CDC Cyber Research Articles

Monte Carlo simulation of U(1) lattice gauge theory

Lattice gauge theory is a technique for studying quantum field theory free of divergences. All the Monte Carlo computer calculations up to now have been performed on scalar machines. A technique has been developed for effectively vectorizing this class of Monte Carlo problems. The key for vectorizing is in finding groups in finding groups of points on the space-time lattice which are independent of each other. This requires a particular ordering of points along diagonals. A technique for matrix multiply is used which enables one to get the whole of the result matrix in one pass. The CDC CYBER 205 is most suitable for this class of problems using random “index-lists” (arising from the ordering algorithm and the use of random numbers) due to the hardware implementation of “GATHER” and “SCATTER” operations performing at a streaming-rate. A preliminary implementation of this method has executed 5 times faster than on the CDC 7600 system.

692 QUANTITATIVE COMPUTERIZED ANALYSIS OF HOUSE-STAFF TRAINING

The larger number of Pediatric residents and the consequent relative decrease in patient population have made it more likely that a resident may not care for an adequate number of patients during his training. At the same time, increased departmental specialization may result in a biased selection of patients managed by a resident.We have developed a Fortran IV program for use with the CDC Cyber 171 computer that records the actual patient population for each resident during his general inpatient experience. The information is collected directly from the supervising residents and is verified with hospital admission and occupancy records. Each patient is listed by his primary diagno sis, chart number, age, sex, dates of admission and discharge and month of care. The patients are sorted by diagnosis providing a cumulative record of the number and variety of patients managed by each resident. The age distribution and daily patient load are calculated and charts are retrievable for review. On our service, a PL-1 typically admitted 30 new inpatients a month and had a daily patient load of 8.4.A resident's individual record is compared to the total patient population of the inpatient service as constructed in this program. Deviations in experience from the complete service and obvious deficiencies are identified. Resident elective time is structured to compensate for these variations. The system is adaptable to register OPD, ER and nursery patients for similar analysis. This record of experience can be combined with monthly subjective faculty assessments and results of the ABP In-Training Examination to accomplish a thorough evaluation of each resident.

Comparison of Solutions to a Three-Dimensional Black-Oil Reservoir Simulation Problem (includes associated paper 9741 )

Summary A comparison of solutions to a three-dimensionalblack-oil reservoir simulation problem is presented.The test of the problem and a brief description of theseven simulators used in the study are given. Introduction Seven companies participated in a reservoirsimulation project to compare the results obtained bydifferent black-oil simulators. The companies werechosen to give a good cross section of the solutionmethods used in the industry. The participants wereAmoco Production Co., Computer Modelling Groupof Calgary (CMG), Exxon Production Research Co., Intercomp Resource Development and EngineeringInc., Mobil Research and Development Corp., ShellDevelopment Co., and Scientific Software Corp.(SSC). The paper presents the text of the problem, acomparison of results in graphical form, and a briefdescription of each model. The descriptions weresupplied by the participants. A variety of computers was used. Amoco usedIBM 3033, IBM 370/168, and Amdahl V/6. CMGused Honeywell 6000 DPS, and Exxon used Amdahl470/V5 and IBM 370/168. Intercomp used Cray-1and Harris/7. Mobil and SSC used CDC Cyber 175, and Shell used Univac 1110/2C Level 36. Thenumber of time steps and the central processor timesvaried considerably. Those interested in the actualvalues should contact the individual companies. Except for Shell, all the participants usedsingle-point upstream mobility weighting. Shell used twopoints upstream. Constraints and data are given inthe text. Statement of the Problem Areal and cross-section views of the reservoir aregiven in Figs. 1 and 2. The grid system is given in Fig.1. Stratification and reservoir properties are given inFig. 2. The reservoir is initially undersaturated. A gasinjection well is located at Grid Point (1, 1), and aproducing well is located at Grid Point (10, 10).Pertinent data and constraints are given in Table 1.PVT properties and relative permeabilities are given in Tables 2 and 3. The participants were asked tomake the runs and report the results described below. Runs To Be Made Case 1 Let the bubble-point (saturation) pressure beconstant with a value equal to the original value. Case 2 Let the saturation pressure vary with gassaturation - i.e., this is a variable saturation-pressurecase. The PVT lines at pressures above the calculatedsaturation pressures are parallel to the original line. Results To Be Reported The following results are to be reported.1. Plots of:a. Oil rate vs. time.b. GOR vs. time.2. Report annually and at abandonment:a. The pressures of the cell where the injectorand producer are located.*b. Gas saturation at Grid Points (1, 1, 1), (1, 1,2), (1, 1, 3),(10, 1, 1), (10, 1, 2), (10, 1, 3), (10,10, 1), (10, 10, 2), and (10, 10, 3).3. Report at the end of 8 years:a. Tables of gas saturation.b. Tables of cell pressures. *c. Tables of saturation pressures for the variablesaturations-pressure case. * JPT P. 13^

A Practical Method for the Rational Design of Ship Structures

A practical, rationally based method is presented for the automated optimum design of ship structures. The method required the development of (a) a rapid, design-oriented finite-element program for the analysis of ship structures; (b) a comprehensive mathematical model for the evaluation of the capability of the structure; and (c) a cost-effective optimization algorithm for the solution of a large, highly constrained, nonlinear redesign problem. These developments have been incorporated into a program called SHIPOPT. The efficiency and robustness of the method is illustrated by using it to determine the optimum design of a complete cargo hold of a general-purpose cargo ship. The overall dimensions and the design loads are the same as those used in the design of the very successful SD14 series of ships. The redesign problem contains 94 variables, a nonlinear objective function, and over 500 constraints of which approximately half are non-linear. Program SHIPOPT required approximately eight minutes of central processing unit time on a CDC CYBER 171 to determine the optimum design.

A critical comparison of several programming language implementations

AbstractThe implementations of Algol 60, FORTRAN, Pascal and Algol 68 provided on the CDC Cyber 73 of the Stichting Academisch Rekencentrum Amsterdam are compared on both qualitative and quantitative aspects. This paper concentrates on those aspects which are likely to be of interest to language designers and implementers, and to users of other computer systems.

The 25th and 26th Mersenne primes

The 25th and 26th Mersenne primes are 2 21701 − 1 {2^{21701}} - 1 and 2 23209 − 1 {2^{23209}} - 1 , respectively. Their primality was determined with an implementation of the Lucas-Lehmer test on a CDC Cyber 174 computer. The 25th and 26th even perfect numbers are ( 2 21701 − 1 ) 2 21700 ({2^{21701}} - 1)\;{2^{21700}} and ( 2 23209 − 1 ) 2 23208 ({2^{23209}} - 1)\;{2^{23208}} , respectively.

ROUT

ROUT is a computer program written in FORTRAN which simulates some of the characteristics of a mazerunning router for a two-layer printed circuit board. The objective of this program is not to actually route from-tos, but rather to aid in studying the effects of various wiring rules. For this reason, from-to points and wire segments are generated and placed using pseudo-random techniques. No actual connections are made. Rather, congestion is developed on the board as if the pseudo-randomly generated from-tos had actually been routed. ROUT has been developed and implemented on the University of Texas, Austin's CDC CYBER computer system for use in a graduate course in Digital Design Automation.

Var: A program to calculate the semivariance function in three orthogonal planes

The geostatistical concept of semivariance is used in VAR to calculate and plot variograms in three orthogonal planes. Distances between all points of pairs are calculated and the vector-angles between them are resolved into components in the X - Y, Y - Z, and Z - X planes respectively. Semivariance values for each distance calculation are assigned to the appropriate angular sector in each plane and variograms for all sectors are plotted in all three orthogonal planes on a CDC CYBER 73 computer.

A mathematical model for the computational determination of parameter values of anthropomorphic segments

A model is presented for determining by computation the parameter values of anthropomorphic segments from a battery of 242 anthropometric measurements taken directly from the subject. The model consists of 17 segments, includes the shoulders as separate entities, and offers the following advantages over previous models: it subdivides segments into small mass elements of different geometrical structures, thus allowing the shape and density fluctuations of a segment to be modelled in detail; in general, no assumptions are made on segmental symmetry, and principal axes transformations are performed whenever necessary; the model differentiates between male and female subjects (exomorphic differences, different density functions and mass distributions), adjusts the densities of certain segmental parts according to the value of a special subcutaneous-fat indicator, and fully accounts for the specificities of pregnancy and obesity; the input data errors are drastically reduced by performing direct anthropometric measurements rather than indirect measurements from photo images; the overall accuracy of the model is better than 3% with a maximum error of about 5%. The features listed above are confirmed by comparing experimentally determined parameter values (volumes, masses, coordinates of mass centroids, principal moments of inertia, coordinates of segmental origins) with model predictions for four different subjects. The execution time of the computer program executing the model is 0.515 seconds on a CDC Cyber 174 digital computer. A sample print-out is listed in the Appendix.

Application of a Minicomputer for Direct Numerical Control of Multiple Machine Tools

Numerical control is the directing of a process through the execution of a pre-coded instruction sequence. Direct numerical control refers to the real-time control of machine tools by the direct coupling of a computer to one or more machine tool controllers. In this paper, a specific direct numerical control system is presented and the corresponding implementation issues are discussed. This system uses an HP/2100S mini comouter with a 16K main memory to control up to 16 machine tools. Blocks of information are sent upon request to the machine tools, via a common bus, and the status of the entire system is maintained in tables in order to enable the control program to keep track of each part program execution. The presentation of the system's hardware and software features is followed by a discussion of the host support system which consists of programs written for a CDC Cyber 74 computer to facilitate the development and maintenance of part program libraries. Several possible extensions to the system, including the incorporation of fault tolerance features, are pointed out in conclusion.

Is APL really processing arrays?

This paper describes a study of the performance of APL. The study was carried out by placing code in an existing APL interpreter (APLUM running on a CDC CYBER 175 computer) to monitor the distribution of data types, element counts, and ranks of arrays. The effect of reference counts on data blocks was also monitored. Although the results of the study may not be surprising, they are very interesting. For example, it turns out that nearly half of all allocations of data are made for scalars or one-element vectors, and that although the over-all average number of elements in APL arrays is 30, most users achieve average element counts of around 10. From these figures, it can be shown that set-up time dominates the execution of APL expressions.

Experiences with PASCAL in an introductory course

Beginning September, 1975, the Computer Science Department at the University of Western Ontario switched from FORTRAN to PASCAL as the primary language in the first year introductory course. The motivation for such a change was to introduce into the first year course concepts and issues which were becoming central to Computer Science. In particular, the instructors involved in the course wished to a) emphasize algorithm design, effect of choice of representation on design and the creation of well-structured programs, b) expose students to the use of a variety of control structures and data structures, c) introduce, at least informally, the notions of complexity and correctness. It was felt that not only would such a course provide a good introduction to Computer Science for all students taking the course, but that those continuing with Computer Science as a chosen field would be well-prepared for subsequent courses. Since PASCAL could satisfy these requirements and a good implementation (on a CDC Cyber 6600) of the language was already available, it was the logical choice.

Maximizing coefficient alpha reliability while maintaining validity

Maximizing coefficient alpha reliability while maintaining validity

Computer-aided estimates of drainage charges - a Dutch case study.

When regional water management is regulated centrally on a cooperative basis, methods must be developed whereby drainage charges can be estimated from agreed and objective data. Reported here is a method to compute and map, from data on altitude, the supposed open water level after a given discharge and soil characteristics for available moisture, the charges to be levied. Independent 'maps' are compiled for each of the three attributes by computer mapping techniques. The area is split into a large number of small cells (here 12.5 m X 15.5 m), for each of which the altitude is estimated by interpolation from known spot heights. For the same cells the supposed open water levels are approximated by interpolation from a number of calculated water levels in streams and ditches. Known watershed divides are introduced as barriers to interpolation. Four soil moisture classes are generated. The three 'maps' provide for each cell the basic data required by an allocation algorithm which evaluates the value for altitude, water level and soil moisture class. For each cell the appropriate drainage charge class was determined. A payment class map is generated by the allocation and can be printed on the computer lineprinter, each class being given a distinguishing printing symbol. The method is demonstrated for a 210-ha test area under the jurisdiction of the Regional Water Board Salland. Further extension to include a cadastral map to allow automated invoicing procedures are discussed. The system is implemented on a CDC Cyber 72 computer and uses SYMAP (Harvard University, USA) as one module. (Abstract retrieved from CAB Abstracts by CABI’s permission)

Multistage multicomponent separation calculations using thermodynamic properties evaluated by the srk/pr equation of state

AbstractA computational procedure using thermodynamic properties evaluated by the Soave‐Redlich‐Kwong' (SRK) equation of state or the Peng‐Robinson (PR) equation of state is developed for multistage multicomponent separation processes. The procedure makes use of the Ishii‐Otto convergence scheme with the proposed modifications for application with the equation of state. A new technique is developed to improve an initial estimation of temperature profile in a column. The computational procedure is tested on CDC CYBER 172 computer for a wide range of problems including lean oil absorbers, reboiled absorbers and fractionators and a debutaniser operating in a natural gas processing plant in Alberta. The computational results are compared with those obtained using the Chao‐Seader model for generating the required thermodynamic properties. It is concluded that the present computational procedure can be used more reliably over a wider range of processing conditions such as those involved in the cryogenic separation of natural gases and in the separation of heavy hydrocarbon mixtures.

A mini-micro computer link promotes distributed processing in scientific and engineering computations

Engineering and scientific work often requires complex computations of different types that should be carried out on different types of computers on the one hand, but must communicate and support each other on the other. To facilitate efficient execution of such computations, a network composed of mini and micro computers (DEC PDP 11/40 and Intel 8080) with the support of a large computer (CDC CYBER 73) has been built and is described in this paper. The network enables the realization of involved scientific computations that could not be carried out by any one of the computers on its own. A general description of the entire set-up is followed by a more detailed discussion of the micro-mini link, and examples of projects using the network.

Direct gravity formula for the Geodetic Reference System 1967

A Direct Gravity Formula, polynomial in latitude, has been obtained from the series expansion of the closed gravity formula of Somigliana by a telescoping procedure. The use of the 7 coefficient formula gives a result as accurate as the closed expression. With 6 coefficients it gives an order of magnitude better accuracy than that of the widely used Chebyshev approximation and with 5 coefficients a result that is accurate to better than that of the conventional form of computing theoretical gravity. The derived approximation is not only simpler than other forms, but also at least 11 and 17 times faster on the CDC CYBER 74 computer than the Chebyshev approximation and the closed gravity formula respectively.

Algorithm for Optimal Activated Sludge Design

Nonlinear design equations are used to model an activated sludge tank, final settling tank, gravity thickener, anaerobic digester, vacuum filter, pumping and aeration in terms of the design variables activated sludge solids retention time, hydraulic retention time, recycle ratio, thickener underflow suspended solids concentration, and digester solids retention time, for steady-state conditions. Nonlinear cost equations are used to model the cost of the individual unit processes in terms of unit process sizes and capacities. The Box-Complex algorithm, a numerical search technique, is then used to determine the optimum sizes of the unit process to minimize the sum of capital and operating costs for a specified design inflow, effluent quality, and equipment limitation. Less than 3 sec of execution time on a CDC CYBER 173 using an FTN compiler are required to determine the optimum.

Simulateur digital des convertisseurs de puissance à thyristors

This paper deals with a digital simulator for the analysis of any thyristor converter. The analysis is based upon a topological method of formulation of state equations and power thyristors and diodes are modelled as inductances. The value of each thyristor inductance changes from a finite value equal to the di/dt limiting series inductance, while conducting, to an infinite value when blocked. The simulator takes into consideration the presence of the snubber circuits associated with each thyristor element. The simulator library which has been named ATOSEC I has been developed having in mind the creation of a tool for the design by computer aided analysis of power electronic converters. It has been presently implemented on a CDC CYBER and will be available to prospective users from the authors.

Numerical solution of Kramers-Kronig transforms by a fourier method

Title of program: KRONIG Catalogue number: ACMN Program obtainable from: CPC Program Library, Queen's University of Belfast, N. Ireland (see application form in this issue). Computer: CDC CYBER 72-26; Installation: Computing Services Unit, University of New South Wales. Operating system: KRONOS. 2.1.1 LEVEL 393 Programming language used: FORTRAN High speed storage required: 31k 8 No. of bits in a word: 60 Overlay structure: none No. of magnetic tapes required: none Other peripherals used: line printer (unit 3) for testing. No. of cards in combined program and test deck: 101 Card punching code: IBM029