previous | contents | next

118 Part 2 The instruction-set processor: main-line computers

Section 1 Processors with one address per instruction

organ requires additional temporary storage and introduces a synchronizing problem, and hence it is not being considered for the first model.

Since, at the beginning of the problem, the computer is empty, facilities must be built into the control for reading a set of numbers from a wire when the operator presses a manual switch. As each number is read from a wire into Ac, the control must transfer it to its proper location in the Selectrons. The CC may be used to count off these positions in sequence, since it is capable of transmitting its contents to FR. A detection circuit on CC will stop the process when the specified number of numbers has been placed n the memory, and the control will then be shifted to the orders located in the first position of the Selectron memory.

It has already been stated that the entire memory facilities of the wires should be available to the computer without human intervention. This means that the control must be able to select the proper set of numbers from those going by. Hence additional orders are required for the code. Here, as before, we are faced with two alternatives. We can make the control capable of executing an order of the form: Take numbers from positions p to p + s on wire No. k and place them in Selectron locations v to v + s. Or we can make the control capable of executing some less complicated operations which, together with the already given control orders, are sufficient for programming the transfer operation of the first alternative. Since the latter scheme is simpler we adopt it tentatively.

The computer must have some way of finding a particular number on a wire. One method of arranging for this is to have each number carry with it its own location designation. A method more economical of wire memory capacity is to use the Selectron memory facilities to remember the position of each wire. For example, the computer would hold the number t1 specifying which number on the wire is in position to be read. If the control is instructed to read the number at position p1 on this wire, it will compare p1 with t1 and if they differ, cause the wire to move in the proper direction. As each number on the wire passes by, one unit is added or subtracted to t1 and the comparison repeated. When p1 = t1 numbers will be transferred from the wire to the accumulator and then to the proper location in the memory. Then both t1 and p1 will be increased by 1, and the transfer from the wire to accumulator to memory repeated. This will be iterated, until t1 + s and p1 + s are reached, at which time the control will direct the wire to stop.

Under this system the control must be able to execute the following orders with regard to each wire: Start the wire forward, start the wire in reverse, stop the wire, transfer from wire to Ac, and transfer from Ac to wire. In addition, the wire must signal the control as each digit is read and when the end of a number has been reached. Conversely, when recording is done the control must have a means of timing the signals sent from Ac to the wire, and of counting off the digits. The 26 counter used for multiplication and division may be used for the latter purpose, but other timing circuits will be required for the former.

If the method of checking by means of two computers operating simultaneously is adopted, and each machine is built so that it can operate independently of the other, then each will have a separate input-output mechanism. The process of making wires for the computer must then be duplicated, and in this way the work of the person making a wire can be checked. Since the wire servomechanisms cannot be synchronized by the central clock, a problem of synchronizing the two computers when the wires are being used arises. It is probably not practical to synchronize the wire feeds to within a given digit, but this is unnecessary since the numbers coming into the two organs Ac need not be checked as the individual digits arrive, but only prior to being deposited in the Selectron memory.

6.8.2. Since the computer operates in the binary system, some means of decimal-binary and binary-decimal conversions is highly desirable. Various alternative ways of handling this problem have been considered. In general we recognize two broad classes of solutions to this problem.

First: The conversion problems can be regarded as simple arithmetic processes and programmed as sub-routines out of the orders already incorporated in the machine. The details of these programs together with a more complete discussion are given fully in Chapter 9, Part II, where it is shown, among other things, that the conversion of a word takes about 5 msec. Thus the conversion time is comparable to the reading or withdrawing time for a word- about 2 msec-and is trivial as compared to the solution time for problems to be handled by the computer. It should be noted that the treatment proposed there presupposes only that the decimal data presented to or received from the computer are in tetrads, each tetrad being the binary coding of a decimal digit-the information (precision) represented by a decimal digit being actually equivalent to that represented by 3.3 binary digits. The coding of decimal digits into tetrads of binary digits and the printing of decimal digits from such tetrads can be accomplished quite simply and automatically by slightly modified Teletype equipment, cf. 6.8.4 below.

Second: The conversion problems can be regarded as unique problems and handled by separate conversion equipment incorporated either in the computer proper or associated with the

previous | contents | next