. ~f 5 titles. Technique of modeling of formation and compression of frequency modulated signals in devices on surface acoustic waves has been implemented on the M-222 computer. Calculation time when 1024 points are specified is about two minutes. COPYRIGHT: Gosudarstvennaya publichnaya nauchno-tekhnicheskaya biblioteka SSSR (GPNTB SSSR), 1981 8545 CSO: 1863/246 18 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY ADDITIONAL EXCERPTS FROM 'ALG~RITHMS AND PROGRAMS', FEBRUARY 1981 Moscow ALGORITMY I PROGRAMMY in Russian No 2, Feb 81 pp 1-132 [Excerpts] 398. Gavrilov, G. K., "Efficient Programming Technology," VOPR. RADIOELEKTRQN. SER. EVT, 1980, issue 8, pp 135-141, bibl. of 8 titles. Recommendations on imrpoving labor productivity of programmers and improving the quality of programs. 485. Brodskaya, I. M. and Kamynin, S. S., "Algorithm for Recognition of Partially Visible Objects," Moscour, 1980, 34 pages, Preprint/Academy of Sciences of USSR,' IPM [Institute of Applied MathematicsJ, No 107, bibl. of 8 titles. Size of training program is 1110 single-address instructions, and that of the recognition program is 3124. About 14500 words of main storage is required to operate the program. 505. Pikel'ner,,B. L. and Romanov, V. D., "Dynamic Display of Multiprogramming in YeS OS;" EKELTRON, TEKHNIKA. SER. 9. ERONOMIKA I SISTEMY UPR., 1980, issue 3(36), pp 31-32, bihl. of 4 titles. A method of obtaining a graphic presentation of the multiprogramming factor as a function of time has been implemented in the PL/1 language. Program size is 76K bytes, and time for construction and printout of daily diagram is three to five minutes. 511. Zaytseva, Zh. N., Yablonskaya, 0. V. ar.d Bodryagin, V. I., "Procedure for Selection of Printed Conductors for Layout," VOPR. RADIOELEKTRON. SER. EVT, 1980, issue 8, pp 124-126, bibl, of 3 t3.tles. PT/1 program for selection of arrangement of cor.ductors for layout, which supports minimization of conflict situations in designing printed circuit boards, with a size of 232 statements takes 22 sec~nds of machine time. - 513. RADIOTEK:TIIKA: RESP. MEZHVED. NAUCH.-TEKHN. SB., Khar'kov Institute of - Radioelectronics, issue 55, Khar'kov, Vishcha shkola, 1980, 109 pages, bibl. at end of articles. ~ 19 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR O~FICIAL USE ONLY From contents: V. M. Bezruk, V. A. Omel'chenko and N. A. Federov, "Comparative ; Analysis of Adaptive and Correlation Methods of Receiving Discrete Signals;'pp 3-11; V. A. Omel'chenko et al., "Recognition of Random Signals Based on Digital Analyzers of the Spectrum," pp 11-17; Yu. A. Kichigin, Z. N. Muzyka and S. A. Sokolov, "Study of Statistical Characteristics of Signal-to-Noise Ratio at Output of Radio Receiver with Nonlinear Interaction of Signal and Noise in Its First Cascades," pp 18-24; S. F. Simovskaya and 0. M. Lebedeva, "Analysis of Correlation Function of Random Radio Signal by Method of Digital Modeling on Carrier Frequency," pp 79-84. 517. Brodskiy, A. Ya. and Yevseyeva, L. M., "Software for Data Retrieval System for Scientific Research Institute," ELEKTRON. TEKF3DTIKA. SER. 9. EKONOMIKA I SISTEMY UPR., 1980, issue 3(36), pp 26-29. - 528. Alekseyev, K. V., "Subsystem for Automated Synthesis of Checking Experiments for Discrete Devices," TEORIYA KONECH. AVTOMATOV I YEYE PRIL., Academy of Sciences of Latvian SSR, Institute of Electronics and Computer Technology, 1980, issue 11, pp 79-91, bibl. of 4 titles. _ Size is about 3000 instructions. Test for checking of circuit of 23 elem~ants with 7 inputs, 7 signal branching points and 1 output, consisting of 14 sets, obtained in 6 minutes and 40 seconds. Key words: methodology, FORTRAN-4, Nova 2/10, checking experiments, discrete devices. 529. Grishina, I. B., "'Library Statistics' Program Package," ELEKTRON. TEKHNIKA. SER. 1. ELEKTRON. SVCh, 1980, issue 7(319), pp 74-75, bibl. of 3 titles. Describes the FORTFAN program package "Library Statist;Lcs" for acquisition and storage of statistics on use of programs from common libraries. 562. Kol'dyayev, V. I., "Kinetics of Charge In~ection in 1~IDP (metal-insulator~;~emi- conductor] Transistors with Floating Gate," ELEKTRON. TEKHNIKA, SER. 3. MIKF.OELEK- TRON.,1980, issue 2(86), pp 14-20, bibl. of 17 titles. 563. Kol'dyayev, V. I., "Kinetics of Charge Relaxation from Insulated Layers of Polycrystalline Silicon," ELEKTRON. TEKHNIKA. SER. 3. MIKROELEKTRON., 1980, issue 3(87), pp 14-19, bibl. of 11 titles. Difficulties o� depolarization of insulated conducting layers, prerequisites of kinetics of charge relaxation. Mathematical model of phenomenum has been con- structed on the assumption of Schottky and Fowler-Nordheim emission. Model of - charge spread has form of a system of algebraic differential equations that are solved with a FORTRAN program. Computation time of one integration for t from 0 to 100 years is about 2 minutes. 564. Kondrat'yev, I. I., "Mathematical Processing of Gamma Spectra of Samples of Aerosols Using Correlation Filtering Method," TR./DVNII, 1980, issue 91, "Prohlems of Hydrometeorology of Eastern Siberia, the Far East and the Pacific Ocean," pp 78-83, bibl. of 5 titles. 2~ FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY 567. Malinovskaya, Ye. V. and Malyshev, Yu. A., "Calculation of Resistance of Layer of Pinch-Resistor of Semiconductor Microcircuits," F.LEKTRON. TEKHNIKA. SER. 3. MIKROELEKTRON., 1980, issue 3(87), pp 20-27, bibl. of 6 titles. Algorithm forms basis of FORTRAN program that performs either "single" analyses, i.e. calculation of resistance of layer ~rith fixed parameters of technological process (time is 2-3 seconds) and voltages, or calculation of dependency of resistance of layer on minimal back bias and voltage on resistor with fixed technological parameters (time is 1-2 minutes). Storage size is 22K bytes. 568. Man'kin, I. A., Usherovich, B. L. and Kontorin, Yu. F., "Program for Calcula- tion of Nonlinear Conditions of LBV [Traveling Wave Tuve] on Chain of Connected Resonators," ELEKTRON. TEKHNIKA. SER. 1. ELEKTRON. SVCh, 1980, issue 7(319), pp 75-76, bibl, of 6 titles. Described is FORTRAN program for calculation of nonlinear output characteristics of LBVR (efficiency factor, gain factor, wave phase) in a band of frequencies as a function of geometric and electric parameters of instrument. Calculation time for three points of operating range of systems with numher of resonators not exceeding 20 does not exceed 15 minutes. 571. Mirgorodskiy, Yu. N.; Myachin, G. M. and Rudenko, A. A., "Program for Calcu- lation of Dynamic Characteristics of Silicon Planar Schottky Diode," ELEKTRON. TEKHNIKA. SER. MIKROELEKTRON., 1980, issue 3(87), pp 90--92, hihl. of 5 titles. Length of FORTRAN program is 550 conditional statements, reserved size of storage is 32 sheets of main storage. Calculation time of one point on pulse characteris- tic of planar Schottky diode is 20 seconds. 572. Mirgorodskiy, Yu. N.; Myachin, G. M. and Rudenko, A. A., "Program for Calcu- lation of Static VAKh [Volt-Ampere Characteristic] of Silicon Planar Schottky Diode," ELEKTRON. TEKHNIKA. SER. 3. MIKROELEKTRON., 1980, issue 2(86), pp 98-100, bibl, of 2 titlea. 575. Pasynkov, I. G. and Skoryupina, I. V., "Interactive Subsystem for Correcting Double-Sided Printed Circuit Boards Designed by Using 'Avtomat-2' System," Moscow, 1980, 13 pages, Preprint/IAE, No 3312/16, bibl. of 6 titles. "Correction" subsystemfor editing printed circuit boards in interactive mode using VT-340 type alphameric display. Program package is written in FORTRAN-GDR and MADLEN languages. 576. "Application of Method of Finite Elements to Calculation of Temperature Fields in Assemblies of Electronic Instruments," by I. M. Bleyvas, A. I. Zhbanov, V. S. Koshelev and V. N. Shevtsov, ELEKTRON. TEHI~NIKA. SER. 1. ELEKTRON. SVCh, 1980, issue 7(319), pp 72-74, bibl. of 1 title. Average time for solving one problem is one to five minutes. 21 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY 598. Dydychkin, A. Ye., "Method of Calculation of References of Typewritten Characters with Computer," VOPR. RADIOELEKTRON. SER. EVT, 1980, issue 8, pp 167-174, bibl. of 2 titles. Describes a FORTRAN program for automatic construction of references. Program run time is proportional to number of points approximating the character and ranges Trom 3 to 12 minutes. 599. Dydychkin, A. Ye. and Kalinin, A. V., "Modeli-~g the Process of Generating Typewritten Print," VOPR. RADIOELEKTRON. SER. EVT, 1980, issue 8, pp 161-166, bibl. of 4 titles. Method of reducing sample size for teaching and assessing validity of automatic reading machines based on using FORTRAN program to generate binary images of type- written print. b15. Yasenovets, A. V., "Software for Formatting and Editing of Text for Wang Mini- computer," in the book "Tipovoye proyektirovaniye vtoroy ocheredi otraslevykh - podsistem ASPR" [Standard Design of Second Phase of Sector Subsystems for Automated System of Plan CalculationsJ, Moscoc~, 1980, pp 108-111. 643. Litavrina, V. N. and Nikitin, S. A., "Language for Describing Topology of SVCh GIS [Super-High-Frequency Hybrid Integrated Circuits]," ELEKTRON. TEKHDTT:ZA. SER. 10. MIKFtOELEKTRON. USTROYSTVA, 1980, issue 5(23), pp 53-56, hihl. of 1 title. Language for coding SHF microcircuits is used in a system for automated design of topologies of integrated circuits and forms a part of the GRAF language for description of graphic information. 682. Kazansk~y, A. V.; Khaykin, A. P. and Trusenkov, S. T., "Application Program Package for Man-Machine Analysis and Forecast of Temperature of Surface Layer of - Ocean, TR./DVNII, 1980, issue 91, "Prohlems of Hydrometeorology of Eastern Siberi.a, the Far East and the Pacific Ocean," pp 39-43, bihl. of 9 titles. 683. Yukhnovets, B. P, and Burzayeva, Ye. I., "Checking Topology and Complex Patterns in Designing Hybrid Film Microcircuits," TR./NIITEPLOPRIBOR, 1980, collec- tion 89, "Technology of Manufacture and Automated Design of Transducers of Heat and Powe~ Parameters," pp 56-63, bibl. of 4 titles. riethod of checking topological patterns for minimally permitted dimensions has heen implemented in a system for automated design of film hy4~id microcircuits--SAPR "PLATA"--MSSP. Program run time when checking topology containing 400 points on the M-4030 computer is 15 seconds. 692. Tikhonov, Yu. V., "Software for 'INTRAS' Interactive Graphic System," in the book "Trudy MFTI. Seriya Aerofizika i prikladnaya matematika" [Transactions of the Moscow Physico--Technical Institute: Series Aerophysics. and Applied Mathematics], Moscow, 1980, pp 144 -146, b~ibl. of 3 titles. 2.2 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2047/02/09: CIA-RDP82-00850R000404060041-6 FOR OFF . 697. Afanasov, S. G.; Kochetkov, V. I. an~ Cherednik, V. I., "Second Zone of Oscillator with Brake Field," ELEKTRON. TEKHNIKA, SER. 1. ELEKTRON. SVCh, 1980, issue 7(319), pp 21-25, bibl, of 6 titles. Results of theoretical research of nonlinear 'interaction of electron flow witti high- frequency field of working interval of oscillator with brake f ield (GTP) in second zone of oscillation. System of equations for nonlinear unidimensional theory of GTP has been solved on the BESM-6 computer hy the Cauchy-Euler method. 701, Karimberdiyeva, S., "Some Studies o~ Daily Course of Components of Heat Balance of Active Surface," TR./SRED~TE,{~Z. REG10N. NII, 19$0, issue 68(149), "Prohlems of Hydrodynamics of Atmosphere," pp 32-42, bihl. of 14 titles, Describes algorithm for solving equation of heat balance on BESM-6. 702. Mazilov, V. N. and Kashik, S. A., "Application of Clastic Analysis in = Mineralogical Studies," in the book "Osnovnyye metody matematicheskoy geologii i resul'taty issledov~niy" [Basic Methods of Mathematical Geology and Results of Research], Yakutstc, 1980, pp 66-70, hibl. of 2 titlea. - 703. Mayorov, S. A. and Rudenko, A. A., "Method of Calculation of Dynamic Character- istics of MDP ~~iaximum nyiiamic Error] of Capacitor in Two-Dimensional Approximation on Large Signal,t� ELEKTRON. TEKHNIKA. SER. 3. MIKROELEKTP.023., 1980, issue 2(86), ' pp 79-83, bibl. of 4 titles. Example of calculation on BESM-6 computer of nonstationary distrihutior~s of potential and carriers of charge for solving problem of calculation of dynamic VAKh [volt- ampere characteristics] and VFKh [expansion unknown] of IrIDF of capacitor in two- dimensional approximation. 704. Mirgorodskiy, Yu. N. and Rudenko, A. A., "Method of Calculating Static Characteristics of Silicon Planar Schottky Diode," MIKROELEKTRON. I POLUPROVODNIKO- VYYE PRIBORY: SB. STATEY, 1980, issue S, pp 194-199, bibl. of 10 titles. Calculation of volt-ampere characteristics (VAKh) of Schottky planar diode is based on solution of boundary problem of nonlinear system of equations of continuity and Poissoa in two-dimensional area for the case when width of the area of space charge at S~hottky contact is small compared to characteristic linear dimension of diode. Cross-section of Schottky diode is covered by spatial network w ith number of assemhlies being about 1000. VAKh calculation time for Schuttky diode on BESM-6 computer is 8 minutes. 708. Vekshina, Ye. V.; Skorik, V, A. and Fursin, G, I., "Machine Modeling of Shift Register with Thyristors with Plasma Bond," ELEKTRON. TEKHNIKA. SER. 3. MIKROELEK- TRON., 1980, issue 2(86), pp 63-66, bibl. of 5 titles. Shift register with instruments with plasma hond (PPS) is base element for develop- ment of logic circuits, parallel adders, triggers and storage. Model of PPS for computer design of transistorized LSI circuits allows determining extremal para- meters of a shift register. ~3 - FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2447/02/09: CIA-RDP82-44850R444444464441-6 FOR OFFICIAL USE ONLY 710. Nechayev, A. M.; Rubakha, Ye. A. and Sinkevich, V. F., "Modeling Failures and Machine Tests in Studying the Reliahility of Pou~er SVCh [Micro~ave] Transistors," MIKROELEKTRON. I POLUPROVODNIKOVYYE PRIBORY: SB. STATEY, 1980, issue 5, pp 98-117, = bibl. of 16 titles. Discussed are machine models of main types of failures of power microwave transis- tors: simulation model of thermal breakdown and physico-statistical model of electrodiffusion failures. Up to 100 hours of BESM-6 computer operation are needed _ to impl~ment the machine tests of transistor structures. 772. Pogodin, A. Ye., Ways of Raising Computer Throughput," VOPR. RADIOELEKTRON. SER. EVT, 1980, issue 8', pp 131-134, bibl. of 3 titles. ~ Raising throughput of softc~are systems built with YeS computer facilities through organization and arrangement of data on 1rID [magnetic disk] storage units to reduce data access time. 792. Sunduchkov, K. S. and Leshchenko, Yu. I., "Automation of Synthesis of Structural Circuits of Low-Noise Receiver Amplifier SVCh [Micro~ave] Devices," ELEKTRON. TEKHNIKA. SER. 1. ELEKTRON. SVCh, 1980, issue 7(319), pp 64-69, bihl. of 14 titles, 796. Kirillov, D. A.; Naumenko, V. I. and Skorokhodov, 0. V., "Development of - Hierarchical Models in the 'Poisk-1' Information System," VOPR. RADIOELEKTRONo SER. EVT, 1980, issue 8, pp 3-9, bibl. of 4 titles. The Poisk-1 documentary data base management system supports efficient retrieval in large data bases. Langua~e for defining hierarchical structures is suggested and - language for manipulating systems is expanded. 810. Belyavskiy, Ye. D, and Gel'ner, V. V., "Numerical Analysis of Nonlinear Distortion of Shape of Nonperiodic Signal in LBV [Traveling Wave Tube]," ELEKTRON. TEKHNIKA. SER. l. ELEKTRON. SVCh, 1980, issue 7(319), pp 69-72, bibl. of 5 titles. Technique for strict (unidimensional) calculation with BESM-4 computer of shape of nonperiodic signal when it passes through LBV, on basis of which is made numeric analysis of nonlinear distortion of short pulses in LBV in nonlinear mode ~aith regard ~ to nonstationary effects. � 834. "Modular Construction of Software for Automated Process Control System for Production of Integrated Circuits," by V. M. Gudkov, V. I. Nadtochiy, A. L. Shabalin and T. P. Shvankova, ELEKTRON. TEKHNIKA. SER. 9. EKONOMIKA I SISTEMY UPR., 1980, issue 3(36), pp 29-31, bibl. of 3 titles. Modular software construction achieves invariance of it with respect to problems and ob~ects of control. Implementation of principles in system for monitoring quality - of process gases, developed on the base on the Elektronika-100I minicomputer, has shown their high efficiency. - z4 ' ;IAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL U: . 835. Molchanova, L. P. and Ostrovskiy, V. I., "Algorithm for ldentification of Process Defects in Microcircuits," ELEKTRON. TEKHNIKA. SER. 10. MIKROELEKTRON. USTROYSTVA, 1980, issue 5(23), pp 49-52, bibl. of 2 titles. Algorithm .~as been implemented in the form of two programs for the Elektronika-100I - cotnputer. Time for analysis of one microcircuit with dimensions of 1 x 1 cm is about 7.5 seconds. 836. Osetinskiy, L. G. and Osetinskiy, M. G., "Software Controlled Process System for Manufacture of Photographic Masks for Large-Scale Integrated Circuits," ELEKTRON. TEKI~TIKA. SER. 9. EKONOMIKA I SISTEMY UPR., 1980, issue 3(36), pp 23-25, bibl. of 4 titles. Organization of storage on ML [magnetic tape] of control information for generators of images--file OS [operating systemJ that supports control of photocomposition equipment from the Elektronika-100I minicomputer. 860. Astsaturov, R. M.; Nikolayeva, G. V, and Sigalov, G. G., "Study of Algorithms - for Control of Heirarchical Storage of Microprograms," VOPR. RADIOELEKTRON. SER. EVT, 1979, issue 12, pp 3-14, bibl, of 8 titles. Using analytic and simulation models, study is made of a numher of algorithms that support fetching microinstructions for execution from fast and slow section of control storage. 879. Samarskiy, A. S.; Pykhtin, V. Ya. and Vasilevich, V. V., "Estimating Complex Throughput of Computers by Analytic Methods," VOPR. RADIOELEKTRON. SER. EVT, 1979, - issue 12, pp 69-79, bibl. of 8 titles. Analytic model for estimating complex throughput, hasic elements of which are the processor and selector channels; procedure for calculating level of multiprogramming with regard to computer resources. Gives an example of calculating throughput for the YeS-1022, -1033 and -1035 medium class computers. 880. Sigalov, G. G, and Kur'yanova, N. I., "Estimating Computer Operation with Separation of Input-Output from Job Processing," VOPR. RADIOELEKTRON. SER. EVT, 1979, issue 12, pp 41-47, bibl. of 4 titles. Technique of using model of computer operation under conditions of separating input- output from process of job processing together with estimate of maximum possible factor of combining operation of devices for selection of numher of input-output devices, degree of multiprogramming and maximum size of job quc~ua on disk. 882. "Structure of Collective-Use Automated Design System," by A. V. Grekovich, A. A. Bozhko, A. B. Trofimov and N. A. F~derov, VOPR. RADIOELEKTRON. SER. EVT, 1979, issue 12, pp 53-59, bibl. of 2 titles. Principles for development of structure of INTE~t-SAPR system, prohlems of organiza- tion of exchange of application program facilities, trilevel principle of system coordination. 25 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY 887. Shklyar, V. B. and Yatsevich, P. F., "Model of Functioning of Heirarchical Storage of Microprograms," VOPR. RADIOELEKTRON. SER. EVT, 1979, issue 12, pp 15-19, bibl, of 2 titles. Probabilistic model has been built on basis of Markov model of random process. COPYRIGHT: Gosudarstvennaya publichnaya nauchno-tekhnicheskaya biblioteka SSSR (GPNTB SSSR), 1981 8545 CSO: 1863/246 ~ 26 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2047/02/09: CIA-RDP82-00850R000404060041-6 FOR OFFICIAL USE ONLY EXCERPTS FROM 'ALGORITHMS AND PROGRAMS', MARCH 1981 . Moscow ALGORITMY I PROGRAMMY in Russian No 3, Mar 81 pp 1-128 . [ExcerptsJ 920. "Territorially Distributed Multimachine Collective-Uss Computer Center," by G. I. Marchuk, Yu. I. Yeremin, G. I. Karpachev et al., Novosibirsk, 1980, 59 pages, (Preprint/Siberian Branch of USSR Academy of Sciences Computer Center, 245). 1036. Zarin'sh, A. Ya., "Noise Resistant Method of Filtration of Satellite Laser Observations," NAUCH. INFORM./AN SSSR. ASTRONOM. SOVET, AN LATVSSR. RADIOASTROFIZ. OBSERVATORIYA, 1980, issue 44, "Use of Satellite Observations in Geodesy and C,eophysics," pp 40-43, bibl. of 2 titles. 1076. Vitrichenko, E. A.; Prokhorov, A. M. and Trushin, Ye. V., "Metody izgotovleniya astronomicheskoy optiki" [Methods of Manufacturing Astronomical Optics], Moscow, Nauka, 1980, 196 pages, ahove the title: Academy of Sciences of USSR, Institute of Space Research, bibl.: pp 138-142. Describes the American CAOS and CCP and the domestic ZEBRA automated process sys- tems in the FORTRAN language. The Z~BRA system consists of two suhsystems: check- ing and technology. The checking aubystem has been implemented by the Hartman - method. Gives texts of programs of instructions for operator and checking examples. 1084. Mironov, N. T. and Bogatyrev, K. A., "Calculation of Ephemeris of Satellites for Observations with First-Generation Laser Ranger," ASTROMETRIYA I ASTROFIZIKA: RESP. MEZHVEA. SB./AN UkSSR, GL. ASTRON. OBSERVATORIYA, 1980, issue 42, pp 83-89, bibl, of 4 titlea. 1097. Trifonov, M. M., "Translator that Forms Programs of Base Elements of GIS SVCh [Microwave Hybrid Integrated Circuits] Based on Experimental Data," ELEKTRON. - TEKHNIKA. SER. 1. ELEKTRON. SVCh, 1980, issue 8(320), pp 67--69, bihl. of 7 titles. Describes FORTRAN program package that by matrices of scattering S or matrices of coctd~ctivity Y, specified in hand of frequencies, forms program of hase element in autocode of BESM-6 MADLEN. 27 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY 1098. Fialovskiy, A. T.; Zaytsev, S. V. and Mikheyev, A. G., "Calculation of Constants of Propagation and Diffraction Losses of Higher Type of Waves in Nonsymmetrical Band Lines," ELEKTRON. TEKHNIKA. SER. 1. ELEKTRON. SVCh, 1980, issue g(320), p 66, bibl. of 3 titles. Iteration algorithm with number of variations of field by thickness of dielectric basis greater than or equal to 1 in the FORTRAN language. Average calculation time of one variant is about 0.08 second. 1115. "Software for Subsystems to Process Images and Charts in Isolines and Numeric Data," Moscow, 1980, 68 pages, (ALGORITMY I PROGRAMMY/AN SSSR. BIEMS, GIVTS ASU - GEOLOGIYA, issue 6 (41)). - Described are FORTRAN programs for systems that process multispectral photogra~hs, halftone images, binary images of schemes of lineaments and erosion net~aorks, and charts of relief in isolines and numeric data. 1146. Batyrev, Ye. V.; Ocheretyannyy, A. N. and Sazonov, A. A., "On-Line Monitor- ing of Operating Capabilit~ and Diagnostic Checking of Systems for Synchronization of Checking and Measuring Systems," ELEKTRON. TEKHNIKA. SER. 3. MIKROELEKTRON., 1980, issue 4(88), pp 65-74, bibl. of 3 titles. Algorithm for checking efficiency of M-6000 computer Assembler language takes no more than 300 instructions, and execution time including printout is no more than 2 seconds. The algorithm for diagnostic checking has been implemented in the BASIC language, and execution time with printout of results is no more than 5 minutes. 1214. Rebrik, S. B., "Movement of Eyes and Processes of Organization of Sensori- motor Activi~:y," ERGONOMIKA. TR./VNIITEKFi~T.ESTETIKI, 1980, issue 19, "Studies of Functional Structure of Actuating Activity," pp 75-89. Studies to define the function2~ t�d~~o:~ship bet4~een groc~s��� of or~anization of instrumental activity and eye movement activeness accompanying it w.ere performed on the M-6000 computer. 1215. Sirotkina, Ye. B. and Apipova, A. I., "Study of Temporal, Preciseness and Spatial Characteristics of Movement under the Effect of Stresses," ERGONOMIKA. TR./ VNII TEKHN. ESTETIKI, 1980, issue 19, "Studies of Functional Structure of Actuating Activity," pp 38-43. Study of inethod of microstructure analysis with M-6000 computer to find oLt selective effect of inedicinal preparation of (kavinton) on structural components of whole activity. 1224. Grigor'yev, A. D.; Sipayev, S. A. and Yankevich, V. B., "Computer Design of Ring-Type Resonators," ELEKTRON. TEKffi~TIKA. SER. 1. ELEKTRON. SVCh, 1980, issue 8(320), Pp 66-67, bibl. of 2 titles. Described is an algorithm for calculation of working E-type of oscillations in ring- type resonators (coaxial) using condition of orthogonality of eigen functions. Results of solving test-problems on the BESM-h are given. 28 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 1237. "Operating Instructions for the CRJE System for Local Displays of a YeS-7906 Complex," by R. S, Triandofilidi, Moscow, 1980, 89 pages (METOL1 REKOMENDATSII PO ~ PROGRAMMIR. NA YES EVM/AN SSSR. TSEMI, issue 18), bibl. of 4 titles. Instruction on preparing jobs for the system, examples of use of job input co~mands, job editing during debugging, obtaining results of an executed ~ob and conversation with central computer operator. - 1242. Przhiyalkovskiy, V. V. and Lomov, Yu. S., "Tekhnicheskiye i programmnyye sredstva Yedir.oy sistemy EVM (YeS EVM-2)" [Unified Computer System Hardware and Software (YeS Comnuter-2)], Moscow, Statistika, 1980, 230 pages, bibl. of 21 titles. 1249. Andreyev, G, G. and Koptseva, N. N., "Machine Method of Smoothing Out , ~ Average Photo Tone over Area of Aerial Photographic Image," TR./GOS. N.-I. TSENTR IZUCH. PRIROD. RESURSOV, 1980, issue 10, "Some Results of Natural Resource Research Using Aircraft and Polygon Facilities," pp 38-41, bibl. of 1 title. Flowchart of program for YeS-1022 computer to correct photo tone of aerial photo- graphic images to eliminate distortions stemming from nonuniformity of function of ? distribution of illumination of lens. _ 1263. Arifov, P. U.; Zakirov, T. A. and Nazirov, A., "Construction of Multimachine System to Automate Experiments on Emission Elect.ronice," VOPR. KIBERNETIKI, UzSSR Academy of Sciences, Institute of Cybernetics and Computer Center, 1980, issue 110, VT V UPR.: SB. NAIICH. TR., pp 8-13, bibl. of 4 titles. Problems of developing a multimachine system based on the YeS-1022, the Elektronika K-200 and CAMAC standard equipment using the principles of unified bus system for exchange of information. 1273. Platonov, A. V., "Complex of Problems of Accounting of Traffic of Physical Assets," UOPR. RADIOELEKTRON. SER. ASU, 1980, issue 1, pp 57-63, bibl. of 3 titles. 1311. Vayda, F. and Chakan', A., "Microcomputers," translated from Hungarian, Moscow, Energiya, 1980, 359 pages, bibl. of 32 titles. (~rganization, principles of operation and fields of application of microprocessors Intel 8008, 8080, 3000, PPS25, PPS4, MCS4, and questions of designing microcomputers based on microprocessors. COPYRIGHT: Gosudarstvennaya publichnaya nauchno-tekhnicheskaya biblioteka SSSR (GPNTB SSSR), i981 8545 CSO; 1863/246 ~ 29 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2047/02/09: CIA-RDP82-00850R000404060041-6 FOR OFFICIAL USE ONLY EXCERPTS FROM 'ALGORITHMS AND PROGRAMS'~ APRIL 1981 Moscow ALGORITMY I PROGRAMMY in Russian No 4, Apr 81 pp 1-112 [Excerpts] 1462. Batalov, A. F.; Bobkov, V. A. and Ovchinnikova, L. S., "Program Package for Operating c~ith the UGD-43-1 Graphic Display," ELEKTRON. TEKHNIKA. SER. 1. ELEKTRON. SVCh, 1980, issue 4, p 112, hibl. of 4 titles. The BOB program package is designed for operating with the UGD-43-1 display, working under control of the Elektronika-100I minicomputer connected to a BESM-6 by communi- cation channel in DISPAK DOS. 1470. Bogdanov, N. B. and Kupenova, T. N., "Programs for Analytic Continuation of Holomorphic Functions beyond Linear," Dubna, 1980, 13 pages (SOOBSHCH./OIYAI, No R11-80-621), bibl. of 4 titles. NAC program package is written in FORTRAN-4 for the CDC-6500 (with conventional pre- cision) and text is given for the YeS an~ IBM computers (with double precision). 1486. Ashkenazi, I. D. and Korolev, A. N., "Program for Calculating Dispersion Characteristics, Wave Impedance and Width of Microband Line," ELEKTRON. TEI~iNIRA. SER. 1. ELEKTRON. SVCh, 1980, issue 1, pp 122-126, bibl. of 5 titles. Described is the FORTRAN program STRIPI for calculation of base elements in a system for machine design of microwave hybrid integrated circuits. Calculation time for one set of input parameters varies from a hundred microaeconds to three milliseconds with a known linewidth; with a known wave impedance, it is three to fivefold greater. 1488. Bunimovich, B. F. and Sharayevskiy, Yu. P., "Program for Analysis of Frequency Responses of Multisectional LBVO [expansion unknown]," ELEKTRON. TEKHNIKA. SER. l. ELEKTRON. SVCh, 1980, issue 1, pp 119-121, bibl. of 3 titles. Description of FORTRAN program to calculate time lag and coupling impedance of sec- tions of a slow-wave system and gain factor for LBVO with various values of geometric and electrical parameters, varying with the design. Calculation time for one fre- quency response depends on number of sections and number of discrete points within the frequency band (one frequency point is about one aecond). 30 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 1492. Kabakov, L. T. and Rukomoyeva, A. I., "Calculation of Coupling of ~ao Rectang- � ular Waveguides through a Rectangular Aperture in a Common Wide Wall of Finite Thickness," ELEKTRON. TEKHNIKA. SER. 1. ELEKTRON. SVCh, 1980, issue 1, pp 121-122, bibl. of 3 titles. The program has been implemented in the FORTRAN language. Calculation time for one alternative is no more than 20 seconds. 1494. "Program Package, 'Analysis of Linear Transistor Micro~rave Amplifiers'," by V. Yu. Yakovenko, S, V. Dvornikov, V. G. Volnistyy and N. N. Perestyuk, ELEICTRON. TEKHNIKA. SER. 1. ELEKTRON. SVCh, 1980, issue 2, pp 112-114, hibl. of 4 titles. The program package was written by using the FORTRAN-4 and Asserabler languages. Main storage size is 64K bytes. Analysis time for one alternative of a two-stage transistor amplifier at 10 frequency points is 2 to 3 minutes. 1496. Nemirovskiy, E., "Reed-Solomon Codes, Resistant to Phase Inversion in Channel," ELEKTRON. TEKHNIKA. SER~ 10. MIKROELEKTRON. USTROYSTVA, 1980, issue 3(21), pp 60-63, bibl. of 3 titles. DescriUed is a FORTRAN program that includes arithmetic of finite fields and takes about 1.5 minutes of machine time to check for the field GF (28). 1498. "Program for Approximate Analysis and Optimization of LBVO [expansion unknown] in Specified Band of Frequencies," by V. G. Borodenko, V. P. Kiryushin, A. S. Krasil'nikov et al., ELEKTRON. TEI~iNIKA. SER. 1. ELEKTRON. SVCh, 1980, issue 2, pp 105-110, bibl. of 9 titles. Program was compiled in FORTRAN. Calculation time ~t one point of band for electric length v= 10 is ahout 1 second. Problem solving time is 35 minutes. 1525. "Program Package for Sorting Special Type Inform.ation," by V. V. Vishin, S. A. _ Zaytsev, A. M. Maslennikov et al., ELEKTRON. TEKHNIKA. ;iER. l.. ELEKTRON. SVCh, 1980, issue 4, pp 110-111, hibl. of 2 titles. 1526. Maslennikov, A. M., "Programs for Checking Information Being Processed," ELEKTRON. TEKHNIKA. SER. 1. ELEKTRON. SVCh, 1980, isaue 4, p 110, bibl. of 2 titles. 1611. "Matematicheskoye obespecheniye gibidnoy vychislitel'noy sistemy 'Rusalka'. Gibridnyy assembler. Gibridnyy dialogovyy otladchik" ["Rusalka" Hybrid Computer Sys- tem Software; Hybrid Assembler, Hybrid Interactive Debugger], Institute of Control Problems, Moscow, 1980, 63 pages. The hyhrid assembler includes the M-400 computer ASS400A assembler and the hybrid superstructure that supports communication between the converter and interface de- vices and digital and analog parts of the GVS [hybrid computer system], and operates in the real-time mode. 31 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 � FOR OFFICIAL USE ONLY . . 1613. "Elektronnaya vychislitel'naya mashina M-4030: Posobiye pol'zovatelyu" [M-4030 Computer: User Handbook], by K. N. Belousov, N. V. Plakhotnyy, V. S. Mokhonchuk et al., Kiev, Tekhnika, 1980, 247 pages. Information needed for independent solution of practical problems on the M-4030 com- puter. Discussed are questions of M-4030 hardware, ~ob control, library utilities, main operator commands and system messages. Problem solving process is described in Assembler, FORTRAN and PL/1. 1615. "Computer Terminal Software, by A. N. Zakharova, L. S. Ovchinnikova, L. F. Pozdnyakova and T. N. Romanova, ELEKTRON. TEKHNIKA. SER. 1. ELEKTRON. SVCh, 1980, issue 4, p 113, bibl. of 4 titles. Purpose of the user station (AVP) is to organize input of user prob]:ems and sending them over communication line to BESM-6 for solving. User station includes the Elektronika-100I and various peripherals. ~ � 1617. "'TEST-3' Program for Checking Serviceability of BESM-6 Equipment for Communi- cation with Minicomputers by Seventh Direction through Computer Switch," by A. N. Zakharova, S. A. Zaytsev, Ye. I. Karmazina et al., ELEKTRON. TEKHNIKA. SER. 1. ELEKTRON. SVCh, 1980, issue 2, pp 110-111, bibl. of 7 titles. See abstract No 1662. 1618. Sakharchuk, V. I. and Serebrennikov, P. I., "Program for Checking Service- ability and Finding Defects in BESM-6 and 'Elektronika 100/16-I' Interface Harware: Engineering Test of Communications," ELEKTRON. TEKHNIKA. SER. 1. ELEKTRON. SVCh, 1980, issue 4, pp 105-106, bibl. of 1 title. 1619. Sorokin, A. G., "Subroutine for Drawing Text Characters," ELEKTRON. TEICHNIKA. SER. 1. ELEKTRON. SVCh, 1980, issue 4, pp 106-107, bihl. of 5 titles. Subroutine run time for BESM-6 for one access is independent of length of specified - character string and is greater than or equal to 5 seconds. 1620. Bryunin, V. N.; Bulatov, M. Kh. and Galitskiy, A. V., "Automation of Relia- bility Design of Microelectronic Apparatus with Redundancy," ELEKTRON. TEKHNIKA. SER. 10. MIIQtOELEKTRON. USTROYSTVA, 1980, issue 3(21); pp 90-102, bibl. of 9 titles. Automated system for assessing and optimizing reliability (ASON) of modeling and computation of theoretical functions of reliability of complex systems with various - types of redundancy has heen implemented on the BESM-6. Descrihed is YaONS special- ized language for describing and automating synthesis of mathematical reliability models of redundant structures. 32 . , APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2047/02/09: CIA-RDP82-00850R000404060041-6 1621. Dashkovskiy, A. A., "Panel Method Calculation of Flo~ about Wing Profile by Flow of Incompressible Liquid," TR./TsAGI, 1980, issue 2089, "Methods of Calculating Aerodynamic Characteristics of Profiles, Bodies of Revolution and Systems of Bodies: ~ A Collection of Articles," pp 3-18, bibl. of 3 titles. BESM-6 computer calculation time for a profile whose surface is subdivided into 100 panels is about 2 minutes; for 60 panels, time is 50 seconds; and for 30 panels, time is 16 seconds. 1623. I1'inskiy~ A. S, and Zarubanov, V. V., "Program for Calculating Distributions of Currents of Normal Waves of Nonsymmetrical Band Line (MSTRIP)," ELEKTRON. TEKHNIKA. SER. 1. ELEKTRON. SVCh, 1980, issue 1, pp 126-128, bibl. of 5 titles. Progr.am for calculating distributions by band of longitudinal and cross currents of basic and higher types of nonradiating waves of partially screened nonsymmetrical band line. Obtaining all graphs took ahout 2.5 minutes of BESM-~6 central processor operatlon. 1624. Oleynikov, V. I., "Technique for Calculating in Three-Dimensional Approxima- tion Nonlaminar Elliptic Electron Beams in Transit Channels of 0 and M Types of Microwave Instruments," ELEKTRON. TEKHNIKA. SER. 1. ELEKTRON. SVCh, 1980, issue 1, pp 51-61, bibl. of 8 titles. _ 1628. "Machine Design of Electronic Optical Systems Using Optimization Method," by - I. I. Golenitskiy, I. B. Grishina, T. P. Kutsevskaya et al., ELEKTRON. TEHI~NIKA. SER. 1. ELEKTRON. SVCh, 1980, issue 4, pp 98-105, hibl. of 6 titles. Examples of solving problems based on a software package that includes a program for analysis of two-dlmensional electronic optical systems (EOS), a standard program for optimization for the BESM-6 and developed program modules that interpret user language in which problem is formulated. 1641. Takhtamyshev, G, G., "'Tvist' Modeling Program," Duhna, 1980, 8 pages, ~ (SOOBSHCH./OIYaI, No 1-80-640), bibl. of 12 titles. Description of program for determining characteristics of magnetic spectrometers and checking programs of geometric reconstruction on CDC-b500 and YeS-1040 computers. The program traces tracks, belonging to the physical process, through the spectro- meter and forms a unified array of all tripped wires. Program size (together with user programs and package for histograms) is 552008 words. 1643. Shemyakin, V. V. and Krasil'nikov, D. P., "Experimental System for Acoustic Emission Monitoring Based on YeS-1010 Computer: MO [Software]," Moscow, 1980, 24 pages, (Preprint/IAE, No 3333/15), bibl, of 4 titles. 1653. Rabinovich, Ye. B.; Sinitsyna, G. B, and Feynberg, V. Z., "Determining Optimal Parameters of Complex Mark of Matching of Drawings of Integrated Circuits," ELEKTRON. TEKHNIKA. SER. 7. TEKHNOLOGIYA, ORG. PR-VA I OBORUD., 1980, issue 1(98), p~ 73-76, bibl. of 6 titles. M-222 computer was used to obtain table of optimal parameters of complex marks that are used in settings for automatic matching of IC drawings. 33 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2047/02/09: CIA-RDP82-00850R000404060041-6 FOR OFFICIAL USE ONLY 1662. Zakharova, A. N. and Ovchinnikova, L. S., "Program 'Test of Complex of On-Line Graphic Station'," ELEKTRON. 'TEKHNIKA. SER. 1. ELEKTRON. SVCh, 1980, issue 1, pp 118-119, bibl. of 1 title. On-line graphic station (POGr) has been developed on hase of Elektronika computer with display and connected to BESM-6 and Videoton VT-340. Program is intended for checking serviceability of POGr hardware and communication line between the Elektronika-100/I and BESM-6 computers. Program is written in SLENG-3 autocode. 1663. Pozdnyakova, L. F., "Program for Checking Serviceability of Card Input Unit (UWK-601) in the 'Elektronika-100I' Computer," ELEKTRON. TEKHNIKA. SER. 1. ELEKTRON. SVCh, 1980, issue 4, p 109, bihl. of 1 title. 1679. Lazutin, Yu. M., "Struktura matematicheskogo ohespecheniya displeynoy stantsii, sopryazhennoy s EVM BESM-6" [Structure of Softu~are for Display Station Interfaced with BESM-6 Computer], Moscow, 1978, SO pages, (Preprint, USSR Academy of , Sciences, IPM [Institute of Applied Mathematics], No 110), bibl. of 7 titles. Describes satellite computer software, part of display system software; system con- sists of main BESM-6 computer that performs large volume of calculations and the small SDS-910 satellite computer connected to it by communication channel; graphic display is connected to the SDS-910. To increase display system throughput, infor- mation coming from the BESM-6 and peripherals is fiuffered, and requests for opera-- tion with the devices are queued. Information is reduced to form ready for use by the FORTRAN program. COPYRIGHT: Gosudarstvennaya publichnaya nauchno-tekhnicheskaya bihlioteka SSSR (GPNTB SSSR), 1981 8545 CSO: 1863/246 34 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY ABSTRACTS FROM 'APPLIED COMPUTER SCIENCE', NO 1, 1981 Moscow PRIKLADNAYA INFORMATIKA in Russian No 1, 1981 (signed to press 20 Feb 81) pp 229-230 METHOD OF MATRIX IDENTIFICATION AND GOAL ANALYSIS OF AUTOMATED DESIGIr' SYSTEMS IN INDUSTRY [Abstract of article by V. N. Vasil'yev and V. N. Lashkov] (Text] L. Zadeh's conceptsof fuzzy sets are used in the work to derive an algorithm for identification of elements of source information. ~VALUATION OF QUALITY OF OPERATIVE-CALENDAR PLANS ~Abstract of article by A. N. Klimov] [TextJ Method of checking plans worked out to perform construction work using production-technological models is descrihed. AUTOMATIOI3 OF PLAIVI3ING OF TECHNOLOGICAL PROCESS MONITORING SYSTEM USING LOCAL CRITERIA OF OPTIMIZATION [Abstract of article by Yu. N. Shchepin and V. I. Ostrovskiy] [TextJ Prohlem of sfting of monitoring operations in a technological process is solved according to criterion of average cost of item. The developed method achieves reduction in number of consfdered alternatives of monitoring plan based on Iacal er3teria of optimfzation. - INTERACTIVE IMAGE-TYPF DATA BASE MANAGII~IENT SYSTEM (WITH EXAMPLE OF 'ACCOUNTING' PROBLEM) ~Abstract of article by V. I. Shakhmatov, A. A. Senchuk, B. A. Vorob'yev, M. Yu. Kamstcitskiy, V. P. Timoshchenko and T. N. Bol'shakova] EText} Tnteraetive automated system for creating and updating data hases on mini- computers is described. Example of wage calculation is given. 35 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000400460041-6 FOR OFFICIAL USE ONLY D~SCARTr,S LANGUAGE AS SOURCE LANGUAGE FOR SPORA SYSiEM [Abstract of article by I. 0. Babayev, F. A. Novikov and T. I. Petrushina] [Text] SPORA [Specialized Software for Work in Astronomy] is a software system for work in the applications pragram field that includes archives (data bank), a number of PPP [applications program packages], monitor and translators for the basic pro- gramming languages. The system is developed for the BESM-6 but is easily transferable to another machine. F-LANGUAGE, FORMALISM FOR REPRESENTATION OF KNOWLEDGE IN INTELLIGENT INTERACTIVE _ SYSTEM [Abstract of article by V. M. BryabrinJ [TextJ Formal language for representation of knowledge in an interactive system is described; it is based on user communication in natural language. This formalism has been implemented within the bo~nds of the DILOS [Dialogue Information Logical System] system at the USSR Academy of Sciences Computer Center. - PARMA, NETWORK DATA BASE MANAGEMENT SYSTEM BASED ON CODASYL RECOI~AIENDATIONS [Abstract of article by V. P. Kosso, I. Ye. Kuznetsov and T. N. Sumarokova] - [Text] Described are the architecture, functional capabilities and service facili- - ties of the network SUBD [data base management system] developed in 1976-1979 on - the basis of CODASYL [Conference on Data Systems Language] recommendations (YaOD [Data Description Language] of 1973 and YaI~ [Data Manipulating Language) for Cobol in 1975). DSP DATA PROCESSING SYSTEM [Abstract of article by V. A. Ivanov, V. A. Kondratenko and V. I. Buyanov] (Text~ Descr_ibed is a data processing system oriented to data processing in an ASUP [automated production control system]. Given are basic system functions, principles of implementation as we11 as the system language. Some evaluations of system use are given by results of industrial operation. VIRTiIAL STORAGE ACCESS METHOD IN YES OS OPERATING SYSTEM [Abstract of article by G. N. Staroverova and T. V. Korotayevaj [Text] Described are data set structure, capabilities of access method and data set service program with VSAM [virtual storage access merhodJ organization. 36 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000400460041-6 FOR OFFICIAL USE ONLY QUESTIONS OF COMPLEXING OF PROGRAMS IN YES OS IN ASSEMBLER AND PL/1 LANGUAGES USING MULTITASKING FACTLITIES ~.Abstract of article by V. M. Malyshev] [Text] Author discusses problems arising for progratmners in joining subroutines in the Assembler and PL/1 languages into a common program for execution of it in YeS - OS. IMPLEMENTATION OF APL ON YES COMPUTERS [Abstract of article by I. I. Malashinin and A. I. Kononov] [TextJ Described is implementation of APL interactive interpreter on terminal system for YeS computers. APPLICATION OF BINARY TREES TO ORGANIZATION OF LARGE FILES OF INFORMATION (SURVEY) [Abstract of article by V. A. Yevstigneyev] [Text] Author ~urveys foreign publications on organization of informat~on and _ searching in it by using binary trees. MUSSON MULTIMODEL DATA BASE MANAGEMENT SYSTEM [Abstract of article by M. R. Kogalovskiy, V. V. Kogutovskiy, K. I. Makal'skiy and M. M. Vinogradov] - [Text] Authors discuss main concepts of implementation of the multilevel data base management system MUSSON [Multilevel Self-Describing General-Purpose System], its architectural features and interfaces to external level. Major concepts of the _ conceptual model of data and the mechanism of maintaining it are considered. COPYRIGHT: Izdatel'stvo "Finansy i statistika", 1981 8545 CSO: 1863/229 37 FOR OFFIGIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00854R000440060041-6 FOR OFFICIAL USE ONLY PARMAy NETWORK DATA BASE MANAGEMENT SYSTEM BASED ON CODASYL RECOMMEENDATIONS Moscow P:;IKLADNAYA INFORMATIKA in Russian No 1, 1981 (signed to press 20 Feb 81) pp 104-117 [Article by V. P. Kosso, I. Ye. Kuznetsov and T. N. Sumarokova from collection of - articles "Applied Computer Science", edited by V. M. Savinkov, Izdatel'stvo "Finansy i statistika", 15,000 copies, 232 pages] [Excerpt] Initial Specifications The PARMA data base management system (DBMS) was developed at the NIIUMS [Scientific = Research Institute of Control Computers and Systems] (in Perm') in the period 1976 to 1979. Development was based on the concepts for building data bases (BD) worked out by the Conference on Data Systems Language (CODASYL). The versions of the Data Description Language (YaOD) of 1973 [1] and the Data Manipulating Language ~ (YarID) for Cobol of 1975 [2] were taken as bases; Cobol and Assembler were taken as the languages included in the system. In developing the DBMS, decisions made in similar foreign systems were taken into account and additional sources were used _ [3]. English mnemonics were kept for all language facilities. Since CODASYL developments are rather well known [7], special terms are used below without strict definition. The DBMS is designed for operation in the environment of YeS OS, versions 4.0 and higher with any configuration of the operating system except RSR; control tables of - YeS OS declared in versions 4.0 and 4.1 are used as the monitor. The system is programmed in Assembler and has about 80,000 instructions. About 400 tracks of 5050 ZUPD or 200 tracks of 5061 ZUPD, respectively, are required for the DBMS libraries. Data bases may be created on crnnputer systems with 256K or more of main storage. Additional facilities are provided by an interface with the "Kama" system which provides the capability of operating in the teleprocessing mode; in this case, 512K of main storage is required. COPYRIGHT: Izdatel'stvo "Finansy i statistika", 1981 8545 CSO: 1863/229 38 - -FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2047/02/09: CIA-RDP82-00850R000404060041-6 DSP DATA PROCESSING SYS~EM Moscow PRIKLADNAYA INFORMATIKA in Russian No 1, 1981 (signed to press 20 Feb 81) pp 117-135 [Article by V. A. Ivanov, V. A. Kondratenko and V. I. Buyanov ~rom collection of articles "Applied Computer Science", edited by V. M. Savin~COV, Izdatel'stvo "Finansy i statistika", 15,000 copies, 232 pages] _ [Excerpt] The L`SP system is an applications program package oriented to pragramming and organization of data processing in the sphere of production control. Purpose of the DSP System DSP operates on YeS computers under control of the operating system (OS) and supports the full technological cycle of data processing: creation and maintenance of a data base, including input and update, conversion of structures and types of data, monitoring of input data and changes; search and extraction of data according to specified conditions of selection; arithmetic and logic processing of data; and _ generation of responses (documents) with rather complex structural and sophisticated design components. The following factors were taken into consideration in developing the DSP system. 1. To one extent or another, the specific facilities for implementation of the capabilities listed above are offered by the DBMS's that are more widespread in our country. However, the advantages of a DBMS are achieved to the full extent only when the direct access storage size connected aimultaneously to the computer system is sufficient to hold the entire data base. The now widespread diak storage units with a capacity of 7.25 Mbytes for medium and large systems are not very suitable. Thus, the extensive distribution o� YeS computers, on the one hand, and the lack of supporting them ~zith direct access storage units with a large capacity (29, 100 and 200 Mby.tes), on the other hand, maintain the relevance of the current use of magne- tic tapes as media for storage and processing of data in computer systems based on YeS computers. 2. Maintenance of continuity of management systems created (transferred) for the YeS computers with systems on the high-volume second-generation computers--the "Minsk-32." 39 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY 3. Necessity of creating simpler facilities, compared to existing ones, for man- computer communicati~n which ~rill allow enlisting for practical programming of ASUP [automated production control system] problems lot~-skilled specialists (including even non-programmers), reducing personnel training time and prot~lem programming gnd debugging time. 4. Providing the capability of direct use of data stored on magnetic tapes for the "Minsk-32" computers together with data in the YeS computer formats ~tithout preliminary conversion. COPYRIGHT: Izdatel'stvo "Finansy i statistika", 1981 8545 CSO: 1863/229 !~0 � , APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY , VIRTUAL STORAGE ACCESS METHOD IN YES OS OPERATING SYSTEM Moscow PRIKLADNAYA INFORMATIKA in Russian No 1, 1981 (signed to press 20 Feb gl) pp 135-145 [Article by G, N. Staroverova and T. V. Korotayeva from collection of articles "Applied Computer Science", edited by V. M. Savinkov, Izdatel'stvo "Finansy i statistika", 15,000 copies, 232 pages] [TextJ In this article, we discuss the basic principYes of operation and service capabilities of a new method of accessing data on direct access devices develflped _ within the framework of the YeS OS operating system. This method of access, called the virtual storage access method (VSAM) is used in the virtual method of operation of the YeS OS (SVS mode) and achieves a high rate of data processing. All earlier access methods existing in YeS OS are designed for direct access devices oriented to processing data in computers with a limited amount of main storage. Therefore, the entire data base management system, and access methods in particular, were oriented to minimizing the size of main storage required to support a specific type of access to achieve a sufficiently high rate of data access. The physical record--the smallest processing unit for a direct access device--was adopted as the unit of exchange between external (storage on a direct access device) and main storage. To access a record, the programmer used the address of the physical record on the device in the form of the cylinder and track address; there- fore, for certain types of processing, the programmer was tied to the physical - characteristic: of the devices. The access methods developed earlier (BSAM, QSAM, QISAM, BISAM, BDAM and BRAM) are possible in YeS OS for the MI~'T and MVT modes and are intended for operat3on of computers without virtual storage. In operating systems with virtual storage, the computer user has a considerably larger amount of storage to solve his problems; therefore, the capability of orga- nizing a new method of access--virtual (VSAM)--has emerged. In this method, the unit of exchange between external and main storage is the control interval consist- ing of a certain number of physical records specified by the programmer. Physical records with a length of 512, 1024, 2048 and 4096 bytes are used in the VSAM method. Within the limits of a data set, the physical record length is fixed, urhich allows more efficient use of sector organization of direct access devices. The user of the VSAM access method actually has nothing to do with the physical records; he works with either the logical records or the control interval. Therefore, a new method of addressing records is used--by the relative displacement of the heginning 1t1 FOR OFFZCIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00854R000440060041-6 FOR OFFICIAL USE ONLY of a record from the first record in the data set in bytes. This method is easier for the programmer to understand and is absolutely independent of the physical characteristics of the direct access device. A four-byte value, designated an RBA [relative byte address], is intended for addressing, and this method of addressing allows access to any record in the data set. The address of the first record in a data set equals 0, the address of the second equals the length of the first logical record, etc. The record format in a data set with VSAM organization differs from that of other access methods. The prograuimer in general does not specify whether his records will be fixed, variable or undefined length since the structure of the control interval allows determination of the address of the next logical record. The control inter- val consists of a certain number of logical records and their descriptors. Several control intervals are grouped into a control area, a larger unit of storage of a data set. Data set records are key- or entry-sequenced. In the key-sequenced data set, the computer user can obtain access to a record by its key or by the relative byte address (RBA). If a data set is entry-sequenced, access to the record is possible only by the RBA. A key-sequenced data set has two components--an index and data. The data component contains the logical records themselves, while the index component establishes the correspondence between the record keys and their addresses. The index has a multi- level structure. For the index component, the length of a physical record always equals the length of a control interval. The data component and the index component (if it exists) describing it are called - the structure group. The structure group for a key-sequenced data set consists of the index and data components, while that for an entry-sequenced set has only the data component. Organization of a data set in ascending key sequence is similar to index sequential organization, while that in entry sequence is similar to sequential and direct organization. In index sequential organization that provides access to a physical record by its key, hardware facilities in the direct access devices are used to search by record key for one greater than or equal to the key. From this stems the necessity of = positioning records in sequence of ascending physical addresses on the medium. - When new records are added in the middle of the data set, the later records are transferred from the tracks to an overflow area so as to not disrupt the sequence of ascending addresses. In VSAM organization, no keys of physical records are used at all and the possibili- - ty emerges of more efficiently including new records in the data set organized in ascending key sequence. If there is free space in the control interval, the record is inserted in the space needed; if there is not enough free space for the ne~z record, the control interval is divided in half, half of it is utritten into a new control interval, and the new record is inserted in the needed space. If there is no free interval in the control area, the control area is also divided in half and half of its control intervals are transferred to a new control area. In the process, 1~2 . , APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00854R000440060041-6 FOR OFFICIAL USE ONLY only the indexes have to be copied, establishing correspondence between the record keys and their new addresses. Thus, the overflow areas inherent to index sequen- tial organization do not occur, and all the shortcomings of index sequential organ- ization are eliminated. In later versions of the VSAM access method, there will be a capability for multi- key structure of a data set, when in addition to a key for the data set, additional keys may be inserted without recopying the data records, but only after building additional indexes that establish the correspondence between the new keys and addresses of the records. The VSAM method provides for both sequential and direct processing of records, and ~ also implements all the capabilities offered 'oy the other access methods for direct - access devices. For data sets organized in ascending key sequence, VSAM allows obtaining direct access to a record by its key or address, adding a new record, erasing a record and changing record length. For data sets organized in entry sequence, each new record is added at the end and insertion of records in the middle, changing record length and ohliteration of records are not permitted. The VSAM method allows sequential, partially sequential or the direct type of data processing, and at the same time combining the different types of data set process- - ing is permitted. In sequential or partially 3equential processing, the programmer may specify the beginning of processing in the form of a key or address of a record, and the key may be incomplete. Partially sequential processing is used for a data set organized in ascending key sequence when not all records in the data set have to be processed, but the required records can be requested in ascending order of their keys and retrieval at the index of the lower level, beginning with the key of the preceding record, can be performed faster than direct retrieval of the key beginning at the index of the higher level. Sequential processing of a data set may be performed by key and address. The VSAM method in operating with data sets supports one language of access--macro- instructions for building control blocks and obtaining from them information and macroinstructions for processing requests. The macroinstructions for building control blocks are: ACB, used to build the access control block during program translation; EXLST, designed to build an exit list for user programs during program translation. One can specify an exit to a program for processing logical input--output [IO] errors, a program for processing IO errors associated with equipment functioning, a program for checking access permission and a program for recording events; RPL, which organizes the request parameter list during program translation; GFNCB, which builds the request control block, a list of exits or a list of request parameters during program execution; 43 FOR OFFTCIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY MODCB, used to modify the access control block, exit list or request parameter list during program execution; SHOWCB, used to read the fields of the access control block, exit list or request parameter list during program execution; and TESTCB, designed to test the fields of the access control block, exit list or request parameter list during program execution. Request processing macroinstructions are used to write, read or erase data records in the data set. The main macroinstructions in this group are: GET for reading data records; PUT for writing data; ERASE for erasing data records; POINT for setting at the record needed; CHECK for suspending execution of processing before completion of request; and ENDREQ for ending the request. To establish a link between the proceasing program and a specific data set, the programmer has to build in his program an access control block (ACB) which in the VSAM is used instead of the data control block (DCB) used by other access methods. In the access control block, the name of the DD statement of the job control language is assigned, and in the DD statement when.the job is started, the data set name that will be processed is indicated. Before starting processing of a data set, the ACB must be opened using the macroinstruction OPEN and closed at the end of processing by using the macroinstruction CLOSE. Access to indexes in a data set organized by ascending keys is performed in one of twc ways: the structure group is opened and the macroinstructions G~TIX (read index) and PUTIX (write index) are used; one index component is opened and the macroinstructions for ordinary data processing are used (GET, PUT, etc.). Processing of one index component is identical to processing of a data set in entry sequence. The index component itself has no indexes and therefore may not be processed by the access method by key. The user may not duplicate or replace the index processing performed by VSAM during ordinary processing of data records. To provide for standard operation with data sets, inatead of a large number of utility programs used for operating with data sets of a different organization, VSAM has a multifunction utility program (AMS).which has its own function request language. AMS defines VSAM data sets, loads records into them, converts sequential or index sequential data sets into VSAM format, lists the VSAM catalog or data records, copies data sets, creates copies of catalog and data set, facilitates data set portability hetween operating systems and erases data sets. ~1~4 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R400440060041-6 ~ . The programmer requests the desired function of the multifunction utility program using instructions of this program and their parameters. The AMS program instruc- tions are divided into two types: functional and modular. _ The main functional instructions that define the operation to actually be performed are: DEFINE for organization of VSAM catalogs, catalog elements describing a data set, and zones of external storage on direct access devices that will be used by the VSAM access method to allocate storage from them for data sets; DELETE for deleting elements in a catalog; ALTER for changing elements in a catalog; LISTCAT for listing catalog elements; PRINT for printing a VSAM data set, as well as index sequential data sets; REPRO for copying data sets, converting sequential and index sequential data sets to VSAM format, converting a VSAM data set and index sequential data set to a - sequential data set, organization of backup copy of catalog and restoration of the catalog from the copy; VERIFY used in detection of an error when the data set is closed to guarantee - the correctness of elements of the catalog that describes the data set; CHKLIST defines the volume of magnetic tape for the checkpoint program; EXPORT for creating a backup copy of a VSAM data set and preparing a VSAM data set and user catalogs for use in another operating system; and - II~ORT for restoring a VSAM data set from its copy and for providing the capabil- ity of using in another operating system the VSAM data sets and user catalogs pre- pared using the EXPORT instruction. The functional instructions DEFINE, ALTER, DELETE, LISTCAT, REPRO and PRINT may be used for processing data sets with an organization differing from VSAM. The modular instructions allow creating conditions for execution of functional pro- grams. The main instructions of this type are: sequence of instructions IF THEN ELSE which check the completion code of the preceding functional instruction and as a function of its value bypass or in- clude execution of the sequence of ~unctional commands; SET changes the completion code; and DO END defines the start and end of a sequence of instructions. The completion code is set when each functional instruction is executed and has the following value: 0-- complete e~ecution of the instruction; 4-- instruction was executed, but a warning was made about the peculiarity of the e:cecution; for example, the required element was not found in printing a catalog - element; ~-~5 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY 8-- instruction was executed, but some of the requested functions were not execu- ted, for example, in deleting a catalog element, the required element was not found; 12 instruction not executed due to logic error; and 16 error in execution, which causes g halt in execution of the entire stream of instructions. This code may be set, for example, when the system could not open a data set for system output or a permanent IO error was detected in system output. In the IF instruction, one can check the actual completion code of the last instruc- tion, designated LASTCC, or the maximum permitted value of the completion code, designated MAXCC. An example of the use of the instruction sequence IF THEN ELSE is: IF MAXCC ~ 8 THEN DO instruction sequence END ELSE DO instruction sequence END The multifunction utility program may be executed as a job or job step, and also may be called from another program. Each data set that is to be processed by the VSAM access method must be defined in a catalog by using the multifunction utility program, i.e. catalog elements must be created for it that describe the occupied zone, name and other characteristics. In the VSAM access method, catalogs are used as the basic source of information on VSAM data sets and the volumes on which they are located. There are two types of catalogs--master and the user catalog. A master catalog is required for VSAM; user catalogs are optional. Each VSAM catalog contains records defining it. System catalog contents point to master catalog contents, and the contents of the latter to user catalogs. A catalog includes information on the space occupied by a data set, indicates data set access restrictions, contains usage statistics and l_inks the RBA byte address ' to physical addresses. All data sets located on a volume must be cataloged in one master or user catalog. A data set is defined only in one catalog. The master catalog may not be located on the main resident volume for the system since if it is damaged it would not be possible to restore it. If a VSAM data set is cataloged in a user catalog, this catalog must be assigned by the statement DD JOBCAT or STEPCAT. The STEPCAT statement defines the catalog used in executing this job step, and the JOBCAT statement in executing the whole job. If several catalogs are needed, they are described as ordinary concatenated data sets in YeS [Unified Series] OS. 46 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIE . The use of user catalogs raises the reliability of system operation. Catalogs may be located on each volume instead of one common catalog. This reduces the number of accesses to a given catalog to the minimum and if some catalog is destroyed the others are still available. When a catalog is created, the programmer may specify a special facility that makes it possible to restore it. In this case, a special area is allocated on a volume for records that contain the information required to restore the catalog when there is a system or equipment malfunction. VSAM obtains storage on a volume and manages it in units called zones. A zone occupies a certain number of tracks and cylinders. One zone may have one or more data set~, and in turn a data set may occupy several zones. Zones are described in a VSAM catalog. A cat3log manages the volumes containing VSAM data set defined in this catalog. All VSAM structural groups and all zones on a volume must be cataloged in the cata- loK that manages the volume. A tag is set in a volume table of contents to indicate that this volume is managed by a VSAM catalog. The latter contains the volume entries for each volume that it owns. Each entry describes the characteristics of the volume, all zones and objects of. VSAM that use storage on the volume. Only one VSAM catalog may own a volume. Volume management using a VSAM catalog does not affect data sets on this volume that have an organization different from VSAM. These data sets may be located on a volume managed by a VSAM catalog and may be cataloged not in this catalog. An OS catalog can be considered an ordinary data set and it may be located on a volume managed by a VSAM catalog. To free a volume from management by a VSAM catalog, one has to get rid of all VSAM objects on the given volume by using the multifunction utility program. Then the DELETE SPACE instruction of the multifunction utility program i~ used to delete all zones and remove the entries of this volume from the catalog and remove the tag of ownership of this volume from the table of contents. A user may process a VSAM catalog as a data set or in a special way. To process a catalog as a data set, a user has to link his processing program with this catalog and use the VSAM macroinstructions (GET, PUT, etc.) for processing. To obtain access by the method special for the catalog, a programmer has to use the special multifunction utility program for data sets (AMS). Thanks to the use of the multi- function utility program for processing a catalog, the capability of performing the following functions is available: creating a catalog (instructions DEFINE MASTERCATALOG and DEFINE USERCATALOG); including records in a catalog (DEFINE); changing records in a catalog (ALTER); deleting records in a catalog (DELETE); and listing the records of a catalog (LISTCAT). 47 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 - FOR OFFICIAL USE ONLY Catalog information pertaining to a specific data set may be obtained hy a process- ing program hy using the macroinstruction SHOWCAT irrespective of ~hether or not the given data set is opened. Information is issued from the catalog specified by the DD statements with the names STEPCAT or JOBCAT, or from several catalogs if the given DD statements define concatenated data sets, or from the master catalog if the JOBCAT or STEPCAT statement is missing or a catalog description is not found in them. To obtain access to a catalog as to a data set, it has to be described by a DD state- ment of the job control language (not JOBCAT or STEPCAT), an access control block (ACB) has to be built for it in the program and it has to be opened by the macro- instruction OPEN. Catalog records may be accessed by key or address. In this opera- tion, it is recommended that catalog records only be read and updated. Since cata- log records have a complex interrelationship, including new records or deleting records from a catalog is not recommended. To add or delete catalog records, it is - recormnended that the multifunction utility program be used. VSAM affords a greater capability for restricting access to data sets compared to other access methods. A programmer may restrict access to structural groups, data or index components, VSAM catalogs and catalog entries for data sets. The capability exists to specify the following levels of data access restrictions: restriction of any access which requires confirmation of permission for execution of - all operations (reading, updating, adding and deleting) with a data set, any index and records of the catalog that correspond to this data set; restriction of access by control interval; restriction of access for updating. This allows reading, updating, adding or delet- " ing of records in a data set only with confirmation of permission for execution of these operations; and restriction of access for reading. This allows reading of data and catalog records with confirmation of permission, but does not allow adding, changing or deleting them. To restrict access to a data set, one must also restrict access to the catalog in which the given data set is defined. Each higher level of access restriction allows execution of a lower level operation. When YeS OS requests permission for access to a data set from the computer operator, a code can be assigned to be used instead of the data set name, so that the computer operator will not know the data set names to which access is restricted. One can specify the number of attempts which a computer operator may use for permission of access to a data set. T_n addition to the standard check for permission to access a data set, VSAM allows a programmer to use an additional check for access permission hy using a user pro- gram. This program is called after YeS OS obtains permission for the given type of access. The VSAM access method permits a programmer to specify permission to access a data set in his program which frees him from a message to the computer operator with _ 48 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 . ~ information on his data sets. VSAM requests access permission from the computer operator only ~zhen access permission is not specified in the program. A great advantage~of the VSAM access method is the simplicity of its use. The pro- grammer has to be familiar with just several macroinstructions and has to indicate a minimum of information on the data set in the program and job control language (substantially less than for other YeS OS access methods). For example, in the DD statement for VSAM, one has to specify only the set name (DSNAME) and its disposi- tion (DISP). The programmer does not have to describe record formats or know the physical organizatlon of the data set; hia programs are device independent and he need not be concerned about blocking of records and bufferization. This reduces outlays for programming and the possibility of errors. The VSAM method exceeds the capabilities of all YeS OS access methods used for direct access devices in providing faster access to data. Therefore, in systems with virtual modes of operation, it will replace all other access methods except the library. Since many programs have now been developed with index sequential organization of data sets, the VSAM accesa method includes programs called ISAM interface with VSAM that permits a program written by using the ISAM access method in fact to process a data set with VSAM organization. All that is required are minor changes in the job control language statements and a description of the data sets by using the multifunction utility program. If there is a good algorithm of randomization for a data set with direct organiza- tion, the use of BDAM may he retained. * * ~ Since the experience of operating with other access methods was taken into account when the VSAM access method was developed, certain shortcomings and inconveniences of the other access methods were eliminated in the design of VSAM. The process of writing reenterable programs carrying on processing of data was made easier; the capability of different simultaneous processing of data sets was introduced, which eliminates the need of repeatedly opening and closing a data set when switching from one form of processing to another, for example from direct processing to sequential; new capabilities for protecting data sets were introduced; and access to information on access restriction is not authorized for even the systems programmer and computer operator, but only for the programmer himself. Therefore, the VSAM access method should find extensive application in data processing systems using the YeS OS operating system in the virtual mode of opera- tion. COPYRIGHT: Izdatel'stvo "Finansy i statistika", 1981 8545 CSO: 1863/229 49 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY IMPLEMENTATION OF APL ON UNIFIED SYSTEM OF COMPUTERS Mosc~w PRIKLADNAYA INFORMATIKA in Russian No 1, 1981 (signed to press 20 Feb 81) pp 155-170 [Article by I. I. Malashinin and A. I. Kononov from collection of articles "Appiied Computer Science", edited by V. M. Savinkov, Izdatel'stvo "Finansy i statistika", 15,000 copies, 232 pages] [Text] The extensive attraction to using computers of specialists in various fields of knowledge has required offering them convenient and on-line facilities for man- machine communication. This pertains to both language and the form of intercourse. In this article, we discuss questions of implementing a terminal system with the APL language which has become widespread abroad and which affords direct user access to allocated com~uter resources through terminals and software facilities. The simplicity o~' the language syntax, brevity and precision of express.tons, power- ful facilities for operating with multidimensional arrays (vectors, matrices, etc.), and the capability of on-line computations and creation of programs (functions) make APL attractive for a broad range of users [1, 7]. This article focuses on APL notation and use of terminals. Also covered are the principles of building the systems and operating it under the control of YeS OS. System Concepts The APL system implements time sharing among a group of remote or local terminals. The terminal is an electric typewriter or console typewriter with a control unit, such as for example in APL ~ 360 [5]. In the APL system, a user exchanges information with the system only through a user - terminal. All statements entered by the user from a terminal are divided into sys- tem commands and APL statements. All information output to the user corrsists of system messages, computation results and program texts [4, 5]. Time sharing is implemented by the system supervisor, and the APL interpreter iden- tifies and processes tnformation coming in from the terminals. The system operates under the control of YeS OS of the MFT or MVT type and takes up one partition or zone of main storage. The API, system functions the following way. Upon connection to the system, the user is allocated workspace of fixed size de- fined at system generation. This space can be named and kept in the user's _ 50 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY personal library. It can subsequencly be called into the active zone for further work. The active zone includes part of main storage and storage on direct access devices used for the swapping [svoping] process. In addition to persor?al librarie.s, the capability exists to create common libraries accessible to all system ~iserg, Storage extents on disk units designed in the standard of sets of Yes OS are used to store user libraries and organize swapping of active work areas. By using a special utility, there is the capability of creating transportable copies on magnetic tape of individual work areas and libraries for storage or transferring them to other installations. Users in the APL system work independently of each other while under supervision of - the APL operator. The latter, in addition to the usual user capabilities in the sys- tem, is given privileged facilities to control the system. The APL system provides an adequate 1eve1 of security and reliability. The APL Language APL (A Programming Language), developed by K. E. Iverson (3], is designed for engi- neers and scientific and technical calculatians although its field of activity is not limited to these areas. The language is unconventional for professional compu- ter users, but easily mastered by nonprofessionai computer users since it is almost , devoid of -thase auxiliary details that any user has to know to solve problems on ~ computers (7, 8]. The wealth of language symbols minimizes terminal input to execute particular opera- ~ tions. APL offers more than 40 built-in functions which substantially eases programming. An important place in this language is allocated to facilities for working with vectors and matrices of numbers and symbols. There are facilities for operating with text constants. _ The mode of immediate interpretation allows the user to use the terminal as a calcu- lator, and the program definition mode allows input, correction and reproduction of a program text. ' In executing programs (functions), APL allows following the course of their execu- _ tion and stopping it. I.n the APL system, there is the capability of writing interactive problems. The main set of functions o�~fer.ed by APL are called primitive functions. Based on them, it is possible to create any number of additional functions; thus, the language is easily expanded. * Shirokov, F. V., "Basic Concepts of APL," in the hook "Algoritmy i organizatsiya resheniya ekonomicheskikh zadach" [Algorithms and Organization for Solving " Economic Problems], Moscow, Statistika, issue 14, 1980. 51 - FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000400460041-6 FOR OFFICIAL USE ONLY A distinguishing feature of APL is the simplicity of data representation and the absence of specifications in defining a variable, since the value assigned to it . characterizes the variable. In addition, there is no traditional priori~ization of operations in APL expressions--all operations are of equal priority, and an expres- sion is computed sequentially from right to left. The number of inessages to the user has been held to a minimum and does not exceed 40. The APL Language for the Unified System of Computers The main difficulties in using APL for the YeS computers are the lack of a domestic - terminal with the proper keyboard and means of communication. Therefore, in using APL on domestic equipment, there arose first of a11 the problem of changing its symbulic notation. � The symbol of the most widespread version of APL is described in Backus normal form the following way [6J: APL symbol letter ~ number - ~ special symbol ~ blank - letter A~a~~~D~E~f~G~H~I~~IKI L~MIN~O~P~QIRISIT~UIV~ W X Y~Z~ A~B C~D~E~F~G~H~I~J~K~ L~M~N~O~P~QIR~S~T~U~VI~I~I IZI . number ~ ~ � 0~1~2I3I4I5I6I~I8I9 special symbol ~~~~~-~~w~~~V~~~"-~ ~ ~~'~x~ ?IWIEIPINI ~ I~ Iz IoI~I~IFI�Irl~l-Ivlol~lolc i~I~ I~ I=I=I - nlulllTl i I; I: I~I , I/I~I~I~I~I~I~I~IoIBI~!' II I~Ifl~l~ blank : . = vacant position Thus, the APL alphabet consists of 131 symbols. All APL symbols are divided into simple and combined (ligatures). Simple symbols are input from the APL terminal by one press of the keys, combined--by printing a simple symbol over another. _ When input from the terminal keyboard, each APL symbol is converted into a single value in the internal representation of the APL system, from which it is recoded ~ into the external representation only when output to the terminal. The APL terminal keyboard contains f38 simple characters (table 1). ~ The majority of the special symb~ls have a specific functional purpose. The symbols C~ ~ W a do not detine internal functions of the APL interpreter and are intended for subsequent implementations of APL. The illegal symbol obtained - by a combination of. the characters 0, U and T has a special meaning. The combinat.ion symbols include: ABCDEFGHIJKLMIVOPQRSTUVWXYZ m G~~1~ ~ O 8~! _~f'~ ~ ' input of whicli is effected by entering the first simple symbol, backspacing and entering the second simple symbol (table 2). 52 ~ ~ APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR U~rICIAL USE ONL'~C Table l. Correspondence between Simple Symhols of APL Language and Its Modifica- tion for Unified System of Computers and YeS-7077 Type Terminals DKOI Replacement Main Use of Symbol in APL APL Symbol Symbol Name Symbol Expressions - A- Z letters of alphabet A- Z identifiers of work areas, functions (programs) and - variables 0- 9 numbers 0- 9 numeric values double period none symbol - superscript ~ sign of negative number ~ less than ~ operation of checking for correspondence to "less than" ~ less than or equal none operation of checking for correspondence to "less than or equal" = equal = operation of checking for correspondence to equality ~ greater than or equal 6[B] operation of checking for ~ correspondence to "greater than or equal" ~ greater than ~ operation of checking for correspondence to "greater - than" # not equal xC [Zh] operation of checking for correspondence to inequality = V logical OR 1i [L] logical "OR" operation logical AND H[I] logical "AND" operation - minus - subtraction operation - divide % division operation + plus + addition operation X multiply multiplication operation ? question ? generation of random numbers c~ omega none symbol E epsilon 9[E] determination of coincident elements of two vectors or matrices p rho ~b operations with dimensionality - of vectors and matrices ~ tilde u [Ts] logical "NOT" operation T up arrow ro[Yu] selection of vector elements 53 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-04850R000400060041-6 FOR OFFICIAL USE ONLY DKOI Replacement Main Use of Symbol in APL APL Symbol Symbol Name Symbol Expressions y down arrow none dropping of vector elements ~ iota fl[Y] generation of sequence of integers, determining coordi- nates af elements of one vector in another O circle @ trigonometric functions and operations with pi * asterisk * raising to power operation - right arrow bl [Y] jump to marker or exit from program left arrow 3[Z] assignment of value to variable a alpha none symbol r maximum T[Gj finding largest of two quantities L minimum ~1 [Ch] finding smallest of two quantities underline see table 2 p nabla ~y (U] beginning and ending of function (program) definition 0 delta A[D] see table 2, tracing of program ~ jot ~ outer product of matrices ' apostrophe ' definition of character string U quad II [P] request for data input into program ( left parenthesis ( changing order of computation in APL expression ) right parent hesis ) [ left bracket [ enclosure of subscripts ] right bracke t ~ (coordinates) of elements of a vector or matrice c subset none symbol ~ inclusive set none symbol ~ intersection A [Ya] symbol U union none symbol 1 base IiI [Sh] recoding of vector into scalar T top iq [Shch] recoding of scalar into vector 54 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00854R000440060041-6 FOR OFFICIAL USE ONLY DKOI Replacement Main Use of Symbol in APL APL Symbol Symbol Name Symbol Expressions ~ stile ! absolute value or finding residue (of a number modulo another) ; semicolon ; index separator; mixed output separator : colon label separator in program \ backward slash ~[F] "expansion" of matrix across. last coo~'dinate / slash / "compression" of matrix across last coordinate , comma , "merging" of matrices, ravel- ing of matrix into vector - , period , separator of whole and frac- tional part of a number, generalized product of matrices When a combined character is entered, actually three code combinations are sent to the channel: the first symbol, the backspace and the second symbol. The same occurs when a combined symbol is output to the terminal. In addition to the codes for the set of simple APL symbols and the [BSJ (backspace) code, the APL interpreter distinguishes the code combinations between the special keys [LF] (line feed) and [CR] (carriage return) which are used in correcting text - and completing entry of information from the APL terminal. The majority of YeS computer terminals based on the "Konsul-260.1" typewriter con- tain the following set of DKOI symbols [2]: The Latin alphabet: ABCDEFGHIJKLMNOPQRSTUVWXYZ; the Russian letters that differ fror~ the Latin: brAxcs~i~tnny~bu~ulu~3~oR; [BGDZhZ IYLPUF' Ts ChShShchEYuYa]; the numbers: 0 1 2 3 4 5 6 7 8 9; and the special symbols: +-#%*l?~1&~~)/_:;~~~-~,. space Based on this set of symbols, the notation for the YeS APL language was built for the YeS-7077 type terminal. S5 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY Table 2. Correspondence betuzeen Combined Symhols of APL Language and Its Modifica-- tion for Unified System of Computers and YeS-7077 Type Terminals CuMBnnbi A/IA Cun~Bonbi n~odu~uyupo d N020 ' AJb/KU A/!A 6 C(,(o(cmaen.v KOM !~N!!- /YOM !/N!/QO . OC/I/Q /lA~!!{UE Main Use of Symbol in APL Language BOqb/ uM- CUMQO/76 ~ n0u Ob~00dC ~~OOe nii npu 1 3 4' _S O * ~ ~ @ ~ logarittun function t t t ' factorial function, combinations function . . . . v~, ~ n~ u [L & Ts] logical NOR operation _ n~, ~ ~ N$ u [I & Ts] logical NAND operation - ~ _ ~ ~ ~ $ _ "compression" across first coordinate ~ _ ,t ~ ~ $ _ [F & "expansion" of matrix across first coordinate ~ ~ ~ ~ @& 1 rearrangement of matrix elements across ~ first coordinate O- O ~ ~n rearrangement of matrix elements acroas last cQOrdinate O ~ ~ ~ @ & ~ [@ & FJ rnatrix transposition ~ ~ ~ ~ & I . [D & ascending sort ~ ~ ~ y& ~ ,[U descending sort B ~ n&�~ .[P matrix inversion, solving a system ~ � of linear equations ~ ~ n~' [P literal input into program � ~ . A d ~ [Ya & comment line in program 0~, ~ ~ y~~ [U & Ts] lock to inhibit printout of . program text T 1 I ll~ W,d W. [Shch & Sh] system dependent functions nv~v. noay~ . . a~~d- an~d~ m: underscored alphabet - oum: q_ A A Ad_ _ g_ & B~_ identifiers of variables _ z _ ~ Z d _ Cne . cneq. sFecial symbol cuni~on cuMBon symbol �or re~ection of literal input into 0 U T m 0 b U dr T program Key: 1. Component symbols 4. Component symbols for input 2. Combined symbols 5. APL symhols 3. Combined symbols when output 6. Symbols of modified APL language 56 FdR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000400460041-6 FOR OFFICIAL USE ONLY Table 3. Correspondence b.et~een APL Symbola and Symbols on the Keyboard of the "Video~ton-340" Display I CuMeon CUM~O%l CU~OO/1 C(LM~OA ~~0/1 ~ nnuif C MOav Cr[Mdan~2 A/ l A d t~ c n n e, v A A ~ u c ~ e A c A/tA' Sucnne �BUdC~ M~tldCO- N 1 NBILdCO- ~ ~ 1~ m N- 340" ~ 1) maN-340" ~ 1~ mny-J40 maH-940" A A A A d_ � c ~ [F] e d B Bd_ . 3)-(NaBx) ~ ~ [Ya] C C ~ Cd_ < < ~ U D D D d_ S E E ~ ~ � E _ = "I 1 -yr ~ ~ Shch ] F F ~ f F d_ Z 6[ B] I G ~ G 66_ > > ~ . N H ~ H h_ ~ * ' v n [L] ~ ~ ~ J J ~ J d- N lIJ K K .11 K 4_ . - - : C L ~ L d,_ 9~0 � t M M ' Md_ + + ~ ~ ~ N N ~ N d_ X # . P P P P~_ ' W ~j E ' L& Ts] Q Q Q~4_ E f H&u fI & Ts] ~ ~ \d i s s ~ s a_ P i~ ~Ts~ @a . T� T T d_ K3 [Yu] 9 d- V ~ V V V~= o Vl fY] ~ j [D w W W W_ @ ~ ~ U~ ~ - Y Y ~ Y -~r ~ [Y] d : . P & r ~ ~ ? ~ d_ [Zl ~ Ad I ~a&& ~ rd- - r~ tc~ ~ yd ~~fU~~ & T~~ _ W [ Shch & Sh J o ~~,~~~eer.) y ~Ch] (5) pI~I d 2 2 v Y(UJ (6) Om~'a~ d~BEoJ input 4 ~ o. (Dl p ed,BoQ output 5 5 ? fl ~P~ ~ ~ B~oB' input 6 6 < < output ~ ~ ~ ~ B6iDod 8 8 C C ~ ~ BQnd input 9 9 : : npodtn OtltPllt Q Baeod Key: 1. APL Symbol 5. uppercase 2. "Videoton-340" display symbol 6. loFrercase 3. superscript 7. space 4. underline 57 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000400460041-6 FOR OFFICIAL USE ONLY In doing so, these were the guidelines follo~ed: - keep the visual perception of the modified language program close to that of the APL program; wherever possible, give a meaningful value to the special symbols; keep the combined symbols in the language; expand the language with the capability of using the Cyrillic alphabet; and keep the continuity of the libraries of programs written at existing installations in the standard APL language. Table 1 shows the correspondence of the simple APL symbols to the symbols in the modified language; table 2 shows the correspondence between the combined symbols and their component elements. From these tables it is evident that no change was made to the ma~oirty of symbols coincident in graphic design. The YeS APL language contains all the Russian let- ters, which allows use of them in literal conatants and comments, but in other cases they have a special meaning. Thus, the letter t'( [P] ("pechatayte" [type!]) is used to request data input into the program; bI [Y] ("vYyti" [to exit]) is used - to jump to the marker or exit the program; 3[Z] ("zanesti" [to make an entry]) is used to assign values to variables; and H, f~ and [I, L and Ts] are used to define the logical operations AND, OR and NOT ("otritsaniye" [negation]) [i = and; ili = or]. ' The combined symbols in the language are implemented the following way: On input, one types the triads of symbols of the first sim~le symbol, the symbol & instead of the backspace and the second simple symbol; on output, the symbol & is replaced by the backspace code and the actual image of the combined symbol is printed out. The correspondence between the APL combined symbols and those used in the YeS APL modified language is shown in table 2. To generate the carriage return code (since no code combinatiun is generated when the carriage return key is pressed on the "Konsul-260.1"), the key is used which does not cause printing of the symbol, but sends the corresponding symbol code to the computer [2]. The APL symbols w.~ a c~ ~ < have no equivalent in the modified language, but since they are rarely used, their absence has aLnost no effect on the capabili- ties of the language. When a program written at another installation is printed out, these symbols are replaced by blanks. The modification, discussed above, to the APL notation was also used in implement- ing the display version of the terminal system. In doing so, the restrictions men- tioned were eliminated and the language was given a number of important additional facilities for communication with a terminal. In this case, the terminal used can be either the YeS-7066 display from the YeS-7906 complex, connected to the multi- plexer channel, or the "Videoton-340" display. 'Phe "Videoton-340" as an APL Terminal Interest in using the "Videoton-340" display as an APL terminal arose due to a num- ber of factors: high output rate, no mechanical elements, high quality, broad 58 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 capabilities in text editing and input-output management, standard set of Latin and Russian characters and special symbols, large information capacity of the screen, and the capability of printing information from the screen on a special alphameric device. In addition, the technical capabilities of the display allow using it at a : considerable distance away from the computer without special software or hardware for communication [9]. - The main ob.stacle to using the "Videoton-340" displays (or the "Videoton-YeS-7168") with YeS computers heginning with the model YeS-1020 and up is the impossihility of connecting them directly to the computer channel. This required development of a special group switch that allows connection of up to 48 displays to the YeS computer multiplexer channel. In doing so, the instruction set corresponding to the YeS-7077 console was selected as the most suitable; this afforded continuity ~ith the APL implementation for the YeS-7077 and compatibility with the sufi~et of instructions for the YeS-7906. The group switch is the control unit for the displays connected to it; it also con- verts j.nformation coming in from DKOI code into KOI-7 code for the "Videoton-340" unit and vice versa. Since the APL system does not use a YeS OS access method, terminals engaged with the APL system are not distributed by the OS scheduler and do not require description when OS is generated. However, it makes sense to descrihe them anyuiay as devices of the YeS-7077 or YeS-7066 type with addresses of the multiplexer channel from the interval of O50 to 07F. Such substitution allows employing software, which uses the subset of instructions for the YeS-7077 and YeS-7066 devices, for operation with the "Videoton" displays. Thus, when a"Videoton" is described as a YeS-7077, it can be defined as a YeS com- puter operator console, and if described as a YeS-7066, it can be used as a terminal for the interactive systems "Kama," SRV [time-sharing systemJ, SPOK [system of pro- gramming of teaching courses] and certain others. The "Videoton" display keyboard connected through the group switch allows transmis- sion of 99 code comhinations. These are the codes of the 26 Latin letters, the 20 Russian letters drawn differently from the Latin, the 10 numbers, 28 special symbols including the space, and the 15 control symbols for the t~rminal. The APL language notation for YeS-7077 type terminals, shown in table l, was revised to account for the capabilities of the "Videoton" terminal and became that shown in tahle 3. The modified version of the language was given the working name of APLVIDEO. Since it is not possible physically to display the combined characters on the screen, they are represented in APLVIDEO by triads of symbols both on input and output. Some of the APL symbols were replaced by the symbols (code combinations) of the ter- minal control keys that are shown in table 3 as rectangles with the corresponding letters. For convenience of operation from the terminal, a numher of the APL control symbols are interpreted differently on input and output to the terminal. Thus, the symbol 59 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY [BS] for the lower case of the prototype terminal at input is represented by the character & or the [F---J of the "Videoton"; ~hen this symbol is output, the back- space situation is simulated: the cursor is shifted to the left one position which is equivalent to using the key. The symbol "blank" corresponds to pressing the special key [ J on the display key- board. It is shown as a space on the screen. However, when correcting function definitions, the key must be used; pressing this key generates the "blank" code _ too, but in doing so the symbols above the cursor are not replaced by spaces: the entire line being corrected remains visible while the cursor is moved below it to the position to be corrected. Features of User Operation from the "Videoton-340" Terminal in the APLVIDEO System The user operating procedure from the "Videoton" terminal is somewhat different from user operation with a special APL terminal. The display is ready for operation in the system when it is in the ONLINE status. To put the display in operation, the keys ~pOWER] and ~ ON J are used. [LINE] Information input operations from the display depend on whether a given terminal is a user terminal or an APL operator terminal (when only one terminal is connected in the system, it is automatically considered the APL operator terminal). A user begins text input directly at the place indicated by the cursor. But the operator has to first send an interrupt to the system using the key [RETURN] and only after this can he begin input of text. In both cases, information input has to be ended by sequentially pressing the keys [LINE] fETX]. [FEED] This mode of terminal input can be called direct input. In the process, each symhol that appears on the screen is sent through buses to the computer channel. Thus, it is not possible to correct input text by replacing ~nd inserting. However, APL offers its own capabilities for correction of text just input if the input end " signals [LINE] ~ETX] have not yet be~n sent to the system. For the display, this [FEED) capability is used as follows: When an erxor in text is detected, the cursor is moved by the [F--] key to the first incorrect symbol from the left; it is then erased by the [DC] key; then the corrected remainder of the text is again input from this position. In addition to direct input, using the "Videoton" display as a terminal for an APL _ system allows composing text in the offline mode. For this, the display is placed OFFLINF by pressing the appropriate key. Then text may be composed starting from any line. In the process, one can make use of all the offline editing facilities by using the keys [E-~j [-~j] [ ] ( ] [ERASE] [DLJ [IL] [DC] [IC], and the symbols corresponding to them are not considered part of the text. Text composition is ended by the combination [LINE] ~ETX]' ~hen the cursor is set under the first line [FEED] of text and the [SEND] key is pressed. This operation causes sequential reading of the composed text and sending it into computer storage; in doing so, the display is ~ put into the ONLINE state. When the symbols [LINE] are encountered in the cumposed [FEED] 60 . , APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000400460041-6 ~ FOR OFFICIAL USE ONLY text, the first such symbol in the line, reading from the left, is considered the end of input from the given line. To correct programs by APL facilities using the "Videoton," the keys (--1] and [y~ or (DC] are used. The cursor is moved by the key to the position to be cor- rected for insertion of the necessary symbols in the line. Pressing the [DC] key right after the number of the program line to be corrected causes this line to be removed from the program text and the line numbers to be renumbered. In correcting lines containing combined symbols, remember that they occupy three - positions on the screen, but just one position as a language symbol; consequently, in moving the cursor, it is necessary to bear in mind the actual number of APL symbols passed over by the cursor. Output to the screen from the system starts, as a rule, at the position where the cursor is set. In doing so, the amount of output may exceed the space remaining to the end of the screen, and also in general may exceed the size of the whole screen. In this case, after the last line of the screen is filled, output continues on the first line, wiping out the information already on the screen. The following facilities are available to prevent this. First, the [ROLLJ key can - be used to shift the text up a line when the last line on the screen is filled. In the process, the very top line on the screen is lost. This procedure is convenient with small amounts of input-output, for example when working with an interactive program in the "question-answer" mode. Second, the information on the screen may be printed out on the special "VT-343" ATsPU [alphameric printer] or one similar to it by pressing the [PRINT] key (the display is placed OFFLINE), and then the screen is _ erased by the (ERASE] key in expectation of output. And third, when programming problems that give output, output size limitations can be considered and the output control facilities discussed below can be used. Expanding the Capabilities of Programs for Input-Output Control Since certain "Videoton" display control symbols are fully legitimate symbol5 in the APLVIDEO language, they may he used in text constants and programs, which yields new capabilities for controlling input-output and graphic programming. Let us discuss several examples. In the examples, we use the following text varia- hles to control the screen to whictt values have previously been assigned: ERASE - erase screen and set cursor at home; IiOME - set cursor at beg~nning of screen; LF - line feed (blinking symbol J); F.TX - end of text (blinking symtaol C); CD - cursor down one position; CU - cursor up one position; CL - cursor left one position; and t;K - cursor right one position. 61 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY Example 1. Output text and erase screen [3= Z= synbol for specification or assignment; see table 3] TXT 3'*---SAMPLE OF OUTPUT TEXT---*' ERASE, TXT in the process on a clear screen on the first line will appear the information: *---SAMPLE OF OUTPUT TEXT---* Example 2. Output text at screen line desired without erasing screen [Russian y= nabla, function definition symbol; Russian b= rho, shape/reshape; see table 3J y LINE N [1] HOME - return cursar to home position; [2] (N-3)bCD - move cursor down needed nunnber of lines taking automatic line feed into account; [3] TXT - output of text; y NE 8 - c.all of function which outputs value of variable TXT on line 8 - of screen. Example 3. Input of text on any screen line and output with storage of line feed [16 = blank] POINTER 3' 16~5161616 ~---INPUT NAME' input pointer LIST 3" empty vector of list of names y INPUT N [1] ERASE [2] (N-3) b CD [3] POINTER, CU [ 4 ] NAME 3 [ (1 = Q ) [5] LIST 3 LIST,NAME,LF y Comments on statements: POINTER and LIST are variables defined in the mode of immediate execution of APL; N is the screen line number, reading from top, on which will be output the phrase POINTER and in the first position of which wi11 be input the name NAME; 1- 2 setting cursor at needed line; � 3 output of text-pointer of input and compensation for automatic line feed; 4 input of text information in field on screen indicated by pointer and entering of it in NAME; 5 adding to list of names by next name with storage of line feed. With repeated operation of the program, a list of names is filled in that can be output in the form of a column of names of operations: LIST 62 F()R OFFI(;IAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY Example 4. Input of information from display screen using SEND mode Let SCREEN be a text variable containing several lines of information which have to be corrected to suit the job requirements. The program CORR will make it possible to do this operation: CORR "CREEN y CORR TX;SC [Y = nabla,~7 ] ~i~ sc 3 Tx f3 = F-- ] [2] ERASE,SC,LF,ETX (3] HOME [4] 12 b CD [b = rho, P ] [5] `161d~PUT DISPLAY OFFLINE, PERFORM CORRECTION AND SEND FROM FIRST LINE',HOME - ~6) sc 3 t'~&' (ri&' = l'~ ] Tx sc [8] ERASE,TX _ y Comments on statements: 1 - copy text information into temporary variable SC; 2 - output text, which is te:ct-sample to be corrected, to screen; 3-5 - output message on line 16 and set cursor home; 6 - stop program for input. At this time, display is put OFFLINE, offline correction is made, and then screen is read in SEND modP; 7 - return value to initial variable; 8 - check reproduction of corrected variable on screen. Example 5. Random walk of cursor over screen starting in its cenier Y RANDCR N; I; CONTRL; COMPAS [~t = V] ~ 3 0 [3 = E-] (2~ CONTRL 3 " ~ [3] COMPAS 3 CL,CR,CU,CD [4] CONTRL 3 CONTRL,COMPAS[?4],'+' [5] I 3 I+1 [6 ] bI 44t~i (I ~ N) [bI = -~j; = X (multiplication) ; Iri = (7] ERASE;'NUMBER OF MOVEMENTS OF CURSOR=';N iota] [ 8 ] HOME [9] 7 b CD [b = rho] [10] (40 b'-'),CONTRL Y Comments on statements: 1-6 - form control vector CONTRL which includes symtaols for cursor movement, and character for indication of its positions on screen; 8-10 - set cursor in screen center and control walk. Implementation of Working Version of APL System The APL system in the initial version was implemented on a YeS-1022 computer ~ith YeS OS in the MFT mode. A YeS-7077 console was used as the system terminal; during operation with APL it was not used as the computer operator console. 63 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY The display version of the system was implementEd with the group switch and 15 "Videoton-340" terminals. The APLVIDEO system operates under YeS OS control in the MVT mode. In both cases, the system uses one priority partition (MFT) or region (MVT) with a size cf 180K. The YeS-5050 and YeS5061 devices have been used as ex- ternal storage. In doing so, for the YeS-5061, 2 extents of sets were used for user libraries with a size of 20 cylinders each and 8 cylinders for the swapping support set. Practice of operating with the APL system oriented to the YeS-7077 terninals has shown that these devices can be used for APL without alteration. Ho~re~er, the existing restrictions in the modified version of the language could be reduced if use of the keys VSh (backspace), NS (new line) and PS (line feed) caused generation of the appropriate code combinations. When operating with 15 users, the APLVIDEO system in the majority of cases provides a response time on the order of 3 seconds with a 36K-workarea for each user; in doing so, use of processor time is negligible and the capability exists for paralle? _ batch processing of user jobs for YeS OS. In addition to those mentioned, other poasible terminals for YeS APL are the YeS- 8570 subscriber stations connected by communication lines to a data transmission multiplexor. References 1. Grenander, U. and Faryber~er, V., "Kratkiy kurs vychislitel'noy veroyatnosti i statistiki" [Short Course in Computational Probability and Statistics], Moscow, Nauka, 1978. 2. Petrova, Ye. P. and Saplin, M. S., "Uetroystvo obmena informatsiyey operatora s protsessorom" [Operator-Processor Communication Device~, Moscow, Statistika, 1976. ~ 3. Iverson, K. E., "A Programming Language", New York, 1962. 4. Katzan, H. J., "APL User's Guide", New York, 1971. ~ - 5. APL ~ 360 User's Manual, IBM, H2O-0683, 1969. 6. Everett, F. A., "A Formal Definition of APL Statement Syntax", APL-75 Congress, Riza, 1975. _ 7. Gilman, L. and Rose, A., "APL: An Interactive Approach", Moscow, Mir, 1979. 8, Pratt, T., "Programming Languages", Moscow, Mir, 1979. _ 9. Ketkov, Yu. L., "Programmirovaniye na BEYSIKE" [Programming in BASIC], Moscow Statistika, 1978, pp 113-116. COPYRIGHT: Izdatel'stvo "Finansy i statistika", 1981 8545 CSO: 1863/229 64 ' ~ USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00854R000440060041-6 F )FFICIE ~E ONLY APPI,ICATIONS UDC 658.815:681.32.06 SYSTEM FOR MONITORING METAL PRODUCT DELIVERIES TO MAJOR NATIONliL ECONOMIC CONSTRUCTION PROJECTS Moscow MATERIAL'NO-TEKHNICHESICOYE SNABZHENIYE, SERIYA 4: PRIMENENIYE MATEMATICHES- KIKH METODOV, VYCHISLITEL'NOY TEKHNIKI I ORGTEKHNIKI V MATERIAL'NO-TEKHNICHESKOM SNABZHENII in Russian No 8, Aug 80 (manuscript received 14 May 80) pp 8-9 [Article by I. B. Ayngorn, deputy department chief, USSR Gossnab Main Computer Center] [Text] The task "Mo.iitoring of Metal Product Deliveries for Construction of Enter- prises for Ferrous Metallurgy, Enterprises for Production of Fertilizer, and Facilities Constructed on the Basis of Compensation and Integration Agreements" has been developed by the USSR Gossnab Main Computer Center and the NIISU [expansion unknown] and put into industrial operation at the USSR Gossnab Main Computer Center. Product ncmenclature now monitored includes four classes of inetal pr~ducts; rolled ferrous metal products, steel gipes, iron pipes and industrial metal goods. Data on the amounts of resources actually r_eceived are sent mont?~ly by union republic gossnab information centers and those of the main territorial admir~istrations of USSR Gossnab from the construction sites to the USSR Gossnab Main Computer Center _ over communication lines. The "Minsk-32" computer is used to receive the data from the communication lines and monitor teletype messages, while data processing is done on the YeS-1040 compu- - ter. Three data bases have heen. built witti the YeS-1040 computer, maintained by the "Oka" ~BMS: "Catalog," "Projects" and "Ministries." The "Catal.og" data base holds permanent data for the task, "Prajects" hoids planned and actual data on metal product deliveries to each sj.te, and "Mi~istries" holds summary data on pl.ans for delivery and actual deliver.ies of inetal products by target functions, construc- tion ministries and main territorial administrations of the USSR Gossnata. The 05 input generator is used to monitor data and convert the f iles into a form convenient for 1_oading into the oata base. "Oka" DBMS facilities were u~ed to build programs to load the itiformation into the = data base, summarize it by target functions, ministries and +territorial administra- tions and printout the output form. The output document for each target function contains summary data or. the ~elivery of each type of resource for all projects ef the target function, including by each uiinistry, and within a ministry--hy each territorial administration and by projects. In the output document are given data - 6; FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040400060041-6 FOR OFFICIAL USE ONLY on the plan for deliveries for the year and for the reporting period, and also data on the actual delivery for the reporting period. Planned for the future is expan- sion of the nomenclature of monitored products, development of new output forms and tying this task to other tasks associateu with supply of materials and equipment for capital construction. Solving this problem achieves elimination of manual monitoring of the fulfillment of assignments, establsihed annually by the USSR Gossnab, to the territorial agencies for delivery of physical resources to the major projects of the national economy. Address for inquiries: Moscow, Orlikov per., 5, GVTs Gossnaba SSSR [USSR Gossnab Main Computer Center]. COPYRIGHT: Tsentral'nyy nauchno-issledovatel'skiy institut informatsii i tekhniko-ekonomicheskikh issledovaniy po material'no-tekhnicheskomu snabzheniyu _ Gossnaba SSSR, 1980. 8545 CSO: 1863/245 � , . 66 _ . , APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000400460041-6 FOR OFFICIAL USE ONLY UDC 681.32:[339.6:662.6/.7] PROCESSING OF ONLINE DATA ON FUEL RESERVES AT MA.TOR USSR ENTERPRISES AND POWER STATIONS Moscow MATERIAL'NO-TEKHNICHESKOYE SNABZHENIYE, SERIYA 4: PRIMENENIYE MATEMATICHES- KIKH METODOV, VYCHISLITEL'NOY TEKHNIKI I ORGTEKHNIKI V MATERIAL'NO-TEKHNICHESKOM SNABZHENII in Russian No 4, Apr 80 (manuscript received 29 Dec 79) pp 2-3 [Article by Ye. G. Torchinskiy, deputy departmen~ chief, M~3n Computer Center for USSR State Committee for Material and Technical Supply] [Text] The task on processing of online data on fuel reserves at major power sta- tions and enterprises in the country has been worked out and placed into industrial operation at the USSR Gossnab Main Computer Center. The task is designed to receive, process on a computer and output to receivers reports on the availability - of fuel reserves at more than 600 ma~or enterprises and power stations of 14 ministries. This problem is solved in two stages: in the f~.rst, special forms of normative provisions containing a list of the enterprises and types of fuel used by them are circulated. The forms of normative provisions are formatted and printed on the YeS-1040 computer at the USSR Gossnab Main Computer Center. Based on the filled-out forms r~ceived from the ministries in September of each year, correction and forming of the library of normative provisions and dictionaries of enterprises and types of fuel is per- formed on the "Minsk-32" computer. Processing of the online information and report formatting is performed in the second, main stage. Information is acquired from the enterprises and power stations by the information centers oF the territorial agencies of the USSR Gossnab. The information received is recorded on the specLal form for online reporting. The data on the form is punched on perforated tape and transmitted by teletype to the USSR Gossnab Main Computer Center. The information goes direct~y into the "Minsk-32" computer and is checked; when errors are detected, information on the errors is printed on the subscriber's teletype. After the information is received and checked, an information file is formatted and prc>cessed and the output documents are printed. 67 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY The "Dialog" software system, developed by the Information Di~patch Center of the USSR Ministry of the Coal Industry, the "Minsk-1560" communication channel interface apparatus and the "Minsk-32" computer are used to receive, check, output error information to the subscriber, record on magentic tape and form intermediate files. A COBOL software system for the "Minsk-32" computer, developed at the Main Computer Center, is used to form thk main information files and library of normative provisions and dictionaries, to process information and print the output documents . Data for the problem are also used by the Information Center of the USSR Gossnab Main Computer Center to submit information to the Central Staff of the USSR Gossnab by the "Wang-2200" minicomputer. The "Minsk-32" computer magnetic tape is recopied into the YeS computer codes and then recorded on the NID [magnetic disk] of the "Wang-2200" minicomputer. Consolidated data are output to a video ~ terminal and for printing. The result of solving the problem is a report on fuel reserves consisting of two parts: consolidated data on fuel reserves for the entire list of enterprises and power stations and types o f fuel by each ministry; a list of enterprises and power stations with a critical f uel reserve (three days for the Ministry of Power and _ Electrification, five days for the others). Implementation of the tas k has made it possible to intensify monitoring the development of fuel reserves and to improve expeditious management of the process of distributing fuel to th e major enterprises and power stations in the country. Address for inquiries: Mo scow, Promyshlennyy proyezd, 3, GVTs Gossnaba SSSR [USSR Gossnab Main Comput e r Center). COPYRIGHT: Tsentral'nyy nauchno-issledovatel'skiy institut informatsii i tekhniko-ekonomicheskikh i ssledovaniy po material'no-tekhnicheskomu snabzheniyu Gossnaba SSSR, 1980. 8545 CSO: 1863/245 68 ? APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAI. USE ONLY PUBLICATIONS , ABSTRACTS FROM JOURNAL 'AUTOMATION AND COMPUTER TECHNOLOGY', JULY-AUGUST 1981 Figa AVTOMATIKA I VYCHISLITEL'NAYA TEKIiNIKA in Russian No 4, Jul-Aug 81 pp 97, 99 UDC 681.3.06:519.872 SOME RESULTS OF ANALYSIS OF A PROBABILISTIC MODEL OF TELEPROCESSING SYSTEMS [Abstract of article by S. F. Yashkov] (Text] A complete analysia of teleprocessing syatems described by the system M] G] 1 is given with a discipline of a uniform partition of the processor in which each request present in the system is serviced simultaneously with the others at a variable rate equal to 1/k in the case of k requests at the given time. A new analytic method of studying a system is deveioped; it is used to derive stationary distributions and features of basic characteristics. A number of practically im- portant characteristics is presented in a form convenient for calculations. New regularities of system behavior are established that substantially expand the capabilities of applying the dervied reaults for analysis of computer network models. Bibl, of 9 titles, 1 table. UDC 681.324 STRUCTURE OF A SYSTEM FOR DATA EXCHANGE BETWEEN PROCESSES IN COMPUTER NETWORKS - [Abstract of article by J. Nemeth] - (Text] Computer network services are effected by processes that exchange data with each other. In a hierarchical system, data exchange between processes has been studied less than at the other levels. This effort is devoted to a study of = this question. The model is based on the principle of the "meeting place" both to effect switching and to fully establish communications. "Meeting place" location � does not depend on terminal location. Communication control system does not depend on the system effecting data exchange. Practical use is illustrated by examples. Bibl. of 12 titles, 8 figs. 69 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2447/02/09: CIA-RDP82-44850R444444464441-6 FOR OFFICIAL USE ONLY UDC 681.324:519.872 MODEL OF A DISTRIBUTED TELEPROCESSING SYSTEM [Abstract of article by A. A. Atnosov] [Textj Analytic model of the transport eecommunicationsrnet~orkgandsthe system of cussed that considers the topology of th information transmission matrices. The communications network is represented as a system of interacting transmission paths (virtual connections) modeled by queueing networks of a special structure. ~A 8~a~eSeOfn hetpathslusing thegprinciple ofprob- lem of finding the final vectors semilocal balance. Bibl. of 6 titlea. UDC 681.32:519.713 CALCULATION OF LOWER-BOUND ESTIMATE OF NUMBER OF TERMS IN DNF SYSTEM DESCRIBING LOGIC STRUCTURE OF AUTOMATON [Abstract of article by E. E. Lange] [Text] A simple procedure is proposed for calculating the lo~er bound estimate of the number of different terms in a system ~StDucturesof atfiniterautomaton.~fThe - Boolean functions that describes the logic calculated estimate is the best in the sense.of achievability. The calculation pro- cedure consists in constructing some finite undirected graph and determining its chromatic number. Reducing graph dimensions in constructing it is considered to simplify determination of the chromatic number. The procedure can be used to esti- mate the possihilities of implementation of an automaton to be synthesized ~ith a pragrammable logic array with specfied dimensions and to estimate the quality of solutions to the problem of optimization of the logic structure of an automaton that are derived by using approximation methods of logic synthesis. Bibl. of 7 tiltes, 3 tahles. UDC 681.323:518.5.001 IMPLEMENTATION OF PARALLEL POLYNOMIAL APPROXIMATION IN MULTIPROCESSOR STRUCTURE [Abstract of article by 0. N. P'yavchenko, Ya. Ye. Romm and I. F. Surzhenko] [Text] A method is proposed to implement in a multiprocessor structure a parallel polynomial approximation in a time proportional to the logarithm of the degree of the polynomial, uzith a proportionality factor equal to the time of one multiplica- tion. Applications to implementation of functions, division and linear differential equations are given. Variants of apparatus implementation are given. Bibl.,of 14 titles, 2 figs., 1 table. 70 . , APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY UDC 681.3-181.48 - MICROPROCESSOR IMPLEMENTATION OF DIGITAL CONTROL AUTOMATA (Abstract of article by B. M. Kagan, G. Kh. Novik, V. G. Persheyev and M. I. Shamrov] [`rext] Discussed are features of implementation oriented to logic control of digi- tal control automata with microprocessor systems (M-systems). Different approaches to construction of M-system programs are given. Methods of program implementation of logic converters are analyzed. Bibl. of 5 titles, 5 figs. UDC 519.873:681.324 APPLICATION OF RELIABILITY THEORY TO PROBLEM OF EXPERT EVALUATION OF COMPUTER SYSTEMS [Abstract of article by V. I. Levin] [Text] Discussed is the problem of aggregating individual orderings Yi of one and and the same set of objects by different experts i= 1, n. In contrast to the classic statement, the author considers the link between Yi and the "true" ordering of objects Y(by introducing the probability q of error by an expert in evaluating the relationship ? i~ between two ob~ects). The author suggests aggregating the evaluations Q i~ according to the majority rule (when G~ = 0.5 according to the majority rule). It is shown that for Q.~ q all these rules when n-~ O'~ are asymptotically devoid of errors, but only the majority rule with the probahility of 1 precludes contradiction or uncertain situations. Aggregating the evaluations Yi into the evaluation Y is done by summing the collective evaluations of the paired relationships for all pairs of objects. It is shown that for that rule of aggregation, the probability of a collective error of the first kind (transposition of objects in Y) and of the second kind (formation in Y o~ loops) approaches zero when n-ajoo . Bibl, of 4 titles, 1 table. UDC 681.326.74.06:621.3.049.774.2 METHOD OF CONSTRUCTING CHECKING EXPERIMENTS FOR LARGE-SCALE INTEGRATED CIRCUITS OF MAIN STORAGE [Abstract of article by V. G. Totsenko and B. V. Strelyayev] [Text] A method is suggesr_ed for constructing checking experiment (PT) for MOS LSI memory based on a general approach called "checking the hierarchy of properties." The checking experiment is specified hy two sections of input-output sequences (VVP), one of which is designed to check the electronic framec~ork (EO), and the other to check the memory matrix and read/write amplifiers. When huilding the first section of the input-output sequences, assumptions are made on the possible nature of manifestation of any malfunctions of the electronic framework. The second sec- tion of the input-output sequences is not built in advance: in huilding the first section, consideration is given to the necessity of including in it sections of the 71 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY checking experiment that test the storage elements of the matrix. Completeness of the proposed test is shown. Length of the suggested test for the memory equals 6N + 2N + 1) where N is memory size. Bibl. of 9 titles, 1 fig. UDC 681.324 ENUMERATION OF ELEMENTARY SECTIQNS IN A COMPUTER COMMUNICATION NETWORK [Abstract of article by Yu. K. Apraksin] ~Text] Discussed is the problem of finding the set of all elementary sections in a computer communication network. Used as a model of the computer communieation network structure is a graph of communications; its structure is specified by a matrix of incidences. The suggested algorithm for.finding the elementary sections between two apexes of the graph is based on using a graph contraction operation and is a simple procedure for transforming the matrix of incidences into the matrix of elementary sections. Bibl. of 7 titles, 3 figs. UDC 621.376 SIGNAL PR~7CESSING IN DISCRIMINATOR SYSTEMS CONSIDERING NOISE [Abstract of article by G. E. Vasarin'sh and E. Kh. Khermanis] ~Text] Discussed are probability aspects of signal processing in discriminator systems with regard to intrinsic noise. Assumed as known are the laws of distribu- tion of additive noise of each discriminator. Probable characteristics of applica- tion of discriminator systems are dervied. Examples are cited. Bibl. of 4 titles, 5 figs. UDC 621.391.244:517.587 SIGNAL ANALYSIS IN TERMS OF REAL EXPONENTS (Abstract of article by A. A. Kochetkov, V. V. Krylov and D. M. Ponomarev] [TextJ Signal analysis by exponents with real indexes is discussed. Described is a method of determining the signal relaxation spectrum based on constructing an in- - verse filter in aggregation with a Fourier transformation. The method permits relatively simple construction of a digital analyzer of the relaxation spectrum. An algorithm is given to calculate spectral.dgnsity~of a signal specified by its discrete samples using a fast Fourier transform. Bibl. of 4 titles, 3 figs. UDC 681.324 RATE CAPABILITIES OI~' SYNCHRONOUS INFORMATION TRANSMISSION BY 'CHANNEL-TO-CHANNEL' ADAPTERS [Abstract oE article by V. A. Red'ko and F. A. Sklyarevich] [Text] Discussed is synchronous process of indormation transmission using a "channel-to-channel" adapter having a sectionalized buffer ~ith limited storage. 72 ~ APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000400460041-6 FOR OFFICIAL USE ONLY The mathematical model of the system in question is an SMO [queueing system] in which request service time and time between sequential entries of requests have constant components. A technique is offered for approximate analysis of such queueing systems to derive estimates of rate capabilities of synchronous informa- tion transmission. Example of application of technique is given. Bib1. of 5 titles, 4 figs. UDC 681.324 SYNTHESIS OF MULTIMACHINE STRUCTURES FOR AUTOMATION OF SCIENTIFIC RESEARCH (Abstract of article by G. P. Vasil'yev, G. A. Yegorov and V. I. Shyaudkulis] [Text] Based on an example of a specific multimachine hierarchical complex, authors discuss solution to problems of synthesis of logic and physical structures of such systems. They substantiate composition and structure of system software and organization of interaction of its various elements. They discuss interfaces between user programs and system softt~are. They point out the possibility of using the suggested logic structure in complexes of various configurations. Bibl. of 8 titles, 7 figs., 1 table COPYRIGHT: Izdatel'stvo "Zinatne", "Avtomatika i vychislitel'naya tekhnika", 1981 8545 - CSO: 1863/248 73 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000400460041-6 FOR OFFICIAL USE ONLY I COLLECTION OF PROBLEMS ON ~'IACHINE PROCESSING OF ECONOMIC INFORMATION A~ioscow SBORNIK ZADACH PO MASHINNOY OBRABOTKE EKONOMICHESKOY INFORMATSII in Russian (signed to press 23 Feb 81) pp 2, 109, 132-133, 135 [Annotation, table of contents, tables 18 and 36 from book "Collection of.Problems on riachine Processing of Economic Information", by Gerta Kirovna Begotskaya, Vasiliy Petrovich Kosarev, Boris Mikhaylovich Rudzitskiy and Em~na Alekseyevna Umnova, Izdatel'stvo "Finansy i statistika", 10,000 copies, 136 pages] [Excerpts] This book contains the problems associated with the design of systems for machine processing of economic information--SMOEI: design of codes, source documents and machine media. The main emphasis is on development by design of com- puter installations (estimate design work). The varieties of information proces- sing systems are treated and problems on evaluating the efficiency of different computer data processing systems are elaborated. This book is intended for students in WZ's specializing in economics. Table of Contents Page Introduction 3 ' Chapter 1. Economic Information 5 1.1. Classification of Economic Information 5 1.2. Economic Problems and Properties of Economic Information 9 1.3. Requirements for Economic Information 12 Chapter 2. Structure and Evaluation of Economic Information 14 2.1. Identif ication of Units of Information 14 2.2. Structure of Economic Information 17 2.3. Determining Volumes of Economic Data 23 2.4. Qualitative Characteristics of Economic Information 25 Chapter 3. Information Procedures and Mathematical Apparatus for 28 Description of Economic Information 3.1. Information Procedures 28 3.2. Algorithmization of Economic Problems 32 3.3. riathematical Apparatus for Description of Economic Information 39 Chapter 4. Organization of Design of Machine Processing Systems for 45 Economic Information 4.1. Design of Nomenclature Codes 45 74 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 4.2. Design of Source Documents and Machine Media 59 4.3. Design of Media of Result Information 79 G.4. Design of Technological Processes for Machine Processing of Economic Information 88 Chapter 5. Organizational Hardware for Machine Processing Systems for Economic Information 109 5.1. Characteristics of Comp;~ter Installations 109 5.2. Building Computer Networks and Centers 111 5.3. Organization of Computer Installations 115 5.4. Efficiency of Systems for Machine Processing of Economic Information 124 Table 18. Table of Computer Center Categories by Staff Wages Nuiaber of Points Quantitative and Qualitative Characteristics 2 4 6 8 10 12 Volume of work performed, 1. to 200 X thousands of rubles 2. 200-500 X 3. 501-800 X 4. 801-1200 X 5. over 1200 X Problem composition by 1. unlimited X performance schedule 2. current X 3. on-line X Functions performed 1. problem prograir,ming and algorithmization X 2. automation of programming and i:.~~ouuction of economic-+mathematical methode X 3. organization of ASOD (automated data proceasing syatem], methodological directio� and coordination of work of related computer centers X Classification of to group I over 18 points computer center to to group II from 11 to 18 points wage groupa to group III to 10 points - 75 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FUR OFFICIAL USE ONLY Table 36. Standards for ~~timating Labor Input and Cost of Operations Average Opera- Hourly Hourly tor's Norm of Output Hourly Amortiza- Norm Rate, tion, Type of Unit of (N) ~ubles Rubles, Type of Operation Equipment Measurement (Ch) (Cha)_ 1 2 3 4 S 6 1. Document reception document 700 0.4 line 2. Pricing Iskra-111 operation 650 0.34 0.02 3. Pricing supervision " " 750 0.34 0.02 4. Check pricing by supervisory tabular form "Correction cf Errors" " " 6,000 0.4 0.02 5. Automatic pricing EVP80-2 card 5,000 0.4 C.215 6. Automatic card selecting or mergin$ RPM80-2M " 15,000 0.34 0.294 ' 7. Card punching P80-6 " 220 0,4 0.016 8. Alphameric card punching PA80-6 " 200 0.4 0.055 9. Check card punching by verification K80-2 " 290 0.4 . 0.016 - 10. Check alphameric card punching by verification KA80-2 " 250 0.4 0.032 11. Addition on tape SDV-107 operation 1,200 0.34 0.004 12. Counting method of verifying punched cards-- comparison of totals, row-column correction of errors SDV-107 (word) 800 0.4 0.004 13. Counting method of card punching verification-- ~ combination of operations 11 and 12 SDV-107 " 800 0.4 0.004 14. Reproduction and � duplication of cards PR80-2 card 5,000 0.4 0.2 15. Interpretation of cards RM80-1 " 4,500 0.4 0.3 16. Perforation by graphic marks PS80-1 " 5,000 0.4 0.137 17. Repunching of erroneous carda P80-6 " 100 0.4 0.016 18. Balancing tabulation T-SM " 4~200 Q.4 0,15 "for printing" 76 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00854R000440060041-6 FOR OFFICIAL USE ONLY 19. Balancing tabulation with pricing T-5M " 4,000 0.4 0.5 20. Balancing tabulation "for total" T-5ri " 6,000 0.4 0.15 21. Checking of balancing tabular forms and entry of check sums batch of in journal Iskra-111 documents 50 0.4 0.02 22. Card sorting S80-SM card column 20,000 0.34 0.19 23. Card sorting SE80-3 " 28,000 0.34 0.055 r 24. Tabulation of suramaries ` "for printing" T-SM card 3,100 0.4 0.15 25. Same, with total punching PI80-1 " 2,8Q0 0.4 0.207 26. Tabulation of sununaries "for total" T-5M " 3,500 0.4 0.15 27. Same, with total punching PI80-1 " 3,200 0.4 0.207 28. Tabulation of alpha- meric summaries TA80-1 " 4~200 0.4 0.35 - 29. Same, with total PI80-1, - punching TA80-1, " 3,900 0.4 0.407 30. Output of tabular form tabular summaries Iskra-111 fo~ 1 0.4 0.007 31, Tape punching STA-2ri word 2,80~ 0.4 0.025 3'L. Automatic collation of , two tapes KSU-1 " 15,000 0,4 0.048 33. Automatic copying f~om tapes to cards BLP-1 card 4,800 0.4 1.125 34~ Card input to computer YeS computer " 30,000 35. Card output from computer YeS computer " 6,000 Note, In operations 12, 21 and 30, the modeJ.s of the machines are used in a limited way. But in estimating their amortization, it is necessary to account for full labor input oF the operations since at this time they are not used for other operations. For a11 machines~ except the reproducer, collating and interpreting machines, elec- tronic computer perforator, reading perforator and devices for compying from perfor- ated tapes to cards, the hourly amortization norm is based on two-shif~ operation... COPYRIGcIT: Izdatel'stvo "Finansy i statistika", 1981 8545 CSO: 1863/257 77 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040400060041-6 FOR OFFICIAL USE ONLY - UDC 681.32:621,32 rIULTIFUNCTIONAL AUTOMATA AND THE ELEMENT BASE OF DIGITAL COMPUTERS - rioscow MNOGOFUNKTSIONAL'NYYE AVTOMATY I ELENiENTNAYA BAZA TSIFROVYI~H EVM in Russian 1981 (signed to press 18 Dec 80) pp 2-4, 239-240 ~ [Annotation, foreword and table of contents from book "~fultifunctional Automata and the Element Base of Digital Computers", by Valentin Aleksandrovich Mishchenko~ Valeriy Dmitriyevich Kozyuminskiy and Aleksandr Nikolayevich Semashko, Izdatel'stvo "Radio i svyaz 6,000 copies, 240 pages] _ [Text] Presented systematically in this monograph are the theoretical and practi- cal aspects of the synthesis of general-purpose and multifunctional digical auto- mata--~a new class of digital devices that permit implementation of speciric func- tions based on the principle of multifunctionality. The element base, implemented - on the basis of these automata, permits building high-speed computers and h,omo- geneous computing media. Described in ~the book are methods for sWnthesis of tilese automata, and numerous examples of jmplementation of concrete structures are given. The monograph is intended for scientific workers, engineers, post-graduac~ students and those in upper-level. courses of WZ's studying the questions of computer technology and electronics, Foreword The great and critical problems facing domestic computer engineering are forcing specialists in this field to look for n~w rnethods of synthesis of computer struc- turc~. It is now clear that t~ie future belo~gs te multiprocessor compute~ systems for parallel processing of information, in wh~ch the questions of regular.-?ty c~f the nedium play an essential role, affecting the durability, reliability, adaptability to manufacture and a tlumber of othe~r pzactically important indicators of c~mputer. ~ structuzes. ~ _ The abundan~ possibilities of con~temporar~ and future microelectronics arQ setting an economically itnportant task for theoreticiana and practitioners: development of a set of LSI circuits t'riat will meet the needs of a broad range of users and take , into account the future develog:r:ent of computer ~triictures. In extensive use now as these LSI circuit;s are micropr.~cessors in combination with programmable logic arrays. 78 ' ~ ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040400060041-6 _ FOR OFFICIAL USE ONLY Ici this book, the authors offer the reader another class of automata that can be used both in combination with programmable logic arrays and microprocessors and as stand-alone units, These are general-pu~pose and multifunctional automata wich pract:ically unltmired functional capabilities L�hat pernit building super fast com- puters and homogeneous computer media. Preliminary studies made by the authors o� this book have ~ndicated their practical implementability [57-60]. The very groblems of general-purpose and multifunctio~al structures cannot be = cal.led absolucely new. In the domestic and foreign li.terature, there are a number of .largely theoretical. works on this topic. These are the werks of Soviet experts: E. V, Yevreinov, I. V. Prangishvili, E. A. Yakubaytis et al., as we11 as of foreign authors [53-56]. Nevertheless, the authors believe that ~i1is relatively new direction is in need of certain popularization and des~rves greater attention from developers. ' In this book, the authors present their own original canceptiun of gener.al-purpose a~d niultifunctional structures and cite fundamentally ne~o rrovisians and results ot synthesis or general-purpose and multifunctional st:tor~atao In contrast to - knuwn works, the authors have set themselves the tas?c of bxinging theoretical results to practical schemes and techniques for their synthes~s t~-~at can be used in c;esign work. At the same time, the book has a cleari S~~:p?-essed monographic nature since it generalizes a number of known works by the aui:hors listed above and pr~ ~ents systematically the main achievements in this field. The features o~ general-purpose and multifunctio:-ia? a~stou~ata that afford a stri.ct mattiematical description of the process of the~r s}Tr.~~esis permit solving a number. of known problems that now bother computer desi~ners und developers. Here are some of them: - l. Ttie large nomenclature of LSI circuits required for building raodern computers. General-purnose and mu].tifunctional automata permifi reducing this nomenclatu~e to the minim�� for many ideologies for building architectures for fiture digital com- put~rs. 5~lving this problem has a clearly expressed economic n~ture since deve- - lopr~ent of modern LSI circuits is rather expensive. 2, Increasing the output of suitable circuits in series production. This problem ;.s solved by the broad functional capabilities of the element base, which permits improving the te~hnology of production that does not vary over an extended period and therPby increasing the output of suitable circuits. 3, Sharply reducing the volume of reserve components and raising the repairability~ ciurability and reliability of computers. This problem is solved by developing in~erchangeable modules o� minimal nomenclature with controllable functionality witli invariable structure and links, which ultimately permits putting reserve elcmctits in the machine itself. 4, i�i~ichine synthesis of elements and devices for computers. This problem may be ~ solv~~ci on the basis of the theory of general-purpose and multifunctional automata t},ro~i~;h the properties of uni.versality of these automata, since a general-purpose :nodule m~iy "generate" a large number of digital microstructures, T~ ~ rtader will find in this book a mo~e detailed pres~ntation of the individual - ~~~rts of these problems. - 'I 9 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 r FOR OFFICIAL USE ONLY ~ This book was written '.~y u tezm of authors. The foreword and sections 1.1, 2.4, - 2.5, 5.7 and 10.3 were wri+;,en by A. tlishchenko, V. D. Kozyuminskiy and A. N. ~ Semashko; chapters 2, 3 and 7-1~ were written by V. A. Mishchenko and V. D. - Kozyuminskiy; ar.3 chapters 1 and 4-6 were written by V. A. Mishchenko and A. N. Semashko . - The authors and publisher will be very grateful for comment$ and remarks on the book's contr~nt; these should be senc to: Moskva, Glavpochtamt, a/ya 693 ~ [Moscow, M;~in Post Office, Box 693], "Sovetskoye radio". ~ Table of Co~:ltents Page Foreword 3 Multifunctionality and Some Problems of Cor.,uter Technology 1.1. Problems of Development of Comp+iters and Their Element Base 5 1,2, Main Trends in Element Base Development 9 1.3. Analysis of State of Theory of Multifunctional Automata ~ 12 1.4. Discussion of Results 13 Ceneral-Purpose and riultifunctional Digital Automata 2,1. General-Purpose and Multifunctional Automata. Concepts and Definitions 14 2.2. Mathematical Models of U- and M-Automata 16 2,3. Structural Synthesis of U-Automata 20 2.4. iiain Characteristics of U- and M-Automata 26 2.5. Classification of Pi-Automat~ 31 2,6. Disr.ussion of Results 34 Two Methods of Synthesis of M-Automata _ 3.1. Structural Synthesis of M-Automata by Canonical Nethod 36 3.2. Synthesis of M-Automata Based on Source Model 45 3.3. Comparacive Evaluation of M-Aufiomata Synthesis Methods 51 3,4. Discussion of Results 54 Synttiesis of Trivial Un - Automata ~+,1, Analysis of Functional Capabilities of Logic Elements S5 4,'L. Statement of the Problem of Synthesis of Trivial Un-Automata 61 4,3. Synthesis of Trivial Ui-Automaton 63 4.4. Synthesis of Trivial UZ-Automata in the Cla~s of Tuning Alphabet A 64 4,5. Synthesis of Trivial U2-Automata in the Class of Tuning Alphabet B 65 4.6. Synthesis of Trivial Uz- Automata in the Class of Tuning Alphabet C 70 4.7. Synthesis of Trivial Un- Automata by Cascade Methods 75 4,8. Synthesis of Trivial Un--Automata by Isolation Method 80 = 4.9. Transformation of Trivial Automata in Various Classes of Tuning Alphabet 1 86 4,10, Synthesis of Trivial Un-Automata with Arbitrary Inputs of Tuning 89 4.11. Discussion of Results 92 Synthesis of Trivial M-Automata 5.1. Statement of Problem of Synthesis of Trivial M-Automata 93 80 . , APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 5.2. Questions of Analysis of Functional Capabilities of Trivial U1-Automata for Realization of Logic Functions g n Variables 95 n 5.3, i~ethod of Synthesis af Trivial Mg-Automata on Basis of Un-Automata 96 5.4. I~iethod of Constructing Networks. Switching N~-Automata 109 5.5. Synthesis of Trivial Mn-Automata by Diagram Methods 113 5.6. Synthesis of Trivial r:n-Automata by Fitting Method 120 5.7. Synthesis of Trivial Mn-Automata by Method of Optimization of Starting Model 127 5.8. Synthesis of Trivial 1~-Automa~a 130 5.9. Discussion of Results 132 Elements of Theory of Selecting Functiona ii~ Solving the Problem of Synthesis of Trivial U- and r~i-Automata 6.1. Statement of Problem of Synthesis of Trivial U1- and M1-Automata on Basis of SPlecting Functions n n 133 6.2, Types of Selecting Functions. Their Interrelation and Properties 133 6.3. Presentation of Logic Functions by Sections of Selection 135 6.4. Logic Aspects of Selecting Functions of the Form sc when omega = 1 137 6.5. Logic Aspects of Selecting Functions of the Form sc when omega~ = 2~-1 141 6.6. Questions of Hardware Implementation of Selecting Functions of the Section of the Form sc 144 6.7. Synthesis of Trivial Un- and Mn-Automata on Basis of Selecting Functions of Section of Form sc 145 6.8. Discussion of Results 149 SyntheGis of M-Automata on Basis of Initial General-Purpose Digital Model 7.1. Definition of Parameters of Structural Schemes That Realize Transformation of A. 150 i 7.2. Selection of Structural Scheme for Initial Digital Model 152 7.3, Definition of Logic Functions and Adjustments of Initial Model 166 7.4. Optimization of Initial Model 170 7.5. Discussion of Results 178 Questions of Automation o� Synthesis of Discrete Devices on Basis of U-M~del 8.1. General Structure of System for Machine Synthesis of M-Automata 180 8,2, Algorithni for Selection of Initial Un-Automaton 182 - 8.3. Algorithm for Definition of Logic Functions Realized by Trivial U1-Automata of Idodel 185 n 8.4. Algorithm for Search of Adjustments of U-Model 188 8.5. Algorithm tor Optimization of Structure of U-Model I~1 8,6. Discussion of Results 198 Examples of Synthesis of M-Automata by Iviethod of Optimization of Initial Mode1 9.1, Synthesis of Single-Digit Adder-Subtracter with Adjustment of C1ass C 199 9.2. Synthesis of Digit of ALU [Arithmetic and Logical Unit] with Adjustment of Class C 203 81 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY 9.3. Syntheais of Digit of ALU with Ad~ustment of Class A 1 206 9.4. Synthesis of M-Automata with Memory on Basis of Trivial U-Automata with Feedbacks n 209 9.5. Synthesis of Digit of ALU of Accumulator Type 214 9.6. Questions of Synthesis of Tabular Devices for Information Conversion 218 9.7. Discussion of Results 222 Evaluation of Efficiency of M-Automata 10.1. Evaluation of Complexity of M-Automata 223 10.2, Reliability of M-Automata and Systems Based on Them 226 10.3. Evaluation of Cost of M-Automata 232 10.4. Discussion of Results 234 Conclusion 235 Bibliography 236 COPYRIGHT: Izdatel'stvo "Radio i svyaz"', 1981 8545 CSO: 1863/257~-A 82 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY UDC 681.32 PAR.ALLEL PROCESSORS FOR CONTROL SYSTEMS Moscow PARI~LLEL'NYYE PROTSESSORX~DLYA ~~~YAYUSHCHIKH SISTEM in Russian 1981 ~ (signed to press 3 Apr 81) pp 2-11, 142-158 (Annotation, table of contents, preface, introduction, chapter 9 and biblio graphy from b~ok "Parallel Processors for Control Systems", by Yakov I1'ich Fet, Energoizdat, 12,000 copies, 160 pages] [Text~ Author deals with questions associated with algorithmic and logic struc- tures of paralle]. processors. He covers various types of homogeneous proces sors that make ir ;'~ossible to substantially raise the speed of systems in solving large- scale computational and informational-logic problems. The processars descriUed can be used for hardware implementation of softwar e func- tions and for developing various functional-oriented systems. _ For engineers and technicians engaged in design of computers and systems, as well as students in VUZ's of the corresponding specialties. Table of Contents Page ~ Preface 3 Introduction. Control Systems and Problems of High-Volume Processing of Information 5 Chapter 1. Specialized Homogeneous Processors 12 ~ 1. Basic Concepts 12 2. Some Examples of Special-Purpose Homogeneous Processors 21 Chapter 2. Special-Purpose Homogeneous Processors Oriented to High Volume Retrieval Operations 31 3. Def{nition of Search Operations 31 4. Processor Oriented to Maximum Search Operation 33 S. Processor Oriented to Retrieval Operation for All Greater and All Lower Numbers 36 6. Processor Oriented to Search Operation by Interval 40 Processor Oriented to Search Operation for Nearest Number 42 8, Basic Operations of Special-Purpose Homogeneous Processors 47 Chapter 3. Ordering of Information in Special-Purpose Homogeneous Processors 50 9. Some Ordering Algorithms 50 83 FOR OFFICIAL USE ONLY ~ APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040400060041-6 FOR OFFICIAL USE ONLY 10. Ordering Based on the Basic ~peration PNB [Maximum Search] (PNM) [Minimum Search] 55 11. Ordering Based on the Basic Operation PVB [Search for All Greater] (PVM) [Search for All Lower] 56 12. Ordering in Specified Limits 60 13. Component Comparison of Two Arrays 63 14. Classification of Numbers 65 Chapter 4. Link between Special-Purpose Homogeneous Processors and Other Types of Para11e1 Processors 6~ 15. Principle of Associative Processing 67 16. Other Variants of Parallel Processing 69 Chapter 5. Associative Parallel Processors 72 17. Associative Processor with Expanded Functions -Processor) 72 18. Some Information on APL Language 76 19. Microprogramming for Special-Purpose Homogeneous Processors 81 20. High-Volume Computations in the 7't -Processor 85 21. Compor.ent Multiplication of Two Vectors 88 22. Accelerated Component Multiplication of Two Vectors 92 Chapter 6. Orthogonal Parallel Processors 96 23. W. Shooman's Design 96 24. OME1V [Orthogonal Mini EmbedmeNt] System 98 25. Orthogonal Processors Based on Special-Purpose Homogeneous Processors 100 Chapter 7. Dynamic Parallel Processors 101 26. Striicture of Dynamic Parallel Processor 101 27. Dynamic Processors Based on Special-Purpose Homogeneous Processors 112 Chapter 8. Data Structure Conversion in Parallel Processors 120 -3 28. Data Structure Conversion Operations 120 29. Some Devices for Data Structure Conversion 126 30. Use of Special-Purpose Homogeneous Processors for Data Structure Conversion 132 Chapter 9. Possible Applications of Parallel Processors in Control Systems 142 31. Autonomous Special-Purpose Processors 142 32. Special-Purpose Attachments 144 33. Functional Modules 146 34. Possibilities of Practical Implementation of Special-Purpose Homogeneous Processors 149 Bibliography 153 Preface In modern control systems, the amount of information and requirements for fast pro- cessing of it are continually increasing. One important reserve for increa~ing the efficiency of control systems is the use of special-purpose digital devices atid, in particular, parallel processors. 84 , APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000440060041-6 The achievements of modern microelectronics are turning paralle? computations and parallel processors into reality. A number of efficient associative and conveyor type parallel processors have emerged. Use of them shows that the effective speed in solving suitable problems is increased several orders. The heiglltened interest in special-purpose and parallel processors has also been reflectecl in the increasing number of articles published in Soviet and foreign scientific and technical journals. In this book, special attention is paid to one type of parallel computing device-- the so-called special-purpose homogeneous prucessors, covered in the first four chapters. Each special-purpose homogeneous processor is a two-di.nensional combina- tional iterative array, matched with matrix storage. The processor is oriented to execution of some group (large-block) operation on an array stored in its storage elements. These basic aperations can have independent vaiue (for example, in infor- mation rctrieval problems) or serve as the basis for microprogram implementation of various algorithms. Also discussed are other types of parallel processors, in some sense "related'' special-purpose homogeneous processors, namely: associative (chapter 5), orthogonal (chapter 6) and dynamic (chapter 7). In connection with organization of parallel processing of large arrays of in'~orma- tion in modern control systems, assuming great importance are questions of rearrange- _ mEnt of arrays and conversion of their structure. The possibility of using some special-purpose homogeneous processors for data structure conversion is showr. in chapter 8. The description of the specialized processors is illustrated by examples of solving individual large-scale problems that are typical for control systems. Questions of application of parallel processors in control systems are also discussed in chapter 9. ~ The author is sincerely grateful to L. V. Kantorovich and D. A. Pospelov for the great interest and attention shocan to this work. The author thanks V, N. Aleyeva, N. A. Zosimova and Ye. V. Suvorov and the editor, - Ye. I.. Piyl', for their valuable comments that facilitated improving the content of the book. The Author Introduction. Control Systems a~id Problems of. High-Volume Information Processing The modern scientific and technical revolution is inconceivahle without the wide- spre~id applicatiun of various automated control systems. The necessity of continual development of control systems, and increasing their functional capabilities, pre- cision and reliability of operation makes enhanced demands on digital computing de- vices--the major component of these systems. The requirement for enhanced efficiency of computers is closely tied to the features of the problems solved in control systems [1]. 85 FOR UFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00854R000440060041-6 FOR OFFICIAL USE ONLY Typical for automated sys~..ms ~f organizational control is the extremely large volume of information (data bases used in large control systems are continually growing and may reach 1010 to 1011 bytes). Mod ern computers (third-generation machines), as a rule, r~,iznot adequately handle the processing of such large files. In automated and automatic process control systems, the volume of information to be processed ;s relatively not large, but usually processing of control actions is required in the real-time mode. With the large rate of flow of processes in the controlled object, this exceeds the capability of existing machines. Finally, special-purpose automated systems can make very high demands on the volume and speed of processing. An example of a special problem with such demands is air traffic control at a modern airport frec atl that the STARAN special-purpose processor, the most powerful of digital machizies now in operation, was developed just to solve such problems [2J). Thus, there is now a discrepancy between the requirements of the problems to be solved in control systems and the capabilities offered by computers in series production. Practical ways of raising computer efficiency are ~ell known: these are raising the speed of the elements, improving software and developing the logic structure and architecture of the machines. As for element speed, almost all possibilities have been exhausted (discounting optica:L elements that may emerge in future generation systems). Experts believe tt~at we can expect only a one-two order of magnitude further increase in computer efficiency by increasing the speed of the electronic elements. SoFtware improvement is aimed basically at speeding up preparation of the prohlems (automation of programming, application of prob lem-oriented languages), but not the solution of them. Actually, on the one hand, the modern operating system facili- tates fuller use of computing system resources, and as a consequence of this, an increase in its efficiency. On the other hand, complex software systems require for their "own" piirposes considerable consumption of storage and time, which reduces the overall posiLive effect. Also, development of these systems entails - large inputs of human labor which leads to high cost of them (especially compared to the rapidly falling cost of hardware). The most important reserve for raising the efficiency of computers is, apparently, the evolution of their architecture. This stems from the majority of problems to be solved in control systems having a special nature; the initial information of these problems is organized into special structures (arrays, tables, etc.). In solvir.g these problems on general-purpose machines having classical architecture that are oriented to a standard set of basic instructions and operate sequentially with scalar values, it is not possible to take advantage of the specific nature of the data and types of processing. Proposals on developing special-purpose softwar e and hard~are t~ solve complex prob- lems with a special nature have heen made relat ively long ago [3, 4]. 86 FOR OFFICIAL U SE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040400060041-6 FOR OFFICIAL USE ONL. the question of expanding the internal machine language by raising the hardware level of interpretation and matching computer instructions with high-level language atakements was cunsidered in detail in [5]. Applying the terminology introduced in [6], one may say that machines with classi- cal architecture have an "elementary education"--within the limits of arithmetic in the early grades. Machines will operate considerably more efficiently if their "education" level is raised--if they are taught hardware execution of the more com- plex and special functions from among those now executed by software (for example, calculation by formulas, implementation of cycles, retrieval of information, editing of data, execution of some functions of operating systems). Improving machine architecture and including special-purpose devices oriented to execution of various complex functions (in particular, parallel processes) in the makeup of the machines entails substantial hardware outlays. Until recently, this was an obstacle to practical use of this important reserve for increasing efficiency. However, the rapid development of integrated circuit technology is creating practical prerequisites for the design and manufacture of machines with a new architecture. It is well know:l that the degree of LSI douhles every one to two years while cir- cuit cost goes down annually by about 30 percent. This makes it possible to produce very complex hardware with very small dimensions at 1o~r cost. Thus, in 1978, Zilog Inc. began marketing a lb-bit single-chip microprocessor, the Z8000, with about 17,500 transistors on a chip the size of 6 x 6.5 mm [7]. Other firms produce similar microprocessors also. By the mid eighties, forecasters expect development of a 32-bit computer with a degree of integration of 100,000 tran:-~:tors per chip [8]. Even greater densities are being achieved in storage devices. A numher of firms are already producing random-access memory of 64K bits on a single chip, and h; the mid eighties, 1M-bit units are expected [9]. Mic:roprocessor LSI circuits will apparently become the main element hase for compu- ters in the next few years. It is necessary to note that in an ideological, structural sense, all microproces- sor sets (microcomputers) now being produced remain r~achines with classical archi- tecture--serial operation. The majority of them are relatively si.mple devices, approaching in the best case in characteristics lhird-generation minicomputers. In the process, the very conservative structural solutions used in the machines made with tt~e new elements do not permit taking advantage of the great possibilities contained in tt~e new Cechnology. Evidently, favorable conditions have arisen for a serious reexamination of computer architecture. The possibilities of raising the "education" level of the machines by raising the hardware leve~., specialization of individual devices and application of hi.gh-volume, group and para11e1 processing have become concrete. Campared to the other ways of improving computers, development of architecture is esr~ecia!ly important since it affords new p~ssibilities for construction of algo- ritl~ms and stimulates development of hetter mathematical approaches to solving ~7 P'OR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY problems. The emergence uf the first parallel processors has led to a considerable number of efforts in which computational methods and algorithms are being revised to fit the capabilities of the new machines [10]. The importance and promise of such efforts are noted in [10]. Experts believe that coordinated, interrelated development of computational methods and structures of computers can lead to a three-four order of magnitude increase in efficiency of solving many problems, in- cluding the major problems to be solved in control systems. What place in the development of computer architecture do the parallel processors discussed in this book hold? To answer this question, we have to make some sort of classification of digital systems. M. Flynn [llJ singles out four types of systems - in his classification that has received wide dissemination: 1. With one stream of instructions and one s~~eam of data (OKOD); classical single- ~ processor machines (von Neumann architecture). 2. With one stream of instructions and many streams of data (OKrID). The same se- quence of instructions is executed simulataneously on many se~s of arguments. These systems are best adapted for execution of component-by-component vector oper- ations. 'l~his ~ype includes the majority of the so-called "matrix" and "vector" processors (array processors). 3. With many streams of instructicns and one stream of data (MKOD). The single stream of data is moved along a"conveyor," at each position of which is executed _ one instruction (operation) from the specified sequence. The stream of results is removed from the output of the "conveyor." MKOD systems have been termed mainline or conveyor systems. 4. With many streams of instructions and many streams of data (MKrID). In this sys- tem, several programs are exP.cuted simulataneously with independent data (although exchange of information between inidvidual elements of the system is also po~sible). The MKIvID type includes multiprogram multiprocessor systems, computer networks and some array processors. Of tl~ese four types of systems, the OKOD are sequential types of machines; the - others employ various forms of parallelism. In accordance with this classification, the parallel processors given further con- sideration here are of the OKMD type. Machines of this type, ho~ever, can tae built differently. Thus, for example, a homogeneous computer system (network) con- sisting of a number of general-purpose computers executing the same program may also fall within the OKrID type. We are interested in a particular type of paralle] processor--proces..ors with a homogeneous internal structure (microstructure). Each of these processors is based on some special-purpose homogeneous processor--a two- dimensional homogeneous structure oriented to implementation of a particular special function anri an information processing algorithm. From the vie~rpoint of the ?nicrostructure, a special-purpose homogeneous processor is a variety of a homo- geneous computer environment [12]. As a rule, it is a two-dimensional combination iterative array [13] matched with two-dimensional storage. In a iunctional sense, a special-purpose homogeneous processor is a problem- oriented digital device (automaton). After the source information (arguments) is input into the processor storage eZements and the Uoundary constants are fed to its external inputs, r~sults are generated after some time (depending on the duration of the transient processes) at specific external outputs. 88 . , APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R400440060041-6 Let us ca11 the operation that the special-purpose homogeneous processor executes in similar fashion (in one cycle) a basic operation, and the set of control signals that implement this basic operation a micro instruction. It will subsequently be shown that the same special-purpose processor may execute several basic operati.ons - (r~icro instructions). From these micro instructions, micro programs are compiled to solve various problems more complex th4n those embedded in the processor struc- ture as its basic operations. Since eact~ basic operation is an operation on an entire array of data, the micro programs of special-purpose homogeneous processors in their speed are considerably more efficient than standard subroutines by which data is similarly processed in OKOD type machines. Thus, a special-purpose homogeneous processor consists of an operating unit (the tc~o-dimensional homogeneous structure proper) and a microprogram control unit. When operated within a control system, it can be considered a special-purpose periptieral. In the process of executing some complex program, the central machine in the system (the general-purpose computer) sends to the special-purpose homo- geneous processor specific group statements and corresponding arrays-arguments. After receiving the task, the special-purpose processor executes it autonomou:-ty and outputs the results. Given in ~14] is a detailed classification of parallel processors, according to which devices similar to special-purpose homogeneous processors (in particular, the so-called Kautz arrays) are classed as associative processors with a high degree of parallelism. It is shown in chapter 4 that there is another link between special-purpose homc- geneous and associative processors (associative storage units), and with that it is suggested that an associative processoti be considered a special case of a special- purpose liomogeneous t,~pe. In addition to associative, cae shall also discuss other variants of parallel pro- cessors--vertical, orthogonal and dynamic, and the links between them and the corresponding special-purpose homogeneous processors will be shown. Chapter 9. Possible Applications of Parallel Processors in Control Systems 31. Autonomous Special-Purpose Processors From the preceding chapters, it is evident that special-purpose homogeneous proces- sors and the other. parallel processors can efficiently solve the variou, problems thrxt are typical for control systems. Spe~:ial.-purp~se proc~ssors ar_e now used as autonomous special-purpose computing and control units, as speci.al-purpose attachments operating in a complex with a standard (general-purpose) computer and as functional modules within multiprocessor data processin~ systems. L~t u~ di~cuss in more detail some features of these applications of special- r~~rnose processors. t~ j FOR OFFICIAL USE ONT,Y APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY As autonomous units, special-purpose processes are widely applied in on-board and ground special-purpose syst~ms, in control machines for some processes and in sys- tems that control operations. Usually of deciG've importance in this case are the sma11 dimensions and weight of the processor as well as the speed of decision-making. In these cases, special-purpose homogeneous processors must be used in the mode of executing their own basic operations. An example is the application of the processor u~ith the basic operation PNB [maximum searchJ for recognition of critical situations in controlled objects. Let us assume that the values of some parameter are continuously coming into the storage elements of an max processor from a number of controlled objects. Sign~als at the right boundary of the proces~~r will always indicate that ohject at - which the regulated parameter reaches maximum value. Instead of an ~ max process~~r, a 1(~ min-processor, adjusted to execute the maximum search operation, can also be used for this purpose. Further, if the maximum per- r..itted value of the regulated parameter is sent to the register of the boundary attributes of the '1~' min-Processor, then us~.ng the basic operation PVB [search for al.l greater] (or PVM [search for all lower]), the objects whose parameters are ouC- side tlie tolerance can be dynamically identified. R A similar. proble~n--dynamic identification of an ohject with the maximal parameter is solved in the unit for allocation of priority in a multichaiinel data processing system. In the implementation of walking robots [71], the coordinates of the robot's legs must continuously be compared to the coordinates of obstacles in the process of con- structing the path of its movement to avoid collision with the obstacles. This problem can be solved using a y' -procESSOr the following way. For each leg, let ehe diEferences between its current coordinate and the corresponding coordinates of all obstacles be calculated continuously (let us note that this calculation is a component-by-component operation between a scalar and a vector). The differences obtained go into the rows of the 1t -processor, and the minimal permitted difference is recorded in its regi.ster of boundary attributes. In doing so, it is evident that a PVM [searct~ for all lowerJ operation will dynamically select the ro~ corres- ponding to a dangerous situaeion--impermissible approaching of a leg to an obstacle, and identify this obstacle. High-volume, parallel processing makes its possible to control the path of robot movement with a large number of legs and otastacles in real time. Another example of an autonomous special-purpose homogeneous processor; P-processor that solves the problem of sorting (grading) items of mass production. IE tl~e boundary values of the permitted zones are recorded in the rows of a -processor and the values of the measured parameter of the items are sent to t~he buses Q,, the signals generated at the right houndary can be used to conti�ol the receiving mechanism of an automatic soxter. ~ 90 FOR OFFICIAL USE OI~TLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY 32. Special-Purpose Attachments These attachments substantially increase efficiency if the special problems to which they are oriented make up a considerable portion of the load on the computet complex. Many examples of successful use of special-purpose attachments both for solving complex computational problems and for information-logic processing of large arrays are well known. The IBM 2938 special-purpose processor is used in a complex with the 360 (370) machines. This processor is oriented to executing a number of high-volume opera- tions: component-by-component addition and multiplication of vectors, scalar pro- duct of vectors, reductive addition and multiplication and others. Based on these operations, efficient micro~programs are built to solve differential equations, pr.oblems of linear algebra, correlation and spectral analysis, etc. The category of special-purpose attachments of a computing orientation should in- clude the vertical processor of the OMEN [Orthogonal Mini EmbedmeNt] system (attach- ment for the PDP-11 machine) and the array processor (quadrant) of the ILLIAC-IV system (attachment for the B-6500 machine). Several examples of the use of special-purpose processors oriented to information logic problems were given in chapter 7. The application of a general-purpose computer in a complex ~zith a special-purpose sorting processor for ordering of large files was discussed in detail in [72, 73]. The special-purpose sorting processor consists of the following units: control unit, storage unit for attributes (associative type), storage unit for addresses and the micro program control unit. The processor is considered an active peripheral con- nected to the selector channel of a general-purpose computer. Access to it is organized as to an external device with a suitable number. The controlling program in the general-purpose computer prepares in main storage the file to be processed as well :is the following information nePded for operation of the special-purpose proces- sor; A1 beginning address of file M in main storage; al beginning address of attribute of first record in file M; 1-record length; r- attribute length; n- number of records in file N!; f- special-purpose processor operation code. When the special-purpose processor is accessed, this information is recorded in the control unit. Then independent operation of the processor begins and it consists of ' two stages: loading and processing proper, In th~ loading stage, the control unit sequentially for each rI file record calculates its beginning address Ai and the beginning address of its attribute ai. The attribute is called from computer main stora~e and entered into the next row of attribute storage, and the address A, into the corresponding row o� address storage. After attribute storage (and addre:;s storage) is f illed, the pr^^Qssing stage begins. At each step of this stage, the speciEic row corresponding to the given operation code f is selected in attrioute storage (for examp:l.e, the row with the minimal attribute if ascending order is executed). The beginning address of the record with the selected attribute is sent irom address storage to the next cell of computer main storage. Thus, in the ~ 9l FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2047/02/09: CIA-RDP82-00850R000404060041-6 ~ FOR OFFICIAL USE OI~TL,Y process oF tlie prc~cc:~sinb, there is formed in main sto~'age a:a~le of beginning addresses of f{.le M recorc~~ ordered by the values of their attributes, Based on this table, the controlling program in the general-purpose. computer copies the records of file M in the needed order, As indicated in [74], by c~ing the special-~~urpose attaclunent, the etficienc; of the system in solvinb various i.nformation retriev~l problec~s is increased by one to two orders of magnitude. Simiiar assesa^~ents of etficiency were obtained as a result of modeling of a data base ~~rocess~r ~57], The si.mplest associative memory wr:~s used as attribute storage in the attactunents described in (72, 73]. If a special-purpose homogeneous processor witt-~ a higher hardware level and greater functional caPabi:._~ies (the y' -processor, for example) is used, then obviously greater efficiency can be obtaitied and the range of problems solved expanded. 33. Functional Modules The decr~ase in cast of LSI circuits and ttie incrEase in level of integration makes it possible to raise the question of more extensive use of special-purpose nroces- sors, paticulariy of including theri as functional-oriented modules with various data processing syste:r,s. It is believed that in this case each functional module performs some procedure assigned to it, thereby implementii~g some subroutine by hardware. If an anticipa- tory review of the overall program is made and data directed for processing to the appropriate functional modules~ then the central prncessor of the system is released from executing a number of subroutines and, what is especially important, from inefficient sequential execution of cycles in implementing high-volum~ op~ra- tions. In this system, hardware supports the software. A natural question arise: how many different functional modules is it necessary to have in a system? ~ It i.s well known that a special-purpose processor executes most efficiently that operation to which it is oriente.d by hardware. There is a contradiction between the attempt to perform each high-volume operation in a functional module designed especially for it and the necessity of a reasonable lim~ttation on the nomenclature of modules. This contradiction can probably be resolved if the following circum- stances are considered. First, each special-purpose homogeneous processor, as a rule, implements not one, uut several basic op~rations; second, in this case, the specific content of a particular operation does not play a role, but only the nature of the computing process necessary for implementation of it. By type of computing process (and consequently, also by requirements for hardware of a functional module), several dif.ferent operations may coincide. Thus, in a certain sense, all compo~ient-by- compor~ent arithmetic operations on vectors are similar. Several elementary func- tiens (for example, sin x, tg x and others) can be implemented in one device, that~ for example, is oriented to computation of polynomials with different coefficients. 92 FOR OFFICIAL U5E ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FC )FFICINL USE ONLY To estimate the number of i:unctional modules that afford satisfactory system eF.ficiency, high-volume statements or procedures can be classified in some way. To this end, the composite and mixed functions c~i the APL language was taken as a list of high-volume statements ~46, 75], This selection is justified because~ for ex~unple, all high-volume statements in the PL/1 language (builr-in functions, ex- - pressions with arrays) are included in the list of APL stat~ments. The same may be said ~f muny procedures, used prac~~.cally in prog:ams, but not having corresponding statements in high-level languages, The analysis ~f high-volume statements in APL in ~75] showed that by nature of pro- cessing, they can be grouped into tha followi:?g five cl~,sses. 1, Component-by Component Computatioaal. ~ This class includes mainly high-volume statements in which the arguments are one or two (or more in rare cases) numeric vectors, the result is also a numeric vector and the processing consists of executing a more or less complex transformation of a computational nature. This transformation is applied independently to each com- ponent (or set of similar components) of vectors-arguments and may iri principle be performed aimultaneously on a11 sets. Best suited for execution of statements o� the first class are special-purpose pro- cessors of the associative fiype (for exatnple, the ?I' -processor) . 2, Reductive Computational. This class includes high--volume statements in which the argument is a numeric vec- tor, tiie result is a scalar and the processing consists of sequential application of some computational operatiou to the components of the vector-argument. Examples of t~lese operations are reductive addition of. an array and reductive multiplication. High efficiency in ex ecur.ing reductive computational operations is achieved in conveyor (mainline) parallel processors ~76], 3. Component-by-Component Logical. Tilis class includes m ainly high-volume statements in which the arguments are two numeric vectors, the result is a logical vector and the processing consists in e~:ecuting sonie logic operation (Eor example, comparison) on some pair of similar � components of vectors-arguments. Processing may be performed simultaneously on ~ a.ll pairs of componE~nts. The ~-processor is best suited of th~ processors described above for execution of cumponent-by-componFnt logic oper.ations. 4. Reductive Logical, This class includes high-volume sCatements in which the argument is an array of bi- nary numbers (or b.it strings); the result is a subset o� the elements of the array- a~-gument (in the particular case, a number or row; associated with the initial array by a specific rel.ationship corresponding to the given statement. Processing cunsists in executing on an argument some global operation that can be called a "scan" of the array, 93 F~i~ OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL US~ ONLY Examples are tne maximuni search and the nearest neighbor search. Operations of this class are performed most efficiently in the appropriate : special-purpose homogeneous processors~ The y' ~processor can be used as the device that performs a sufficiently broad set of reductive logic operations. 5. Data Structure Transformations. The arguments and results are arrays of binary numbers (or biL strings) that have in the general case c~i�ferent dimensions and ranks. Pru~essing consists in chang- ing the structure of the array-argument or sequence of its components. Operatians of this class are efficiently peri,:..med by special-purpose devices for transfonning data structures, In particular~ a nu~?ber of important operations of _ this class can be i~rplemented by using the -processor. Of c.ourse, the above classification is conventional. On the one hand, the high- volume statements could be combined into larger classes, making use of the fact that, Eor esample, in an associative processor in principle, not only component--by- component, but also reductive operations are executable (although with less efficiency). On the other tiand~ perhaps it is advisable to separate out some of the state~aents included in one class (especially in the broad first class) and find for them a more efficient, narrowly specialized hardware implementation. Neverttieless, it may be assumed that the number of functional modules sufficient For a considerable increase in efficiency of a system for high--volume data processing is not large: four to six. That set of special-purpose processors in the aggregate can cover a broad assort- ment of high-volume statements, while achieving sufficient universality of hardware support. Required for efficient system operation, of course, is a good organiza- tion of distribution of tasks, loading of arrays-arguments and exchange of intermediate results. 34. Possibilities of Practical Implementation of Special-Purpose Homogeneous Processors There are now concrete possibilities for the transition to practical manufacture and use of special-purpose homogeneous processors. The main prerequisite for this is the possibility of censtructing large, inexpensive - circuits. This circumstance, associated with progress in the technology of manufacturing integrated microcircuits, has led~ as is well known, to the rapid development and distribution of microprocessors (77]. Series produced microprocessors make it pos- sible to implement microcomputers, the architecture of which differs little from the traditional. Such machines, just as other general-purpose computers, gossess - limited "capabilities." 94 ~ APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-04850R000400060041-6 On the other hand, as the degree ~f integrarion increases, the difficulties grow in designing and testing the irregular circuits of. traditional architectux~e~ as well as in connecting the individual modules with a limited number of leads. - Or�e way to surmount these difficulties is to develop special-purpose, funtional- oriented microprocessor. LSI circuits which especially possess a regular structure. Such devices include the now mastered programmable logic arrays (PLM), special- purpose microprocessors for multiplication, fot executi.ng fast Fourier transforms, etc. Some experts believe the main application of very large integrated micro- circuits will be special-purpose functional devices [9]. - Two-dimen~ional homogeneous logic circuits are a convenient object for implementa- tion with LSI circuits. Therefore it can be assumed that based on the structures of special-purpose hom~geneous processors such as described in this book, it is expedient to develop some special--purpose LSI circuits that supplement the standard series of microprocessors and are compatible with them. An important factor determining the selection of a specific structure for practical implE-~aentation of an LSI circuit is the flexibility of the circuit and the possi- - bility of sufficiently broad application of a type module. From this viewpoint, special-purpose homogeneous processors have definite advantages, With a rational design, a special-purpose homogeneous processor may execute not only that operation for whi~h its logic network is properly oriEnted, but also a number of other operations. I.et us note that the min Processor executes with = identical efficiency tlie basic operations PVM, PNB, PVB, PBM and PS ~search for a11 lo~~er, maximum search, search for a11 greater, search for nearest neighbor lower and co~nci.dence search]; the ~,-processor perf urms the basic operations of compression, expansion and others, There are various ways to expand the functional capabilities of special-purpose homogeneous processors: - combining in one structure the functions of several different processors; using a sequence of elementa'cy (basic) operations of the processor to execute com- plex procedures. This is how, f.or example, the sorting algorithm in the ~f` -proc~ssor or the associative computations in the lt' -processor are implemented; using some storage elEments and (or) external leads for ad~ustment of the structure. Ttie adjustment of t't, -matrices for execution of the different functions to trans�or-cn data strt?ctures can serve as an example. 1'o illust=rcite the flexibil~.ty of special-purpose homogeneous processors, let us list some possibilities of a circuit with the simplest ce11s'-the ~j,-matrix. By construction, tt~e a max Processor is designed to p~rform the basic operation o� maximum search. It is well known that by microprogramming in the A. -processor, various information retrieval algorithms can be implemented. By ad~ustment of the ce].ls, rhe switching function can be ~anpl.emented, - 9ti FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00854R000400060041-6 F'~d OFFICIAL USE ONLY The method of modeling of: associative memflry in an Q. -structure is suggested in [78]. In principle, this makes it possible to organize various high-volume compu- tations by the method of sequential transformation of states. The possibility of implementing random finite automata in an -structure is shown in [23]. A cell of associative memory built on the base of an Q-structsre has four steady states, and thanks to this the Q. -structure can efficiently be used as functional memory. P. Gardner, in [79], discusses tabular methods of computation of functions in homogeneous structures, and the gresence of redundant sfiates in associative meinory cells is used for minimization. As shown in ~80], the ~-structure makes it possible to apply a11 the minimization methods suggested by Gardner and in addi- tion, achieve a number of additional advant^,;es. An essential factor influencing the realization of some structure with LSI cir- cuits is the iimitation on the nuwber of external connections. This requires serious study of r_he design of special-purpose homogeneous processors with regard to specific teatures of the technology applied. In doing so, it is necessary to make use of. all known means to reduce the lead/element ratio (incorporation within the module of circuits for control, encoding~ decoding~ multiplex mode of exchange, etc,), With some types of special-purpose homogeneous processors~ the problem is simplified because a relatively sma11 set of standard vectors is sufficient for control. These vectors can be generated by special circuits included within the design modules and controlled from without by feeding economical codes ("numbers") of the vectors needed, - Solving the problem of leads is substantially facilitated in going from static homogeneous structures to dynamic (chapter 7). This tnakes it possible to consider - that dynamic~ parallel processors are now closest to practical use. BIBLIOGRAPHY 1. Kantorovic'~, L. V., "Paths of Development of Computer Facilities for Solving Large-Sca1e Problems of Optimal P~anning and Control," in the book "Optimizatsiya (sbornik trudov)" [Optimization (Collection of Works)], - Novosibirsk, Institute of Mathematics~ Siberian Branch of the USSR Academy of Sciences, 1972~ issue 6(23), pp 5-7. - 2. "Mu1t:Cprocessor Systens and F~ra11e1 Computing," edited by F. G. Enslow~ trans- late3 from English, Moscow, rlir, 1976, 384 pages. 3. Kantorovich, L. V., "A Mathematica~. Notation Suitable When Calculations Are rIade on a Machine," AOKLADY AN SSSR, 1957, Vol 113~ No pp 738-741, 4. Iiantorovich, L. V., "Prospects of Efforts to Automate Programming Based on a Large-B1ock System~" TRUDY MAT. IN-TA IMENI V, A. STEKLOVA, 1968~ Vol 96, pp 5-].5. 96 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPR~VED F~R RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY ~ 5, Glushkov, V. M.; Barat~anov, A, A.; Kalin~.chenko, L. A. et a1., "Vychislitel'nyye machiny s razvitymi sistemami interpretatsii" (Computers ~ with Developed Systems of Interpretation], Kiev, Naukova dumka, 1970, 260 pages. 6. Zadykhaylo, I. B.; Kamynin, S. S. and Lyubimskiy, E. z., "Questions of Design of Computers from Blocks of Enhanc~d Qualification," in the book "Sistemnoye ~ i teorPticheskoye programmirovariiye (sbornik trudov)" [Systems and Theoretical Programming (Collection of Works)], Novosibirsk, Computer Center, Siberian Branch, USSR Academy of Sciences, 1972, pp 126-135. 7. Sima, Ti. "16-Bit Instrument That Implements Nicroprocessor and Minicomputer Functions,~' ELEKTRONIKA, Russiau translation~ Vol S1, No 25, 1978, pp 24-34. 8. Turuta, Ye. N., "Multimicroprocessor Systems," ZARUBEZHNAYA RADIOELEKTRONIKA, No 3, 1979> pP 3-27. 9. Keypis, R, P,, "Problems of Aesign and Production of Very Large-Scale Integrat- ed Circuits," ELEKTRONIKA, Russian translation, Vol 51, No 24, 1978, pp 29-51. 10. Fadcieyeva, V. N. and Fa~deyev, D. K., "Parallel Computations in Linear A~.gebra," KIBERNETZKA, No 6, 1977, pp 28-40. 11. Flynn, ri. G., "High-Speed Computing Systems," TRANS. IEEE, Russian translation, Vol 54, No 12, 1966, pp 311-320. ~ 12. Yevreinov, E. V,, "Microstructure of Elementary Machines of Computer System," in the book "Vychislitel'nyye sistemy (sbornik trudov)" (Computer Systems (Co1- - lection of Works)], Novosibirsk, Institute of ?4athematics, Siberian Branch, USSR Academy of Sciences, issue 1962, pp 5-28. 13. Hennie, F. C., "Iterative Arrays of Logical Circuits," New Yoxk-London, MIT Press-Wiley, 1961, 241 pages. " n 14. Tliurber, K. J. and Wa1d, L.D., Associative and Parallel Processors,~' ACM COM- PUTING SUKVEYS, Vol 7, No 4, 1975, pp 215-255. 15. Hennie, F. C., "~inite-State rtodels for Logical Machines," New York-London- Sydney, Wi.ley, 1968, 456 pages, 16. Varshavskiy, V. I,; Marakhovskiy, V, B. and Peschanskiy, V. A. et al., "Odnorodnyye struktur.y/ (Analiz. Sintez. Povedeniye)/'~ jHomogeneous Structures: Analysis, Synthesis, nehaviorJ, rioscow, Energiya, 1973, 152 pages. 17. i~~ajithia, J. C, and Kitai, R., "An Iterative Array for Multiplication of Signed Binary Numbers," IEEE TRANS., Vo7. C-20, No 2, 197.1, pp 214-216. 18, Toma, C. I., "Cellular Logic Array far High-Speed Signed Binary Multiplication," IEEE TRANS., Vol C-2G, PIo 9, 1975, pp 932-935. 19. Cui1d, H. H., "Some Cellular l.ogic Arrays for Non-restoring Binary Aivision," }ZADIO ELECTRON, ENG., Vol 39, No 6, 1970, pp 345-348. 20. Cappa, N. and 1?c~nacher, V, C., An Augmented Iterative Array for High-Speed Binary Division," IEEE TRAP+S., Vol C--22~ No 2, 1973, pp 172~175. - 21. Shautrian, A. PI,, "Piatrix Extraction of. Square Root," in the book "Vychi.slitel'- nayr; tekhnika" [Camputer Technology], Leningrad, Leningrad State University, issue 6, 1977, pp 105-111. 97 - FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-04850R000400060041-6 FOR OFFICIAL USE ONLY 22. Fet, Ya, I., Patent 42~+:1.41 (USSR), "Ustroystvo dlya sortirovki informatsii" ~Information Sorting DeviceJ, published in I., No 14, 1974. 23. Fet., Ya, I., "Massovaya obrabotka informatsii v spetsiali.zirovannykh odnorodnykh prntsessorakh" [High-~Volume Infoi-mation Processing in Special- Purpose Homogeneous Processcrs], Novosibirsk, Nauka, 1976, 200 pages. 24. Blokh, A. Sh, and Lades, V, I., "Synthesis of Single-Cycle Circuits, Behavior of Which Is Described by Linear Inequalities," IZV. AN BSSR. SER. FIZ-TEKHN. - NAUK; No 3, 1966, pp 93-99. 25. Berkovich, S, Ya, and Kochin, Yu. Ya,, "Nearest Neighbor Search," AVTOMATIKA I TELEMEKHANIKA~ No 2, 1975, pp 176-178. 26. Fet, Ya. I,, Patent 63~+372 (i]SSR), "~le~^snr assotsiativnoy matritsy pamyati" [Associative Memory Matrix Element]z pu~tished ~.n B. T., No 43, 1978. - 27. Kratko, M, I., "Regular and Stable Iterative Networks~" in the book "Problemy k~'.,ernetiki" [Problems of Cybernetics], Moscow, Nauka~ issue 19, 1967, pp 95-106. 28. Glushkov, V. M.; Gladun, V. P.; Lozinskiy, L~ S, et al., "Obrabotka informa- tsionnykii~~~assivov v abtomatizirovannykh sistemakh upravleniya" [Info~ation File Processing in Automated Control Systems], Kiev, Naukova dumka, 1973, 183 p. 29. Papernov, A. Aa and Podymov, V. Ya., "Metody uporyadocheniya informatsii v - tsifrovykti sistemakh" [Information So~ting Methods in Digital Systems], rioscow, Naulca, 1973, 384 pages, - 30. Flores, I., "Analysis of Internal Computer Sorting," JOURNAL OF THE ACM, Vol 8, No 1, 196], pp 41-80. 31, Martin, W. A., "Sorting," ACM COMPUTING SIRVEYS, Val 3, No 4, 1971, pp 147-174. 32. Krayzmer, L. Borodayev, D. A.; Gutenmakher, L. I. et al., "Assotsiativnyye _ zapominayushchiye ustroystva" [Associative MemoryJ, Leningrad, Energiya, 1967, - 184 pages, 33~ Kautz, W. H., "Cellular Logic-in-Memory Arrays," IEEE TRANS., Vol C-18, No 8, 1969, pp 719-727. 34. Bratal'skiy, Xe. A. and Krupskiy, A. A., "File Sorting Method Using Associative Device," VOPROSY RADIOELEKTRONIKI. SER. EVT, issue 7, 1973, pp 90-93. 35. Batcher, K. E., "Sorting Networks and Their Applications," in AFIPS CONFER. PROC., 1968 SJCC, Vol 32, pp 307-314. 36, Fuller, R. H. and Estrin, G., "Some Applications for Content-Addressable - N?emories," in AFIPS CONFER. PROC., 1963 FJCC, Vol 24, pp ~+95-508. 37, Shooman, W., "Para11e1 Compu~ing with Vertical Data," in AFIPS CONFER. PROC., 1960 EJCC, Vol 7.8, pp 111.-115. 38. Shooman, W., Patent 3277449 (USA), "Orthogonal Computer." 39. Shooman, W., "Orthogonal Processing," in '~Para11e1 Processors, Syste.~?s, Tech- nologies and Applications," Spartan Books, 1970, pp 297-308. 40, Crane, B. A, and Githens, J. A., ~'Eulk Processing in Distributed I,ogic Memory,~~ TEEE TrcANS., Vol EC-14, No 2, 1965~ pp 186-196. 98 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-04850R000400060041-6 FOR OFFICIAL USE ONLY _ 41. Lipovski, G. J., The Architecture of a Large Assvciative Processor,'~ in AFIPS CONFER, PROC., 1970 SJCC, Vol 36~ pp 385-396. 42. Popova, G. ri. and Prangishvili, I. V,, "Associative Parallel Processor. for for Batch Data Yrocessing," A~TOMATIKA T TELEMEKHANIKA, No 1, 1972, pp 171-184. ~ 43. Aleyeva, V. N. and Fet, Ya. I., "Associative Processor for Large-B1ock Proces- sing of Information," in the bootc "Optimizatsiya (sbornik trudov)" ~Optimiza- tion (Collection of Works)], Novosibirsk~ Institute of Mathematics, Siberian Branch, USSR Academy of Sciences, issue 15(32), 1974~ pp 154-177. 44. Prangishvi..li, I. V.; Popova~ G, M.; Smorodinova, 0. G.; and Chudin, A. A., "Odnorodnyye mikr.oelektronnyye assotsiativnyye protsessory" [Homogeneous Micro- electronic Associative Processors], Moscow, Sovetskoye Radio, 1973, 280 pages. 45. Fet, Xa. I., Patent 746728 (USSR), "Zapominayushchiy modul' dlya matrichnykh blokov pamyati" [Storage Module for Matrix Blocks of Me~nory], published in B. I., No 25, 1980. 46. Iverson, K. E., "A Programming I.anguage," New York-London, Wi1ey, 286 Pages. 47. Yakin, S., "APL/360 Reference Manual," Chicago, Science Research Associates, 1968, 160 pages, 48. Abrams, P. S., "An APL Machine," Stanford Univ., AD-706741, 1970~ 213 pages. - 49. Thurber, K. J. and riyrna, J, W,, "System Design of a Cellular APL Machine," IEEE TRANS., Vo1 C-19, No 19r0, pp 291--303. 50. Harrison, ri. J. and Harrison, W. H., "The Implementation of APL on Associative Processor," in "Para11e1 Processing (Proc. of the Sagamore Comp. Confer., 1974)," Berlin-Heidelberg-New Xork, Springer, 1975, pp 75-96. ~51. Y.igbie, L, C., "The OMEN Computers: Associative Array Processors," in COMPCOIV 7'L. 6-th Annuai IEEE Comp. So. Int~ Confer.~ San Francisco, 1972~ pp 287-290. 52. Slotnick, D. L., "Logic-per-Track Devices," in "Advances in Computers," New York, Academic Press, Vol 10, 1970, pp 291-296. 53. Parker, J. L., A Logic per Track RetrievaJ. Systen," in "Proc. of IFIP Congress 71," Amsterdam, North-Holland Pub1. Comp., 1971, Vo1 1, pp 711-716. 54. Coulouris, G, r.; Ewans, J, M, and Mitciiell, R. W., "Towards Content-Addressing " in Data Bases," COMPUT. JOURN., Vo1 15, No 2, 1~372, pp 95-98. 55. Finnila, C. A. and Love, H. H., "The Associative Linear Ar.ray Processor," IEEE TRANS., Vol C-26, No 2, 1977~ pp 112-125. 56, Anderson, G, A. and Kain, R, Y,, "A Content-Addresse~' Memory Designed for Data- Base Applications," in "Proc. 1976 Int. Confer. on Parallel Processing," Ne.w York, IEEE, 1976, pp 191-195, 57. Banerjee, J.; tisiao, D, K. and Kannan, K., DBC--A Database Computer for Very ~ Large Databases," IEEE TRANS., Vol C-28, No 6, 1979, pp 414-429. 58. Su, S. Y. W.; Nguyen, L. H.; Emam, A and Lipovski, G. J., "The Architectural Features and Impl.ementation techniques of the Multicell CASSM,~' IEEE TRANS., Vol C-28, No 6, 1979, pp 430-445. 99 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 FOR OFFICIAL USE ONLY 59. Thurber, K. J., "InCerconnection Networks---A Survey and Assessment," in AFIPS Nat, Comp, Confer,, Vol 43, 1974~ pp 909-919. 60. Kalyayev, A. V., "Odnorodnyye komanutatsionnyye registrovyye struktury" - ~Homogeneous 5witch Register Structures], Moscow, Sovetskoye radio, 1978, 336 p. 61. Feng, T.-Y., "Data Manipulating Functions in Para11e1 Processors ~nd Their Implementations," IEEE TRANS., Vol C-23, No 3, 1974, pp 309-318. 62, Stone, H. S., "Parallel Pro^_essing with the Parfect Shuffle," IEEE TRAN~., ' Vol C-2C, No 2, 1971, pp 153-161. 63. Benes, V. E., "Optimal Rearrangeable Multistage Connecting Networks," BSTJ, Vol 43, No 4, 1964, pp 1641-1656. , 64. Opferman, D. C. and Tsao-Wu, N. T., "Or a Class of Rearrangeable Switching Net- works," nSTJ, Vol 50, No 5, 1971, pp 1579-1618. 65. Kautz, W. H.; Levitt, K. N. and Waksman, A., "Cellular Interconnec:ton Arrays," IEEE TRANS., Vol C-17, No 5, 1968, pp 443-~+51. 66. Waksman, "A Permutation Ne;,work," JOURN. ACM, Vo~ 15, No 1, 1968, pp 159-163. 67. Pease, M. C., "An Adaptation of the Fast Fourier Transf_o~!� to Parallel Proces- sing," JOURN. ACM, Vol 15, No 4, 1968, pp 252-264. 68. Lang, T. and StonE, H. S,, "A Shuffle-Exchange Network with Si~?g:i?fied _ Control," IEEE TRANS,, Vol C-25, No 1, 1976~ pp 55-65. _ 69. Aleyeva, V. N. and Fet, Ya. I., Patent 590747 (USSR), "Dvumernaya r.,dnorodnaya struktura dlya analiza logicheskikh vektorov" [Two-Dimensional Homogeneou~ Structure for Analysis of Logic Vectors], published in B. I., No 4, 1978, p 188. 70. Aleyeva, V. N., "Solving Problems of Linear Algebra with Associative Pro~essnr," in the book "Optimiza~siya (sbornik trudov)" [O~timization (Collection of Works)], Novosibirsk, Institute of Mathematics, Siberian Branch, USSR Academy of Sceinces, issue 18(35), 1976, pp 146-159. 71. Okhotsimskiy, D. Ye. and Platonov, A. K., "Algorithms to Control Walking Apparatus Capable of Surmounting Obstacles," IZV. AN SSSR. TEKHNICHESKAYA KIBERNETIKA, No 5, 1973, pp 3-10. 72. Barsamian, H., "Firmware Sort-Processor with LSI Components," in AFIPS Confer. " Proc,, 1970 SJCC, Vol 36, pp 183-190. - 73. De Fiore, C. R., "Fast Sorting," DATAMATION, Vol 16, No 8, 1970, pp 47-51. 74. De Fiore, C. R. and Berra, P. B., "A Quant~tative Analysis of the Utilization of Associative Memories in Data Management," IEEE TRANS., Vol C~23, No 2, 1974, pF 121-133. 75. Fet. Xa. I., "Hardware Support of High-V~lume Computations," in the book - "Optirnizatsiya (sbornik trudov)" [Optimization (Collection of Works)J, Novosibirsk~ Institute of Mathematics~ Siberian Branch, USSR Academy of Sciences, issue 22(39)~ 1978, pp 115-126. 100 � FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6 ~ 76. Fet. Xa. I.~ "Iniplementation of Group Operations on Rotor Con~eyor Processor," in the book "Optimizatsiya (sbornik trudov)" [Optimization (Collection of Works)J, Novosibirsk, Institute of Mathematics~ Siberian Branch, IISSR Academy of Sciences, issue 6(23), 1972, pp 104-115. 77. Prangi.shvili, I. V., "Mikroprotsessory i mikro-EVM" [Microprocessors and - riicrocomputers], Moscow, Energiya, 1~'?9, 232 pages~ 78, Fet, Ya. I., "Funr.tional Possibilities ~f Simple Computing Media," AVTUi~ATIKA I VYCHISLTTEL'NA1'A TEKHNIKA~ Riga, No 3, 1974, pp 4$-54. 79. Gardner, P. L., "Functional ~temory and Its Microprogramming Implications," _ I~,~i TRANS., Vol C-20, No 7, 1971, pp 765-775. 80. Fet. Ya. I., "Special-Purpose HomogeneoL~s Structures as Programmabl.e Logic Arrays," AVTOMATIKA I VYCHISLITEL'NAXA TEKiR~IKA, Riga~ No 6, 1979, pp 67-72. COPYRIGHT: Energoizdat, 1981 8545 CSO: 1R63/251 END 101 FOR OFFICIAL USE ONLY APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000400060041-6