223 2721 0_7094_II_CE_Instruction_Maintenance_Vol1 0 7094 II CE Instruction Maintenance Vol1
223-2721-0_7094_II_CE_Instruction_Maintenance_Vol1 223-2721-0_7094_II_CE_Instruction_Maintenance_Vol1
User Manual: 223-2721-0_7094_II_CE_Instruction_Maintenance_Vol1
Open the PDF directly: View PDF .
Page Count: 109
Download | |
Open PDF In Browser | View PDF |
Customer Engineering Instruction -Main tenance 7094 II Data Processing System-Volume 1 Introduction Component Circuits System and Functional Components Timing Customer Engineering Instruction-Maintenance 7094 II Data Processing System-Volume 1 . Introduction Component Circuits System and Functional Components Timing Preface Safety This manual contains four sections: IBM 7094 II Data Processing System, Component Circuits, System and Functional Components, and Timing. The purpose of these sections is to: 1. Introduce the 7094 II system. 2. Describe the various circuit cards and circuit configurations. 3. Explain briefly the purpose and operation of the various system components (such as the multiplexor, core storage, data channels, etc.), and internal functional components (such as registers, adders, counters, etc.). 4. Explain the 7094 II timing and cyclic makeup. Customer Engineers who are familiar with the 7094 system may easily skip over the first section by reading only the portion on Instruction Overlap. Condensed logic diagrams used in this manual are as close to actual systems as possible. Most of these diagrams have been converted to positive logic by eliminating any references to + or - levels. In maintaining this positive logic, in-phase outputs are used to indicate an active (conditions met) state from the condensed logic block. Out-of-phase outputs are also used in some cases to simplify the diagrams by eliminating the cluttering effect of convert and invert blocks. The material in this manual is written at engineering change level 253407; however, future engineering changes may change the logic and machine operations from the presentation in this manual. The following manuals pertain to the 7094 II system: The follOwing safety practices should be observed: l. At least two men should be within sight of each other when working on a machine with power on. 2. Safety glasses must be worn when soldering or performing other operations which may endanger the eyes. 3. Use caution when lowering a tailgate. Keep fingers clear of gate slides when sliding a gate into a module. Avoid hitting laminar bus connections. 4. 120 volts, 60 cycles, and 48 vdc are still present inside SMS frame with frame power off and 7618 power on. If it is necessary to work near live power connectors, convenience outlets, or inside the MG unit or core storage control, disconnect power cables, or turn off wall circuit breakers. 5. Discharge capacitors before working on DC power supplies. 6. Always turn off power before replacing a fuse. 7. Replace safety covers that have been removed before proceeding to another operation. 8. Prior to servicing, note and check the following items: Master power switch location ________ Air conditioning switch location _______ Fire extinguishers (C02 type )_ _ _ _ _ _ __ Emergency exit doors 10cationL-_ _ _ _ _ __ Fire control phone number_ _ _ _ _ _ _ __ First aid phone number _ _ _ _ _ _ _ _ __ 9. Remove metal jewelry before servicing the computer. FORM 223-2721 223-2722 223-2723 223-2724 223-6910 223-2551 TITLE 7094 II DPS CEIM Manual Volume 1 (Introduction, Component Circuits, System and Functional Components, and Timing) 7094 II DPS CEIM Manual Volume 2 (Arithmetic Instructions) 7094 II DPS CEIM Manual Volume 3 (Non-Arithmetic Instructions, Overlap, Trapping, Compatibility, 7151-2 Console) 7302-3 Core Storage CEIM Manual 7607 Data Channel CE I-R Manual 7909 Data Channel CE I-R Manual Copies of this and other publications can be obtained through IBM Branch Offices. Address comments concerning the contents of this publication to: IBM Corporation, CE Manuals, Dept. B96, PO. Box 390, Poughkeepsie, N.Y. 12602. © 1964 by International Business Machines Corporation Contents IBM 7094 II Data Processing System . ............. . COMPUTER WORDS. . ... . . . . . . . . . . .. . .. Data Word ....................................... . Instruction Word . . . .. Instruction Addressing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Direct Address. Effective Address. Indirect Addressing .. 7607 Data Channel Command Word. 7909 Data Channel Command Word. INSTRUCTION OVERLAP ....... . BINARY ARITHMETIC .... '. . ...... . Octal Number System .............................. . Addition . , . . . . . . . . . . . . . . . . . . . . .. ............ . Subtraction ..................... . ......... . Complement Method. . . ...... . l's Complement ................. . 2's Complement ....... . Signed Number ................. . Multiplication ........................ . Division ......................................... . Component Circuits . ................. . BASIC D1F CIRCUIT OPERATION ............ . AND-Invert .. . . . . . . . . . . . . . . . . . . . . . . . .. oR-Invert ........................... . AND-oR-Invert ................... . D1F CIRCUIT LOGIC BLOCKS. . . . . . . ........... Micro and Macro Blocks ................. . AND-OR-Invert ............................... . D1F (F-Level) Triggers. . . . . . . . . ...... . +AOI Trigger ................................... . -OAI Trigger. . . . . . . . . ................. . IBR Trigger ................... . Shift Cell-sc. . . . . . . . . . . . . . . . . . . . DOT-oR'ing and AND'ing. COMPONENT CIRCUITS CARD TYPES .... . SMS Single Card. . . . . . . . . . . . . . . .. . SMS Twin Card .. SMS STAN-PAC Card ................... . D1F CIRCUIT SPECIFICATIONS. . . . . . . D1F Logic Block ..... Logic Block Input Specifications. Logic Block Output Specifications. Logic Block Circuit Delays ..... Logic Block Power Supply Requirements ............ . Logic Block Extended Capabilities. D1F Indicator Driver .... Indicator Driver Input SpeGifications. . . . . Indicator Driver Output Specifications ...... . Indicator Driver Power Supply Requirements .. . N-Line to D1F Converter-Terminator. N to F Converter Input Specifications ............... . N to F Converter Output Specifications ............. . N to F Converter Delays .......................... . Power Supply Requirements ....................... . P-Line to D1F Converter-Terminator. . . . . . . ........ . P to F Converter Input Specifications ............... . P to F Converter Output Specifications. . . . . . . . . . . . . . . P to F Converter Delays. . . . . . . . . . . . . . . . . . . . . . . . . .. P to F Power Supply Requirements ................. . D1F to N-Line Converter-Driver. . ........... . F to N Converter Input Specifications. . . . . . . . . . . . . . . . F to N Converter Output Specifications . . . . . . . . . . .... F to N Converter Delays. . . . . . . . . . . . . . . . . . . . . . . . . . . F to N Power Supply Requirements ................. . 7 8 8 9 9 9 9 10 10 11 11 11 13 13 14 14 15 16 16 16 17 18 18 18 19 19 20 20 20 20 21 22 22 22 23 24 24 24 24 27 27 27 27 27 27 27 28 28 28 28 28 28 28 29 29 29 29 29 29 29 29 30 30 30 30 D1F to P-Line Converter-Driver. .. . ................ F to P Converter Input Specifications ............... F to P Converter Output Specifications .............. F to P Converter Delays. . ................... F to P Power Supply Requirements ................. . . . . . 30 30 30 31 31 System and Functional Components . ............. . 33 SYSTEM COMPONENTS. . . . ..................... . 33 7111 and 7109 Central Processing Units ............... . 33 7606 Multiplexor. . . . . . . . . . . . . . . . .. . .............. . 33 Multiplexor Storage Busses ........................ . 34 Multiplexor Storage Bus oR'ing ..................... . 34 Buffer Address Register. . . . . . . . .. . . . . . . . . . . . . . . . . . . 34 7302-3 Core Storage ............................... . 34 7302-3 Addressing ............................... . 36 Input-Output ..................................... . 36 7607 Data Channel ................................ . 36 7909 Data Channel ................................ . 37 7151-2 Console Control Unit ........................ . 37 7608 and 7618 Power Control Unit ................... . 37 FUNCTIONAL COMPONENTS ........................... . 38 Storage Register-sR ............................... . 38 Accumulator Register-AC .......................... . 41 Multiplier-Quotient Register-MQ. . . . . . .. . .......... . 42 Sense Indicator Register-sl ......................... . 42 Instruction Backup Register-IBR .................... . 43 Tag Register-TR .................................. . 44 Address Register-AR .............................. . 44 Program Counter-pc .............................. . 46 Index Registers-xR ................................ . 47 Program Register-PR .............................. . 47 Shift Counter-sc ................................. . 48 Stepping the Shift Counter. . . . . . . . . . . . . . . . . . . . . . . . . 48 Main Adders-AD .................................. . 52 Individual Main Adder Position .................... . 52 Main Adder Bit Carry Lookahead .................. . 52 Main Adder Group Carry Lookahead ............... . 53 Q Carry Lookahead .............................. . 57 Index Adders-xAD ................................ . 58 Index Adder Position. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 Index Adder Bit Carry Lookahead .................. . 59 Index Adder Group Carry Lookahead ............... . 59 Index Adder 3 Carry ............................. . 59 Index Adder Compatibility ........................ . 59 SR Zero Check ..................................... . 63 Memory Selection Circuits .......................... . 63 MAR Switching and Address Controls ................. . 64 Timing .......................................... . MASTER CLOCKS AND PULSES ......................... . CPU Clock Pulse Distribution. . . . . . . . . . . . . . . . . . . . .... . cpu-l A and B Gate Clock Pulse Designation ........... . CP Set Pulse Generation and Distribution .............. . MACHINE TIMING CYCLES ............................ . Master I Time .................................... . Address Gating. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Program Counter Update ......................... . Storage Bus Gating. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Address Modification ............................. . N ext Machine Cycle ............................. . Master E Time .................................... . IA E Cycle ....................................... . Master L Time .................................... . Master II Time. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Master B Time .................................... . B Cycle .......................................... . 66 66 70 70 72 73 74 76 77 77 78 78 82 82 86 86 87 89 Cycle. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. Cycle ........................... '.' . . . . . . . . .. Indirectly Addressed BCW Cycle. . . . . . . . . . . . . . . . . . . .. TCH Command. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. Indirectly Addressed TCH Command. . . . . . . . . . . . . Channel Cycle Times. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. Channel I Time ................................. , Channel E Time. '. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. Channel L Time. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. Channel-cpu Cycle Time Controls. . . . . . . . . . . . . . . . .. Multiple Cycle Time Error Detection. . . . . . . . . . . . . . . . .. BDW BCW WAVEFORMS AND VARIABLE DELAY ADJUSTMENTS. . . . . . . . .. Initial Delay Settings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. Odd and Even Clock Drive Pulses. . . . . . . . . . . . . . . . . . . .. 89 89 89 90 90 93 93 93 93 93 95 95 95 96 Set Pulses ..................................... '. . cpu-l CP Set Pulse Width Adjustment. ................ CP Set Pulse--Clock Pulse Alignment. . . . . . . . . . . . . . . . .. CPU and Channel Memory Select Alignment. . . . . . . . Memory Select and MAR Bus Alignment. . . . . . . . . . . . . .... 96 96 96 97 97 Appendix A: Octal-Decimal Integer Conversion Table ...... .. , ......... , .. ... .... ........ 98 CP Appendix B: Octal-Decimal Fraction Conversion Table . ................. , ...................... 102 Appendix C: Table of Powers of Two. . . . . . . . . . . . .. 105 Illustrations FIGURE TITLE PAGE I ntrod uction 1 General Computer Functional Arrangement .. 7 2 7094 II Data Word ............................. . 9 . .............. . 3 7094 II Instruction Word. . 9 4 Instruction Word Address Modification Fields ...... . 9 5 Instruction Word Count Fields. . . . . . . . . . . . . . . . . . .. 9 6 Data Channel Command Word ......... . 10 7 7909 Command Word Formats ........ . 11 8 Double Instruction Overlap .... . 12 9 Extended Sequence Overlap ..................... . 12 Component Circuits 10 +AI (-01) Circuit .................. . 11 +AND, -OR Circuit. 12 +01 (-AI) Circuit ......... . 13 +OR, -AND Circuit. 14 AND -OR-Invert Circuit 15 Macro and Micro Blocks.. . ....... . 16 Systems Logic vs Condensed Positive Logic Gating Circuitry . . . . . . . . . . . . . . . . . . . . . . . . . .. 17 + A01 Trigger . . . . . . . . . . . . . . . . . . . ...... .. 18 -OAI Trigger. . . . . . . . . . . . . . . . . ........ . 19 IBR Trigger ...................... . 20 Shift Cell. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Shift Cell Timing Chart ........... . 22 DOT-oR'ing and DOT-AND'ing ............. . 23 Twin-Card Locations for SR, AC and MQ Input Gating. 24 Twin-Card Locations for SI, IBR, XAD and XR ....... . 25 BDS Card Locations for SR, AC, MQ and AD. 26 Typical BDS Card Layout. 27 DIF Logic. . . . . . . . . . . . . . . . . . . . .. 28 DIF Indicator Driver ...... . 29 N-Line to F -Line Converter ... 30 P-Line to F -Line Converter. . . . . . . . . . . .. . ...... . 31 F-Line to N-Line Converter. 32 F-Line to P-Line Converter .... 21 21 22 22 22 23 23 24 25 25 26 27 28 28 29 30 30 System and Functional Components 33 7094 II System Configuration .................... . 34 Basic Functional Organization .. 35 Multiplexor Data Flow .................. . 36 7094 II Power Distribution. . . . . ...... . 37 7094 II CPU Data Flow . . . . ......... . 38 Storage Register Position 17. . . . . . . . . . . . . . . . . . . . . . 39 Accumulator Position 17. . . . . . . ............... 40 MQ Position 17. . . . . . . . . . . . . . . ........... 41 Sense Indicator Position 17 ...................... . 42 IBR Position 17. . . . . . . . . . . . . . . . . . . . . . . . . . . .. 43 Tag Register Position 18 ................. . 32 33 35 38 39 40 41 42 43 43 44 18 19 19 20 20 21 FIGURE TITLE PAGE 44 Address Register Position 17. . . . . . . . . . . . . . . . . . . . 45 45 Program Counter Position 17 .................... . 46 46 XRA Position 17. . . . . . . . . . . . 47 47 Program Register Position 9. . ................ . 47 48 Operation Decoding. . . . . . . . . . . . . . . . . . . . . . . . . 48 49 Shift Counter Stepping ......................... . 49 50 Shift Counter Timing Chart. . . . . . . . . . . . . . . . . . . . .. 51 51 Main Adder Position 35 ......................... . 52 52 Group 1 Adder Carry, Generate and Propagate ..... . 54 53 Group 2 Generate and Propagate Lookahead ....... . 55 54 Main Adder Group Lookahead and Q Carry ........ . 56 55 Index Adder Routing. . . . . . . . . . . . . . .-. . . . . . . . . . . . . 58 56 Index Adder Positions. ... . . . . . . . . . . . . . . . . . . . . . 60 57 Group 1 Index Adder Bit Carry and Lookahead .... . 61 58 Index Adder Group Lookahead and XAD( 3) Carry ... . 62 59 SR Input Zero Check. . . . . . . .. . .......... . 63 60 Memory Select Circuitry. . . . . .......... . 63 61 Program Counter MAR Bus Selection .. 64 62 MAR Bus Selection and Switching ................ . 65 Timing 63 Master Clock Logic. 64 5.71 MC Oscillator Output ... 65 Oscillator Output and Even Clock Drive .. . 66 CPU and Channel Cycle Relationship .............. . 67 CPU and Channel Clock Outputs and Controls. . . . . . . . . . . . . . . . . . .. 68 CPU Clock Sequence Chart 69 Channel Clock Sequence Clrart. ................ . 70 CPU Clock Pulse Distribution. .. . .............. . 71 CPU Clock Pulse Designations and Usage. . . . . . . . . . . 72 CP Set Pulse Distribution ....................... . 73 Cycle Time Relationship ........................ . 74 I, E, and L Time Cycle Logic. . .............. . . .................. . 75 I Time Flow Chart. . 76 SB Gating Decision Chart. . ................. . 77 IA Flow Chart. . . . . . . . . . ............. . 78 Inhibit L Trigger Logic. . . . ......... . 79 II Time Condensed Logic. .. . ........ . 80 B Time Condensed Logic ....................... . 81 B Time Sequence Chart ........................ . . ................. . 82 B Time Flow Chart. 83 Channel Cycle Time Logic ................... . 84 Multiple Time Check Circuitry .................. . 85 Clock Drive Pulse ............................. . 86 Even and Odd Clock Drive Pulses ................ . 87 CP Set Pulse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 88 CP Set Pulse-Clock Pulse Alignment. . . . . . . . . . . . . . 89 Memory Select Pulse Alignment .................. . 90 Memory Select-MAR Bus Alignment ............. . 67 68 68 68 68 68 69 71 72 73 74 75 79 78 85 86 86 87 88 91 94 95 96 96 96 96 97 97 • j:. ;1.. .. ". 'J."1 • • . '.' •• '." • .• ' ~_\i~ " . • . 7094 n Data ProCeSsing System " Jid~.. J ~';" ... ...... ... :.~, ... ~~, .. ~~ ... ".., . " , >. ~ ""~ ~,. ' ... • • • ',,:;;" . ",' .,' ••••• " .. . " ) I . . q Ii, Subtrahend -011 -010 - 001 011 011 000 Minuend Subtrahend Difference 100 010 110 100 011 III Minuend complement Subtrahend Complement of difference Minuend> Subtrahend +011 - 010 + 000 011 101 110 Minuend Subtrahend Difference Minuend 011 011 101 010 000 101 With a 1 end carry 1 000 110 Minuend Complement Difference - 1 True difference Introduction 15 same. The last requires the extra step of end-around carry, which is a carry from the highest order around to the lowest order. This carry is required because of the cyclical nature of the number system. The only time it is required is when the minuend is larger than the subtrahend, that is, when the answer will come out a true positive answer. Fortunately, whenever it is required, there is a carry from the leftmost position, which serves as a reminder. 2's Complement In the derivation of the complement system, it was shown that a 2's complement of a number is equal to the modulus minus the number, (M - N). Therefore, to obtain a 2's complement in a six-place machine, the number is subtracted from the modulus, 1 000 000. As an example, find the 2's complement of the numbers 101 001 and 001 101: aDO 101 1 010 ODD .00'1 III = 2's Complement 110 000 = Modulus 00'1 = Number 1 000 = Modulus 101 = Number all = 2's Complement An examination of the numbers and their complements shows that the 2's complement of a number is the same as the 1's complement with a 1 added to it. The 2's complement is therefore formed by obtaining the 1's complement and adding 1 to it. For example, to form the 2's complement of 001101: 0'0'1 110' 110 101 = Number OlD = 1's Complement 010 = l's Complement +1 110' on = 2's Complement To perform subtraction by the 2's complement method: l. Find the 2's complement of the subtrahend. 2. Add this complement to the minuend. The result is the difference in complement form if it is negative and in true form if it is positive. In the 2's complement system, there is no need to end-around carry. Signed Numbers How can negative numbers in complement form be distinguished from positive numbers in true form? In this regard, also, binary numbers offer an advantage with respect to representation. The sign of a number is binary in nature; th:at is, a number is either positive or negative. Thus, a bit representing the sign can be used in addition to the bits representing magnitude. A 0 in the sign bit position can be interpreted to mean that the number is positive. A 1 in the sign bit position can be interpreted to mean that the number is negative. By treating the signs separately from the magnitudes in each operation, the result sign can be predicted. Therefore, the rules of algebra apply in determining the result sign. 16 Multiplication The rules for binary multiplication are similar to those of decimal multiplication. The rules for multiplying two single digits are the same in both systems. These rules are: DXD=D DX1=D 1xD=0 1x1=1 The general procedure when multiplying two multiple digit binary numbers is the same as that in decimal arithmetic. That is, the multiplicand is multiplied by a digit of the multiplier, and the partial product obtained is placed so that the least significant digit is under the multiplier digit. When all the partial products have been found, they are added together to find the final product. The only difference between decimal and binary multiplication, therefore, is in the summing of the partial products. In binary, the binary addition table is used while in decimal, the decimal table is used. As can be seen from the following examples, the method of obtaining partial products and then adding them to obtain the final product is identical to that of decimal arithmetic. Multiplicand Multiplier First partial product Second partial product Third partial product Fourth partial product Final product 10'10 10'.11 1101 1010 0000 1010 1010 10'000010 100.1 1011 0000 0000 1011 1100'.011 1111 1111 1111 1111 1111 1111 11100001 Note the placement of the binary point in the second example. The same rules hold for its placement as hold for placement of the decimal point in decimal arithmetic. The third example also illustrates an interesting point. This is the multiplication of the two largest possible 4-bit numbers. The product is 8 bits long. In other words the largest product that can result from the multiplication of two numbers will be no longer than the sum of the number of bits in the multiplier and multiplicand. If a number is multiplied by the radix of the number system, this multiplication has the effect of shifting the number one place to the left with respect to the radix point. This is true in any number system. For example, multiply 12.51 10 by 10 (the radix of the decimal system) and multiply the number 10.112 by 2 (the radix of the binary system): Number Number times radix 12.51 125.1 10.11 101.1 Binary multiplication, then, is nothing more than a series of add and shift operations. An example of such an operation is given under Fixed Point Arithmetic in Volume 2. Division Binary division is the process of counting the number of times a divisor goes into a dividend. The count of the number of times the divisor may be subtracted from the dividend before a negative remainder occurs is called the quotient. Direct binary division is performed by a series of subtractions of the divisor (actually a multiple of the divisor), just as it is in the decimal system. For example, divide 100011100 by 1110: (bd ehi jk) 10 1 110 }100 011 (a) 11 10 (c) 111 (f) 111 (g) (1) (m) 100.01 100.00 1 0 100 00 ..1l...lQ. 10 In the example, the first step is to place the divisor below the dividend in a position which is as far removed to the left as possible (a), but which will allow a positive difference to result when the divisor is subtracted from the dividend. Since the divisor will go into this many bits of the dividend once, a 1 is placed in the quotient at b in the same column as the lowest order digit of the divisor. The divisor is then multiplied by the quotient digit, and the resulting product is subtracted from the dividend to produce the positive difference (c), called the current remainder. The next digit in the dividend is brought down to the line c. Compare the divisor to line c; note that the divisor is larger than line c, or that the divisor goes into line c o times. Therefore, place a 0 in the quotient at the d position. The next digit of the dividend is then brought down to line c. Comparing the divisor to line c shows line c to be greater. Place a 1 in the quotient at the e position. Multiply the divisor by the last quotient bit to form line f. Subtract line f from line c to start line g. The next digit in the dividend is brought down to line g. Compare the divisor to line g; the divisor is greater, so place a 0 in the quotient at position h. Bring the next digit of the dividend down to line g; by comparison line g is still smaller than the divisor. Place a 0 in the quotient in position i, and place the next dividend digit on line g. Still, line g is smaller than the divisor, so a 0 is placed in the quotient at position j. Placing the next dividend digit on line g now makes line g greater than the divisor. Place a 1 in the quotient at position k, and multiply the divisor by this 1 to form line 1. Subtract line I from line k to start line m. Assuming a quotient has been developed of sufficient length, terminate the operation. The quotient is 10 100.01 with a remainder of 10 (line m). Since the quotients bit is always either 0 or 1, the division process can be reduced to a series of subtractions of the divisor, multiplied by the power of the quotient bit being sought from the dividend. Each time a subtraction results in a positive current remainder, a 1 is placed in the corresponding quotient bit position, and the process is immediately repeated for the next quotient bit. Each time the subtraction results in a negative remainder, a 0 is placed in the corresponding quotient bit. In this case, the current remainder is restored to a positive number by adding the divisor back to it. Following this, the next quotient bit is obtained by the subtraction of the divisor multiplied by the power of the next quotient bit. Since the quotient bits are generated from left to right, the power of each quotient bit is one smaller than that of the last bit generated. This means that as the divisor is successively subtracted from the dividend (or current remainder), the divisor is shifted to the right in relation to the binary point. The division process can therefore be reduced to a process of successive subtract and shift steps. An example of such a process is given under Fixed Point Arithmetic in Volume 2. Introduction 17 Component Circuits +3v All logic in the A and B gates of CPU 1 (7111) is composed of DIF (DIode Feedback) circuitry. This new Flevel circuitry is of the non-saturating type and provides the three basic logical functions of AND, OR, and invert. A collector-to-base diode feedback network prevents transistor saturation and, therefore, allows highspeed circuit operation. Voltage inversion always occurs between the input and output. +6v AND Output Input 1 Basic DIF Circuit Operation Three logical functions are performed by the DIF circuit block: AND'ing, oR'ing, and inverting. Other functions are also performed such as terminating, driving, and converting from one voltage level to another, but these are only "convenience" functions, not logical functions. DIF circuitry allows both the AND and OR logical functions to be performed within the same logic block before output powering and inversion. Not all logic blocks contain both functions, however; some produce just AND'ing, others just oR'ing, while others are simply inverters. AND-Invert The two-legged circuit shown in Figure 10 represents either a + AI or -01 function. The circuit configuration for the + AND and -OR are identical. The polarity designations are adapted to work in negative logic; that is, the recognition of the absence as well as the presence of information. The +AI circuit requires that all inputs must be up (+3 volts) to obtain a down-level output (0 volts). Note that even though this circuit represents only an AND function; an OR diode, D3, is also included. This one-legged OR circuit serves no logical function, but is always a part of the +AI configuration. Figure lla shows the junction point of the two input AND diodes. Only two inputs are shown in this case-three or more could also be used. The resistor (Rl) limits the current How and controls the rise time of the output. If both inputs are at 0 volts (Figure lIb), the polarity is correct for both diodes to conduct. The resultant current How through Rl causes a voltage drop across it to maintain a level of about 0 volts. (Consider the forward resistance of the diode to be insignificant.) 18 Input 2 Input 1 Input 2 I , I I I '+3v Output ! I lov Figure 10. +AI (-01) Circuit If input 1 rises to +3 volts (Figure llc), Dl is cut off because the cathode is more positive than the plate. D2, with 0 volts on its cathode, maintains conduction and the output remains unchanged. When input 2 changes to +3 volts (Figure lId), D2 is cut off momentarily. The junction voltage starts rising towards +6 volts with a rise time effected by the stray capacitance of the associated circuitry. As the junction reaches 3 volts, both diodes return to conduction as a final steady state condition. If the input signals are not of the same voltage levels, the junction voltage assumes the lowest of the input sources. When input 1 falls to 0 volts (Figure lIe), Dl conducts harder, D2 is cut off, and the output follows the input down to 0 volts. When input 2 falls to 0 volts, D2 goes back into conduction to help maintain the 0 volt output level. At the output of every DIF logic circuit is a powering transistor which automatically causes a voltage inversion. When all inputs are down (0 volts), the base-toemitter voltage keeps the transistor out of conduction. In this condition, the output signal level is essentially at that of the collector power source, +3 volts. When both inputs 1 and 2 are at +3 volts, for example, diode D3 conducts and causes the voltage at the base of the transistor to go more positive. In this condition, the transistor conducts and the output signal is essentially that of the transistor emitter, 0 volts. If the active logical input lines are represented as plus (+3 volts) levels, the active logical circuit output is a minus (0 volt) level (Figure 10). If this output is fed into a-AND or -OR circuit, however, the inversion was not a "logical" inversion. Diode D4 in the OR circuit performs a resistance function only and does not affect the logic. D5 is the diode which clamps the collector at a more positive voltage than if the transistor were allowed to conduct fully. By preventing the transistor from going into saturation, better waveshapes are produced to allow faster computer operation. Note that even though this circuit represents only an OR function, two corresponding AND diodes (D3 and D4) are included. These one-legged AND circuits serve no logical function, but are always a part of the +01 configuration. Figure 13 shows the junction point of the two input OR diodes (D5 from Figure 12 has been eliminated at this time). If both inputs are at 0 volts (Figure 13b), the polarity is correct for both diodes to conduct. The voltage drop across the current limiting resistor (R2) sets the correct output level-for explanation purposes, ovolts. If either input rises to +3 volts (Figure 13c), that circuit leg conducts harder. The other diode cuts off and the output follows the input to +3 volts. Normally only one input is active at anyone time. The junction voltage, however, tends to follow the highest of the input signal levels. OR-Invert The two-legged OR circuit shown in Figure 12 represents either a +01 or -AI function; both differ in logical function but are identical in circuitry. The +01 circuit produces a down-level output (0 volts) if anyone of the inputs is up (+3 volts). This inversion, as explained previously for the AND-invert, is caused by the output powering transistor. +6v AND-OR-Invert Figure 14 shows the previously explained AND and OR functions combined into a typical DIF + AOI circuit. This circuit, too, can be considered as a -OAI depending on either the inputs available, or the output level desired. For example, if a minus (-) output level is desired, a +AOI designation would be used; if a plus (+) output level is desired, a -OAI would be applicable. +6v Rl Rl ANO Output Input 1 Ov Input 2 Ov 01 J- 02 .Jt (0) AND Ov +6v INVERT +3v R3 (b) Input 1 +6v I - - - - AND- -:(;:--, +6v I Rl +3v Ov Rl Ov 01 01 - Rl +3v ~ 02 +3v I N I Input 2 (c) Output I I +3v 02 D6 I (d) N R2 +6v Rl } Ov -3v Input 1 Ov -.-J Input I I Input2~ I 1+3v +3v ~~----~~---- ~ Input 2 I I ANOOutPut~ Output I I I I I I ~ (e) Figure ll. + AND, -OR Circuit Figure 12. +01 (-AI) Circuit Component Circuits 19 Ov Input 1 D4 Ov OR Output Ov Input 2 (To Transistor Base) R2 R2 (a) (b) -3v -3v +3v Input 3 +3v ~~~--~~ Ov ---.J I Input4~ I R2 (c) I ~ OR Output ~~--- . ~ -3v Figure 13. +OR, -AND Circuit DIF Circuit Logic Blocks 1. Two 2-way AND'S, 2. Three 2-way AND'S, 3. 3-way and 2-way AND, 4. 3-way and I-way AND, etc. In cases of negative logic (or because of the availability of signal lines ), a +AOI can be represented by a -OAI. In either case, the internal circuitry is identical; only the final active output level will be changed. The +AOI circuit is used in many cases to "gate" information into registers. In these cases the +AOI is replaced by a G within the circuit block. Most circuit diagrams shown in this manual are condensed ALD systems with logic converted to positive (active) logic. Figure 16 shows at comparison of systems logic (a) and condensed positive logic (b). Note that the condensed logic is not concerned with voltage levels, only active logical levels. The out-of-phase (inverted) output of the systems gating circuit feeds a -sc; therefore, the inversion is not a logical inversion. Because there is no logic to the inversion, the condensed logic, therefore, uses an in-phase (active) output from the gating (AO) circuit. Micro and Macro Blocks DIF (F-Ievel) Triggers Both "macro" and "micro" logic blocks are used on the 7094 II systems pages; the type of block used depends mainly on the availabilIty of printing space. Macro blocks can incorporate two or more micro blocks, but only when there are no pin connections and back panel wiring between micro blocks. As an example, consider the AND-OR-invert (AOI) circuit described in Figure 14. Figure 15a shows a macro block condensation of the corresponding three micro blocks at (b). Inputs to the macro block are "pinpointed" to specify input pins so that the AND functions are grouped and easy to separate visually. Of the eight possible inputs to the logic block, for example, the top AND function is pinpointed to input pins 2 and 3; the bottom AND function is pinpointed to input pins 6 and 7. Input pins 4 and 5 provide a visual separation of the two AND functions. Note that the i~-phase output is used from the two AND (+ A) micro blocks. These two blocks represent diode circuitry only (see Figure 14). The out-of-phase output is used from the +01 because the transistor is located at this point. Triggers (also referred to as "latches") act as storage or remembering devices. Once they are turned on, they remain in that state until turned off. Triggers are used in some cases to form registers (tag register and address register, for example); in other cases they are INVERT Input 1 I r --~I I Rl I -------' I Input 2 The +AOI was described in a previous section. Because of the limited number of logic block input pins, only certain combinations of AND-OR functions can be shown without losing the visual spacing. Obviously, then, it is not possible to show two 4-way AND'S. Combinations used include: 20 Output +6v D3 Input 3 D6 I I Input 4 I I I AND Input 1 AN D-OR-I nvert +3v +6v ..----... -----' ~----- Input 2 - - - - - - l Input 3 -.J Input 4 - - - - - . , ; : . . . . . . - - - - . . . : . . . . - - - - - - Output ..+_3v_ _.....1 Ov Figure 14. AND-oR-!nvert Circuit used singly to retain a specific condition (adder Q carry and MQ overflow, for example). Various trigger configurations exist in the 7094 II. + AOI Trigger The +AOI is the most common type of trigger used in the new F-Ievel circuitry of the 7094 II. This circuit, Figure 17, uses a +AOI followed by an inverter. One or more input AND conditions can exist, usually consisting of a data input and a gating input. The output from the inverter is fed back to the input circuit and AND'ed with the reset conditions to form a latch (or hold) circuit. Outputs from the + AOI and inverter blocks indicate the trigger ON and OFF conditions, respectively. In Figure 17a, either data input AND being fully conditioned causes the output of the +AOI to go minus (-) and the output of the inverter to go plus (+). This inverter output is fed back to AND with the reset pulse line forming a hold circuit. When the original data input AND circuit is deconditioned, the hold circuit keeps the trigger on. The trigger remains on until such time that the reset signal goes minus (-) and breaks the hold circuit. Figure 17b shows the trigger in positive condensed logic. In-phase and out-of-phase trigger outputs indicate an ON and OFF state, respectively. Resets are shown at the bottom of the trigger block with an additional section added for each additional reset condition. r + _In-'-pu_t_l_ _ _ _ D _ Input 2 E_ Input 3 Input 4 F - 1 2 3 PinS ----r - -- +AOI I TPU 6 BDS lB IC + F Gate AD Left AC Line 4 + FAD 18 + F Gate AC Left Ln 4 + F AC 18 B Out-of-Phase Outputs +G BBW + F Gate AC Left 2 Ln 4 + F AC 19 6 2C Gate AD - AC AD17 Gate AD Rt - AC AD16 Gate AD Rt 2 - AC ADl5 A In-Phase Outputs 0 A A Gate AD Left AC AD18 Gate AC left ACl8 Gate AC Left 2 AC19 A SC (IB) 0 A A (2C) Figure 16. Systems Logic vs Condensed Positive Logic Gating Circuitry I Sample Tgr r--- +AOI I 1 + F Sample Tgr On '---- 3D + F Data 2 l + F Data 2 Gate + F Sample Tgr Off -4D- 3 _ A. _ _ _O_u-.:tp_u_t ZHQ Internal Connection +A Data 1 -=D-=a.:..::ta~I_G;:;.a::.:t~e---:J.... A 0 Data 2 Input 2 _D.. :,a. . :.ta.. .:.2. . :.G,;."a.:. t. .;,.e---:J.... A (4D) Input 3 -T -U 4 5 6 G-7 8 D + F Gate AD-Rt2 AC Ln3 + FAD 15 vtJ +G + F Data 1 + F Data 1 Gate (a) Input 1 Gil ~ GD ~ B + F Gate AD-Rt AC Ln3 + F AD 16 - F Reset SMS Card lnput/outPut Pins Logic Block (Input(output + F Gate AD-AC + F AD 17 A _ _O_ut-,-pu_t_ +A Sample Tgr (Out-of-Phase) (In-Phase) Trigger Off Trigger On Reset _In.......p_ut_4_ _ _ _ G Z HQ t------t Figure 15. Macro and Micro Blocks Figure 17. +AOI Trigger Component Circuits 21 - OAI Trigger The -OAI trigger is sometimes used when + F input levels are not available and - F levels must be used. Figure 18 shows one example of a -OAI trigger that can be turned on from two sources and reset by one reset signal. Note that all of the active input levels are - F. The overall logic remains the same as the + AOI trigger discussed previously. IBR Trigger The !BR trigger circuit is packaged four to a twin card. On systems pages the trigger is shown as two + T blocks; two blocks are necessary because of the limited number of input pins available on an individual block. The trigger is also a macro block because the output inverter is not individually shown (Figure 19). The outof-phase output of the + T block originates before the inverter; the in-phase output of the +T block originates after the inverter. The address portion of the !BR (21-35) requires an additional input gate. The output of the + AI block is connected to the trigger output at a point in front of the output inverter; therefore, a hold circuit is established. The hold circuit is from pin 6 of the output inverter to pin 6 of the lower + T block. These two pin 6's Sample - F Set Pulse J - F Set Pulse 2 Tgr - F Tgr On -OA I t--__....----I are common points of internal card wiring; there is no back panel wiring as might be concluded from the systems page layout. Shift Cell-SC The shift cell is an element which can accept new data at its input while it simultaneously supplies old data at its output. Shift cells make up the storage register, accumulator, and multiplier-quotient (MQ) register in the 7094 II. These shift cells are a "double-latch" type of circuit where two -OAI triggers are connected in series. Both "set" and "hold" pulses are generated by control circuitry and used to introduce new data. Figure 20 shows the shift cell configuration and sequence chart for setting in a 1 from an initial 0 condition. Note that the active levels of the input data, set, and hold pulse are -F. The appropriate data level must be present at the input to the shift cell prior to arrival of the set and hold pulses. The final output changes on the lagging edge of the pulse. The set and hold pulses are approximately 60 nanoseconds in width and occur at the end of the 175 nanosecond clock pulse. The hold pulse is generated from the inverted output of the set pulse, skewed by one level of circuit delay. If the shift cell initially contains a 0, point B (Figure 20) is plus (+) and O2 is deconditioned. When a minus (-) data signal arrives at the shift cell, 0 1 becomes con- + F Reset 2D 2C - F Tgr Off Figure 18. -OAI Trigger - F SC D r---------------~~~ ~y t- T IBR I 35 +AOI t------t LD TPX ~X t-u TPX L_...L_I _ _ _ _ _ _ _ _ _ I -.J - F Set )icro - F Hold r-------------·I IBR 35 ----I~yo ---+I-'T +T 7-------_~I ----- ~ -..J~ Set Pulse 1 level of delay Hold Pulse : --: I Point B r--121.evels ~~tay 1 level of delay Point C Point D Figure 20. Shift Cell ~ - :..- 2 levels of delay I =1 ditioned. When the set pulse goes minus (-), both inputs are conditioned; point A goes plus (+), point B goes minus (-) and a data bit is set into the first half of the shift cell. The minus (-) signal at point B feeds 0 5 and 0 6 , but the final condition at 0 7 is blocked because the hold pulse has gone plus (+). The incoming data bit is not allowed in the second half of the shift cell at this time because doing so would destroy the old data and defeat the purpose of the shift cell. The minus (-) output of the inverter at point B does, however, feed back to O 2 to act as a hold on the first half of the shift cell when the input data signal is removed. At the end of the clock pulse, the hold pulse goes minus (-) and conditions 0 7 , the final input to the second half of the shift cell. At this time, point C goes plus (+), point D goes minus (-) and the shift cell indicates a 1 output. Point D is also returned to both 0 6 and 0 7 to form a hold circuit for the second half of the shift cell. Figure 21 shows a sequence chart of a shift cell, AC(35) for example, under the following conditions: 1. Initially reset to a 0 state 2. The initial 0 replaced by a 1 3. The 1 replaced by a second 1 4. The second 1 replaced by a 0 5. The last 0 replaced by another 0 Figure 22a shows the DOT'ing function with a threelegged OR and a two-legged AND. The top three-legged OR circuit supplies the collector load for both itself and the unloaded two-legged AND circuit below. Output connection between the two circuits is made by back panel wiring. Note that the logical input and output levels desired (i.e., +F or -F) will determine whether the DOT'ing function is an AND or OR. Figure 22b shows a DOT-OR' ed configuration. Any + F input to the +01 circuit causes transistor T1 to conduct and produce a - F output; also, both inputs being + F at the +AI circuit cause transistor T2 to conduct and produce a - F output. Therefore, if the logical output level is - F, either circuit block produces the required output. Figure 22c shows the same identical circuit as (a) but as a DOT-AND configuration. All inputs being -Fat the -AI circuit force transistor T1 out of conduction and produce a +F output; either input being - F at the -or circuit forces transistor T2 out of conduction and also produces a + F output. If the input conditions are not as just described, either Tl or T2 will conduct and produce a - F output which is opposite to the logical output desired. ,-------------------. +3v I I DOT-OR1ing and AND1ing I I In many cases, the available circuits do not contain enough inputs to satisfy a logical function. To meet this requirement, two or more existing circuits can be DOT'ed together so that the function of one logical block is «extended" into the other. DOT'ing is accomplished by having the circuits share a common transistor load, and because of this, some circuit card types do not contain transistor collector loads. I I Input I Back Panel Input 2 /wirin g I I Clock Output -3v I I I -------1 t------Data Input I I I Set Pulse I I I Hold Pulse I I L _ _ _ _ _ _ _ _ _ _ _ _ _ _~v_ _ _ _-=_...1I Point B (a) Point C ~~~__r-_,~--~~.~-F~I~np~ut~I--. .- ~~__~~ -F Input 2 -F Input 3 Point D t-----hAC(35)= 1 AC(35)= 1 .-----i-------i AC(35)=O AC(35)=O AC(35)=O -F Input 5 (b) Figure 21. Shift Cell Timing Chart Dot-AND -F Input 4 (c) Figure 22. DOT-oR'ing and DOT-AND'ing Component Circuits 23 Component Circuits Card Types Three types of SMS component circuit cards are used to support the DIF circuitry; single, twin, and STAN-PAC cards. In many cases, register positions or similar functions are combined on cards as both a packaging and trouble-shooting convenience. SMS Single Card All electronic components are mounted on the front side of the card and connections to the components are made on the back side by printed wiring patterns. The 16 contacts (labeled A through R) couple the signal and service voltages to the circuit components when the card is inserted into the SMS socket. These single SMS cards form the bulk of the computer logic. They contain: basic circuit elements such as AND'ing, oR'ing, inverting, and terminating; and semi-specialized circuit functions such as adder lookahead and gating. SMS Twin Card The twin SMS card is one physical card which requires the panel space of two single SMS cards. The use of twin cards provides more circuitry in a given space (compared to single cards) and is desirable in high speed circuitry because more operations can be per- Reg Pos S 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 01B4F27-28 01B4F27-28 01B4F27-28 01B4F27-28 01B4F27-28 01B4F27-28 01B4F25-26 01B4F25-26 01B4F25-26 01B4F25-26 01B4F25-26 01B4F25-26 01 B.4F07-08 01B4F07-08 01B4F07-08 OlB4F07-08 01B4F07-08 01B4F07-08 01B3F17-18 01B3F17-18 01B3F17-18 01B3F17-18 01B3F17-18 01B3F17-18 01B3H06-07 01B3H06-07 01B3H06-07 01B3H06-07 01B3H06-07 01B3H06-07 01B3H04-05 01B3H04-05 0IB3H04-05 01B3H04-05 01B3H04-05 * * * * '.\t * * * * * 24 The STAN-PAC card is identified by its vertically mounted components. Resistors, diodes, chokes, and so forth have their top terminal welded to a component mounting strip which clamps to the body of the component for mechanical strength. The strip also provides an electrical path to the adjacent component. Both terminals of the components pass through a hole in the card and are soldered to a land pattern on the reverse side of the card. The 32 contacts on the card (labeled A through Z and 1 through 8) couple the signal and service voltages to the circuit components when the card is inserted into the SMS sockets. BBW TPU * * * * * * * 01B4A05 01B4A05 01B4A05 01B4A05 OlB4A05 0lB4A05 01B4A07 01B4A07 OlB4A07 01B4A07 01B4A07 0lB4A07 01B4A08 OlB4A08 01B4A08 01B4A08 OlB4A08 0lB4A08 * * * * * * 01B4A04 OlB4A04 01B4A04 PIB4A04 0lB4A04 OlB4A04 01B3A25 0lB3A25 OlB3A25 01B3A25 0lB3A25 01B3A25 01B3A22 01B3A22 01B3A22 OlB3A22 OlB3A22 OlB3A22 OlB3A20 0lB3A20 OlB3A20 0lB3A20 0lB3A20 0lB3A20 01B4A06 0lB4A06 0lB4A06 0lB4A06 0lB4A06 0lB4A06 0lB3A24 0lB3A24 0lB3A24 0lB3A24 0lB3A24 0lB3A24 0lB3A23 0lB3A23 0lB3A23 0lB3A23 0lB3A23 0lB3A23 0lB3A21 0lB3A21 0lB3A21 0lB3A21 0lB3A21 0lB3A21 * * * 0lB3A19 0lB3A19 0lB3AI9 0lB3A19 0lB3AI9 01B3A19 0lB3A 18 0lB3AI8 01B3A18 01B3A18 0lB3A18 0lB3A 18 0lB3AI7 0lB3A17 0lB3A17 01B3A17 0lB3A17 0lB3A17 0lB3A16 0lB3AI6 01B3A16 OlB3A 16 01B3A 16 0lB3AI6 * Indicates the use of single cards. Figure 23. Twin-Card Locations for SMS STAN·PAC Card MQ Accumulator BBW TPU * 1 2 3 4 5 * Storage Register TPU formed before the resultant signal must be directed to other cards by way of connectors and back-panel wiring. The 32 contacts on the card (labeled A through Z and 1 through 8) couple the signal and service voltages to the circuit component when the card is inserted into the SMS sockets. Circuitry using twin cards includes: input gating to the storage register, accumulator and MQ; instruction backup register; sense indicator register; index registers; and index adder positions. Register positions and twin card locations are as shown in Figures 23 and 24. SR, AC, and MQ Input Gating * * * * * * * * * * * * * * * * * * * * * * * * * * * The various BDS card register positions and corresponding machine locations are shown in Figure 25. Figure 26 shows a STAN-PAC BDS card. To achieve the fast adder time of one clock pulse (175 nanoseconds), corresponding positions of the storage register, accumulator, MQ, and main adder have been combined on one STAN-PAC circuit card. Figure 26 shows position 35. This card position contains the shift cells for SR ( 35 ) , MQ(35), and AC(35). One input gate (+C) circuit for is also included; the remaining gates for the SR, and AC are on other cards and DOT-OR'ed at the shift cell as shown. The corresponding main adder position also has its input gating circuitry and output lookahead functions (propagate, generate, and exclusive OR) on the same card. Note, however, that the actual adder sum output logic is on a separate card. Note, also, that the top circuit of the bottom adder gate is not used for AD( 35). SR ( 35) MQ, SR,AC,MQ,AD Position Card location (BDS) SR(S),AC(S, P),AD(P) 1 01B4F22 0lB4F21 01 B4F20 01B4FI9 0lB4FI8 0lB4FI7 01B4FI6 0lB4FI5 0lB4F14 0lB4FI3 0lB4FI2 OIB4FIl OlB4FIO 0lB4F09 OlB4F06 01B4F05 01B4F04 01B3F25 OlB3F24 OlB3F23 OlB3F22 OlB3F21 01B3F20 01B3FI9 01B3FI6 0lB3FI4 0lB3FI3 0lB3FI2 OlB3Fl1 OlB3FIO OlB3F09 OlB3F08 0lB3F07 01B3F06 0lB3F05 0lB3F04 2 3 4 5 Sense Indicator Register Card Location (TPW) Instruction Backup Register Card Location (TPX) 5-3 4-7 8-11 12-15 16-19 20-23 24-27 28-31 32-35 01B2H28 01B2H27 01B2H26 01 B2H25 01B2H24 01B2H23 01 B2H22 01B2H21 01B2H20 5-3 4-7 8-11 12-15 16-19 20 21-23 24-27 28-31 32-35 01A3D12 01A3D11 01A3D10 01A3D09 01A3D08 * 01A3D07 01A3D06 01A3D05 01A3D04 Index Adder Position Card Location (TPS) 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 01A3F18 01A3F17 01A3F16 01A3F15 01A3F14 01A3F13 01A3F12 01A3Fll 01A3F10 01A3F09 01A3F08 01A3F07 01A3F06 01A3F05 01A3F04 (a) (b) *Indicates the use of single cards Index Registers Position (c) (d) 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Figure 24. Twin-Card Locations for 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Card Location (TPV) 2I 22 23 24 25 26 27 28 29 30 31 32 01A1G18 01A1G17 01A1G16 01A1G15 01A1G14 01A1G13 01A1G12 01A1G11 01A1G10 01A1G09 01A1G08 01A 1G07 01A1G06 01A 1G05 01A 1G04 SI, IBR, XAD, and 33 34 35 XR Figure 25. BDS Card Locations for SR, AC, MQ and AD Component Circuits 25 : :t]GD~>--D-~"'F +F Gate AD ...... SR +F AD 35 +F Gate AC ~ SR ~---C E r-- F +F SR 35 ,--K- G r-H - _1------. _+.:...F--:G;....a_te_M--:Q-=--...... _S_R_ _ _ _ _ 1 _ _-4-_ _ E ,-L -F {Dot OR of Other Input Gates} ....,D.,--+_-+--------I -F Set SR ~K~+_-+------------~ -F Hold SR ..;.H-=---+_--+---------...J - - --- --- --- -i-- ---- -F {Dot OR or Input Gates} M -F Set MQ -F Hold MQ P N -F {Dot OR of Input Gates} Storage Register {;35} 02.01.09.1 -- -- -- -- rr --~ - - - - - - - - - - MQ 35 V ---4------------------- v- -~F~S~e~t~A--'C~-----------I-T~+---------------------T....:-F_H;...;.o;:.,;l..:;.d;....A.;,.,:C'---_ _ _ _ _ _ I-=U'----i-----------------U ---=sc AC 35 G -- L +F MQ 35 G +F AC 35 ~~_r---------------------- ---- ---- ---- -- -- -- MQ Register {35} 02.04.06.1 ----+-+------------------------ __ Accumulator Register {35} 02.03.08.1 -- -- -- --- -- -- -- -- -- -- -- --- -- -- --- --- --- --- --+F Gate AC ~AD +F Gate Comp AC _R__-+---+--+-+ R ~+--+--+ G ~AD -----+~_+ ~ +G S, 1....-1-- ...---If---J-II --0011 ~ X ________________________________.:..:.X rz~ +F Propagate 35 I I I I {Not Used} -z - - (Not Used) 6 ----------,{Not Used} 7 +-----+---1 AD 35 +F Gate Comp SR +AD I 6 - _2_1--+-+-: +G I I L ~ ~-~y y +F Generate 35 ----4J---- I Q +F Gate SR+AD I ---+--~,--Q I....-F-_ I I I Ground 1 -+-3....:v-o~lt-s---- ~~ +6 volts (MC) -3 volts Ground _5_ _ _8_ _ _J_ _ I I I ~ ~I ~w ~ 1- Main Adder (35) 02.02.19.1 BDS - Card Location 01 B3 F04 Figure 26. Typical BDS Card Layout 26 W +F Exclusive OR 35 Dff Circuit Specifications DIF Logic Block Figure 27 shows a typical AND-OR-invert (AOI) DIF circuit. Diode Dl and the dotted circuitry below it indicates the AND function; diode D2 and the dotted circuitry below it indicates the OR function. Diode D3 produces an additional voltage drop in its circuit and is used because it provides better control than a resistor. Diode D4 provides the transistor collector-tobase feedback network to prevent the transistor from going into saturation. The inductor (L) in the transistor collector circuit is used to improve the output signal rise time by overcoming stray capacitance associated with the output circuit. Inverter circuits are formed by using one-legged AOI circuits. -Dfj --E +AOI F F A-- _KME_+ Output Dl { ---., I ------l - - ---1 --- .... ---~ ----~ ___ ...1 ------I -3v Figure 27. DIF Circuit delays vary as a function of: the number of fanin Signals, the number of fan-out signals, and the total length of back panel wire. The total best-case to worstcase delays vary from approximately 5.0 to 20.0 nanoseconds. These delays apply to the logic block which is performing two logic functions (AND-OR). Logic Siock Power Supply Requirements Nominal Voltages and Tolerances: SUPPLY + 3 volts - 3 volts 3V U +6v Fan-in Logic Siock Circuit Delays +6M --F --G Input Logic blocks Line drivers Indicator drivers Inverter (only one )-equivalent to three loads Logic Logic Siock Input Specifications DC Voltage Levels and Limits: TOLERANCE ± 4 percent ± 4 percent ± 4 percent Overvoltage and Undervoltage Limits and C onditions: 1. The following overvoltages can be tolerated without causing component damage: SUPPLY OVERVOLTAGE LIMIT +6M + 3 volts + 9.0 volts + 3.6 volts - 3 volts - 6.0 volts 2. Any power supply can be open-circuited without causing component damage. 3. Any power supply can be shorted to ground without causing component damage. 4. Power supply sequencing is not required. 5. Logic block cards may be inserted into or removed from the computer with power on without causing component damage. 3.35 v 2.88 v 0.93 v 0.47 v / - Fan-In Capabilities: The logic block can have a fanin of five on the AND circuitry, and five on the OR circuitry. Inputs can be expanded, however, as explained later. Logic Siock Output Specifications DC Voltage Levels and Limits: r 3.35V 2.88v 0.93 v 0.47 v Fan-Out Capabilities: The logic block has a fan-out capability of ten of the following in any combination. Logic Siock Extended Capabilities Eight-Way AND: An eight-way AND fan-in may be used with a maxinlum of one nanosecond delay added per block. Dot-OR Connections: Logic block collectors may be DOT-oR'ed to increase the OR fan-in. Up to two additional collectors may be connected to a logic block output; the DOT-OR' ed collectors will have a single collector resistor and inductor. Each additional DOT-oR'ed collector adds 1.5 nanoseconds to the delay of the original circuit. Frequently all of the circuit diodes are not used. In these cases: 1. AND diodes can "float" 2. OR diodes must be tied to ground Component Circuits 27 DIF Indicator Driver The DIF indicator driver (Figure 28) is a saturating circuit designed to indicate an output level of the DIF logic block. The indicator light will be on when the logic block output is at the "up" (+3 volts) level, and off when the logic block output is at the "down" (0 volts) level. A small amount of "pre-energization" current continues to flow through the indicator lamp even when off and causes a faint glow at the indicator filament. U ~F ---A Power Supply Limitations: 1. Power supply sequencing not required. 2. Duty cycle is 100 percent. 3. Card can be removed from the computer with power on without damage to the circuit. N-Line to DIF Converter-Terminator The N to F level converter provides proper termination for N-line (drift current mode) logic blocks (Figure 29). The converter also converts the N-line level to one capable of driving an F-Ievel (voltage mode) logic block. The circuit is designed to be driven by 200 feet of coaxial cable from a B-type logic block. r------, ~ :~ : +6M R4 r-----1a----vVV.... 1 L Rl 820Q R5 +30v _L~ _ _ I The circuit portion within dotted lines is external to the indicator driver card assembly. 3.3K ---H .J r::l BA-LJ-D-- U _p +3v R8 150Q .. Converter L 3.3ttH +6v +3v +12v :lI----+-~ Out -3v Figure 28. DIF N Indicator Driver p N Indicator Driver Input Specifications DC Voltage Levels and Limits: 3.12V 2.85 v f + 0.93 v + 0.47 v In _ _ " 1 / \ 1 ' , - - _ . Loading: The indicator driver must be considered as at least three loads. -12v Indicator Driver Output Specifications Figure 29. N-Line to F-Line Converter The output specifications for this circuit are determined by the particular indicator lamp being driven. These specifications are for driving indicator lamp PiN 550511 connected as shown in Figure 28. DC Voltage Levels and Limits: Lamp OFF f · 29.2v 24.2v Lamp ON 22.8v 15.0v Indicator Driver Power Supply Requirements Voltages and Tolerances: + 6v - 3v + 30 v 28 ± 4 percent ± 4 percent ± 4 percent (required on the indicator panelnot on the circuit card) N to f Converter Input Specifications Input voltage levels are a fundion of the input current which is set by the driving circuit, the 82 ohm input resistor, and the base-to-emitter drop of the transistor. Fan-in Capabilities: The N to F converter-driver is designed to accept a single N-Ievel input signal. N to f Converter Output Specifications DC Voltage Levels and Limits: + .912v - .380 v f . 3.92V 2.98 v Fan-Out Capabilities: The N to F converter without the inverter output can drive only one logic block. When the converter has the inverter (out-of-phase) output, the capabilities are the same as for the AOI DIF logic block. circuit also converts the incoming signal to an F-Ievel (voltage mode) signal sufficient to drive ten DIF logic blocks. P to F Converter Input Specifications - DC Voltage Levels and Limits: N to f Converter Delays The best-case to worst-case circuit delays for the N to F converter vary from approximately 4 to 160 nanoseconds. Worst-case delays consider worst-case conditions of components, voltages, driver and output loading. Power Supply Requirements Power Supply Limitations: 1. All power supply tolerances are ±4 percent at the circuit. 2. Voltage sequencing not required. 3. Circuit card can be removed without damage to itself, its driver, or its load. Over-Voltage Limits: SUPPLY + 12 volts + 3 volts - 3 volts - 12 volts OVER-VOLTAGE LIMITS 20 volts + + 9 volts - 9 volts - 20 volts These limits assure that breakdown limits will not be exceeded; they do not assume proper delays and levels. - 3.51 v - 3.90v 2.90V f - 3.44 v Fan-in Capabilities: The P to F converter-driver is designed to accept a single P-Ievel input signal. P to f Converter Output Specifications DC Voltages Levels and Limits: + 0.91 v + 0.18 v f 3. 12V 2.88 v Fan-Out Capabilities: The converter block has a fanout capability of ten of the following in any combination. 1. Logic blocks 2. Line drivers 3. Indicator drivers 4. Inverter (only one )-equivalent to three loads P to f Converter Delays The best-case to worst-case circuit delays for the P to F converter vary from approximately 4 to 54 nanoseconds. P-Line to DIF Converter-Terminator This converter-terminator circuit is designed to terminate up to 200 feet of 93 ohm coaxial line driven by a P-line (current mode) logic block (Figure 30). The ~A --P-U +12v P to F Power Supply Requirements Power supply requirements are the same as for the Nline to F converter-terminator. +3v R3 1.62K Dl Figure 30. P-Line to F-Line Converter Component Circuits 29 DIF to N-Line Converter-Driver F to N Converter Delays The F to N-line converter is designed to accept an Flevel (voltage mode) input signal and drive a single N-Ievel (current mode) output (Figure 31). This output can feed up to 200 feet (maximum) of 95 ohm coaxial cable piN 595997. The best-case to worst-case circuit delays for the F to N converter are approximately 6 to 76 nanoseconds. F to N Power Supply Requirements Voltages and Tolerances: - 12 volts + 12 volts {j- ---E F N ~--... A-- Output ± 4 percent ± 4 percent A ±3 volt excursion is permissible on the +12 volt and -12 volt power supplies under fault condJ.tions. DIF to P-Line Converter-Driver The F to P-line converter is designed to accept an Flevel (voltage mode) input signal and drive a single P-Ievel (current mode) output (Figure 32). The output c.an feed up to 200 feet (maximum) of 95 ohm coaxial cable PiN 595997. R7 N -12v P R6 N -12v +12v p Input _--..1\1'1., .......---1 +12v Output N p R3 6.2K _ Input Figure 31. F-Line to N-Line Converter F to N Converter Input Specifications Dl DC Voltage Levels and Limits: 3.38 v 2.95 v + 0.93 v + 0.47 v / Fan-in Capabilities: The F to N converter-driver is designed to accept a single F-Ievel input signal. +12v Figure 32. F-Line to P-Line Converter F to P Converter Input Specifications DC Voltage Levels and Limits: F to N Converter Output Specifications DC Voltage Levels and Limits: The converter output voltage levels depend on whether the terminator is a translating (N to P-Ievel) block, or a non-translating ( N to N -level) block. TRANSLATING NON-TRANSLATING LINE TERMINATOR LINE TERMINATOR - 0.20 v - 0.35 v j +0.32V + 0.21 v ::t_2.64V - 3.24 v - 3.38 v - 3.82 v Fan-Out Capabilities: The F to N converter can drive only one logic block. Output of the converter is an outof-phase signal level. 30 + 0.93 v + 0.47 v f + 5.80V + 4.42 v Fan-in Capabilities: The F to P converter is designed to accept a single F-Ievel input signal. This converter circuit is to be driven by an unloaded DIF circuit. The maximum wire length between the DIF circuit output and converter input is not to exceed 24 inches. The DIF circuit driving this converter is to drive no other loads. F to P Converter Output Specifications DC Voltage Levels and Limits: The converter output voltage levels depend on whether the tenninator is a translating (P to N-level) block, or a non-translating (P to P-level) block. TRANSLATING NON-TRANSLATING LINE TERMINATOR LINE TERMINATOR - 5.98 v - 6.52 v j -5.39V - 6.07 v - 2.75v - 3.12v j -2.16V - 2.63v Fan-Out Capabilities: The F to P converter can drive only one logic block. Output of the converter is an outof-phase signal level. F to P Converter Delays These delays are measured from the input of the DIF driving circuit to the output of the conversion circuit. F to P Power Supply Requirements Voltages and Tolerances: - 3 volts + 12 volts ± 4 percent ± 4 percent Power Supply Limitations: 1. The + 12 volt supply cannot be more positive than +12.82 volts. 2. The -3 volt supply cannot be more positive than -1.5 volts. The best-case to worst-case circuit delays for the F to P converter are approximately 11 to 62 nanoseconds. Component Circuits 31 60 rv208v 7618 PCU 3~ 7109 7111 Arithmetic Instruction Processi ng Sequence Unit Unit (Central Proc~ssing Unit) Power (400'" 208v, 3~) to all frames I 7909 Data Channel 716 Printer 711 Card Reader 721 Card Punch Figure 33. 7094 II System Configuration 32 1009DTU 1011 PTU 1014 RIU Telegraph I/O Units System and Functional Components System Components IBM 7111 and 7109 Central Processing Units Figure 33 is a block diagram of the 7094 II system configuration showing a possible combination of units. Of course, not all units are required in every installation. The final selection would be determined by customer need. Figure 34 illustrates the basic functional organization of the 7094 II. Although there is a slight change in terminology, components and functions are essentially the same as previously described for a general computer (Figure 1). Note, however, that a multiplexor has been added. Arrows indicate the general flow of information. Although the sections can be neatly separated physically, there are many functional combinations not shown in this grouping. Storage is the only functional section that is a separate machine unit. The central processing unit (cpu) of the 7094 II is actually made up of two sub-units-cpu-l and CPU-2. The CPU-l is the IBM 7111 Instruction ProceSSing Unit and CPU-2 is the IBM 7109 Arithmetic Sequence Unit. As these names imply, CPU-l contains all arithmetic and control registers and accepts, decodes, and routes instructions to the rest of the computer. Because of the dense circuit packing, CPU-l also performs a great deal of instruction execution. CPU-2 controls instruction execution for most POD 76 and sense indicator instructions, and I/O operations. :Many of the functions previous associated with CPU-2 are now contained in CPU-I. There is a great deal of interplay and overlap between these two units, and in some sequences it is difficult to determine an accurate functional boundary. The arithmetic section is the calculating section of the computer system. Here, portions of information, either instructions or data, can be transformed, combined, or altered. This section also keeps account of the instruction it is using and the one it will use next. The control section directs the other sections. It tells them what to do and when to do it. Instructions come into the control section from storage. CPU Control and Arithmetic f----Read Storage-- 0 " A ~f-,. A To SR Zero Check (4A) "---- Gate SI +SR SI 17 Gate IBR~SR IBR 17 Gate Op Keys Op Key 17 Gate AD~SR AD 17 Gate AC+ SR AC 17 Gate MQ ~SR MQ 17 ,. ,. (02. 12.47.1) -r- A - 0 A '- --" A (3A)----4 Gate Comp SR ~AD Not SR 17 Gate SR ~AD - ;---~ ~ A 0 ' t--- ;---- ,. A -'" 0 ~ t--A (4B) ...;. To AD 17 ,. '--'-- 02.02.10. I A t--A (2A) SR (17) ~ Shift Cell - Set SR Hold SR (IA) Gate SR Left .....AD -'- .J To AD 16 (tG) " 02.02.09.1 Gate SR 3-17 ..... XAD 02.01.05. I " A ..:. (4G) 0:3".05.47. I SR 17 To XAD 17 ,...--...., Gate SR ~SB r:J (~) MF SB 17 Output ",. 02.01.50. I L{HJ 51 41 SR 17 Late (ToMQ&SI) 02.01.11.1 Figure 38. Storage Register Position 17 Figure 38 shows condensed logic of SR( 17). Inputs to the storage register as a whole can come from: the op keys; instruction backup register (IBR); sense indicator register (SI); accumulator register (AC); MQ register; main adders (AD); index adders (XAD); and either the even or odd storage bus (even or odd memory) . Note that the index adders can be gated into either SR ( 3,.17) or SR ( 21 ~35 ). This choice in gating is helpful when routing the index registers to either decrement or address positions. SR( Q) also receives AC( Q) as temporary storage during certain instructions as mentioned previously. Note that the storage register inputs are also gated to a zero check circuit (Systems 02.12.47.1). During many operations, information is gated to the storage register to make use of this circuit in checking for zero values. Because there is no set pulse accompanying the input data, the present storage register contents are not destroyed or affected. The storage register is the focal pOint for information distribution and register swapping. Therefore, many outputs are gated from various sections of the storage register to various sections of other registers or adders. Units receiving information from the storage register include: the storage bus (the storage register is the only exit from the CPU to memory), main adders, index 40 adders, SI register, MQ, accumulator, and tag register (manual operations). Note that either the true or complement SR values can be sent to the main adders. The true output of the SR can also be shifted left one place as it is sent to the main adders. This shift feature is used during multiplication to effectively multiply by two (2). There is no full-word routing path from the storage register to the accumulator; data transfer of this type is accomplished by routing the storage register to the main adders and from there to the accumulator. SR( s, 1-8) and SR( s, 1-5) have a direct path to AC( S, 1-8) and AC(P, 1-5) respectively. These paths are used during floating-point operations for characteristic routing and convert instructions. SR ( Q) can also be gated directly to AC( Q) when restoring the AC to its original value after certain instructions. Either SR( 3-17) or SR( 21-35) can be gated to the index adders. SR( 3-17) gating is used primarily during class A (TIX, TNX, etc.) and index transmission (LXD, LDC, PDX, and PDC) instructions. SR( 21-35) gating is used primarily as: a routing path for index transmission instructions (LXA, LAC, etc.); POD 76 routing to the shift counter for class and unit decoding; and normal address modification with a specified index register. Each position of the storage register is combined on the same STAN-PAC circuit card with a corresponding position of the AD, AC, and MQ. Accumulator Register-AC (Systems 02.03.00.1-02.03.09.1) . The accumulator is a 39-position register, each position consisting of a shift cell (sc). The register positions are labeled S, Q, P, 1-8, 9P, and 9-35. Positions Sand 1 through 35 accommodate the computer word in standard operations. Positions Q and P are used as overflow positions because the sum of two 35-position numbers can be greater than 35 positions. Position 9P is also an overflow position used during floating-point operations; it replaces the 9 overflow trigger which was used on previous systems. The term accumulator is somewhat misleading because the register is not actually able to accumulate. The Accan contain only one word at a time. When adding, the AC and adders work as a unit to perform the addition, and the AC merely holds the result. Figure 39 shows condensed logic of AC( 17). Inputs to the accumulator can come from: the storage register, main adders, adjacent positions of the accumulator, and the MQ. SR( S, 1-8), SR( S, 1-5), and SR( Q) have a direct path to the accumulator. When the entire storage reg- Gate AD .... AC AD 17 Gate AD Rt-+AC AD16 Gate AD Rt 2 ~AC AD15 ister contents is to be set into the accumulator, however, the main adders are used as a routing path. Note that the main adders can be gated into the accumulator in a variety of ways: direct (true or complement form for positions Q, 1-8), shifted left one position, shifted right one position, or shifted right two positions. Advantage is taken of this shift feature during execution of arithmetic operations. Right and left shift operations cause a particular AC position to receive data from adjacent right or left AC positions; This shifting process proceeds at two positions per clock pulse as long as the shift counter value is two or greater. With a shift count value of one, a single position shift is accomplished. Because of these shifts, routing circuits are available to either the first ( adjacent) or second left/right position. of the AC and MQ registers. Right shifts take advantage of the "shift right 1" and "shift right 2" circuitry from the main adders; for example, the shift is accomplished by routing the AC to the main adders, and from there back into the appropriate AC/MQ positions. Routing paths are available to AC(9, 10) from MQ(34, 35) and used during the initial phase of DFAD instructions when aligning characteristics. Outputs from the accumulator can go to a variety of places: the storage register; main adders (true or com- -A ,...0 ~ - ... , -A A ~C) -'--- Gate AD Left ~ AC ADIS Gate AC Lett ADIS Gate AC Left 2 AC19 ~ ,. ...--..--A 0 r-- ~ A ....,. I-... A (4D) Gate Comp AC ~AD Not AC (17) Gate AC~AD - AC (17) ... "'---""-- Set AC Hold AC ... ... -,...-"" A 0 ... --"" ~ Shift Cell A (4A) To AD 17 -'--- 02.02.10.1 (2C) Gate AC Left 2 02.03.04.1 ... LJA (4H) To AC 15 ... 02-:03.03.1 Gate AC Left .------. A (4B) ~ To AC 16 ,. 02.03.04.1 Gate AC .... SR LJ :I-----(:A) To SR 17 ,. 02.01.05.1 AC17 Gate AC-+ IBR (tG) To IBR 17 ...,. 03.0S.05.1 Figure 39. Accumulator Position 17 7094 II System and Functional Components 41 plement form); index adders (AC positions 30-35 for convert instructions); IBR; and adjacent positions of the AC or MQ during shift operations. The accumulator is sent to the input of the storage register in many cases (without a set pulse) to take advantage of the zero check circuitry. Each position of the accumulator is combined on the same STAN-PAC circuit card with a corresponding position of the SR, AD, and MQ. Multiplier-Quotient Register-MQ {Systems 02.04.00.1-02.04.06.1} The MQ is a 36-position register, each position consisting of a shift cell (sc). The MQ receives its name from functions performed. At the start of a multiply operation it contains the multiplier; after the multiply it contains the least significant half of the product. At the beginning of a divide operation, it contains the least significant half of the dividend; after the divide it contains the quotient. Due to the double latch makeup of the shift cell, the MQ has the ability of shifting left or right. Figure 40 shows condensed logic of MQ(17). Inputs to the MQ can come from the storage register, accumulator, or main adders. The SR outputs may come to the MQ as a full word, or SR( s, 1-5) may be gated to MQ ( 30-35) for convert instruction operations. Inputs from the adders during multiplication are from AD(34-35) to either MQ(1-2) or MQ(9-1O), depending on fixed-point or floating-point operations. The Gt a e MQLft e MQ 18 Gate MQ Rt MQ 16 Gate SR~MQ SR 17 routing paths are also used as inputs to when performing right shift operations. Outputs from the MQ go to a variety of places. Gatings to the storage register are used for MQ store operations (the SR is the only output register to the storage bus) or for register swapping during arithmetic operations. MQ(S, 1-5) are gated to the index adders to produce required core storage addresses during convert operations. MQ(34, 35) are gated into AC(9, 10) during DFAD operations when aligning the fractions prior to the actual addition. Gating is provided for normal right and left shifts within the MQ or between the MQ and AC. Outputs are also available from MQ (s, 1) to MQ (34, 35) for ring shift (RQL) operations. Each position of the MQ is combined on the same STAN-PAC circuit card with a corresponding position of the SR, AD, and AC. AD(34-35) MQ( 1,2) Sense Indicator Register-SI {Systems 02.06.00.1-02.06.11.1} The sense indicator register is a 36-position register labeled S, 1-35. This register serves a variety of functions. It can be used as a set of switches which are set and tested by the program to check the progress or direction of the program. It may also be used to store words or parts of words temporarily; in this way the register is useful in altering and testing words. During some logical or masking operations, SI(S) is referred to as SI( 0); in these cases the first position has lost its identity as a sign and has become just another bit in the overall group of bits being operated on. G at e MQ Left 2 r-- r---- :: A ::10... 0 Gate MQ Left ~ .... A (4D) ~ J~ ~i...-- ~ ... ~ -A M~ (4E) ---~ Set MQ Hold MQ ... -.J (~) I (tH) Gate MQ 17 42 Position 17 To MQ 19 1 MQ~SR ~I ... A (2A) 02.01.05.1 MQ To MQ 18 02.04.03.1 02.04.03.1 Figure 40. To MQ 16 02.04.03.1 Gate MQ Rt 2 Shift Cell I --l 0 ... (:s) 02:04.Ci3. 1 J -A - U Gate MQ Rt Gate MQ Rt 2 MQ 15 Gate MQ Lt 2 MQ 19 To MQ 15 02.04.03.1 A To SR 17 Inv or Set 51 SR 17 Not 5117 Load 51 (To 5R) 5117 Elect Reset 51 Inv or Reset 51 Figure 41. Sense Indicator Position 17 During double-precision floating-point operations the SI register is used extensively for temporary storage of intermediate results. In these cases the S position retains its identity as an arithmetic sign. Each sense indicator position is composed of a single trigger-type position with a delayed output (Figure 41). The delayed output holds the trigger information long enough to allow proper input sampling during operations requiring SI register inversion. The only path for information into or out of the SI register is by way of the storage register. Instruction Backup Register-IBR (Systems 03.08.00.1-03.08.11.1) The !BR is a 36-position trigger register labeled S, 1-35 and is used primarily during instruction overlap operations. As long as the instruction sequence permits, the !BR contains the next sequential instruction to be executed. Because of the simultaneous readout feature from both the even and odd core storages, the !BR is able to obtain a next sequential instruction during the same cycle that the current instruction is being placed into the program register. Simultaneous memory readout also allows the !BR to obtain the next sequential instruction simultaneously with a data fetch (E) cycle of the current instruction. The first case occurs during "double-instruction" overlap and the second applies to "extended-sequence" overlap. The IBR is also used as a working register during double-precision floating-point and ERA instructions. Figure 42 shows condensed logic of IBR position 17. Inputs to the IBR can come from: either the even or odd storage bus when receiving a new instruction; the accumulator during double-precision register swapping or ERA operations; XAD(3-17) to !BR(21-35). In the latter case, the index adder is either returning a modified IBR instruction address or sending the incremented program counter value to the IBR for temporary storage. Input gating is such that the outputs from the IBR immediately follow the inputs from the storage bus. This feature (the same as for the program register to be discussed later) provides the earliest possible decoding of instructions as they are read for core storage, and is necessary for controlling overlap functions. This input gating feature does not apply, however, for data arriving from the accumulator. Outputs from the IBR can go to a variety of places: IBR(S, 1-35) are gated to the SR during both overlap and nonoverlap operations; !BR( S, 1-2) or IBR( S, 3-11) are gated to the program register at a time when the current instruction has completed operation and the overlapping (IBR) instruction is to begin execution; IBR(18-20) is sent to the tag register for indexing indications; IBR(3-17) are gated to the index adders for index register testing or modification during overlapping class A instructions such as TIX, TX1, and TNX; IBR(21-35) are also gated to the index adders so that address modification can be performed by the specified index register and the new address returned to the !BR; IBR(21-34) are gated to the MAR switch for an appropriate instruction or data fetch memory reference. Extensive decoding circuitry at the IBR analyzes the overlapping instruction to determine the various aspects of overlap possibilities and their operation. Gate 5B Even .... IBR A MF 5B 17 Even Gate 5B (4H) Odd~IBR A MF 5B 17 Odd (4G) Gate IBR 3-17-XAD To XAD 17 03.05.47.1 Not Hold IBR AC17 IBR 17 Reset IBR ----~R(4H) A (3A) To 5R 17 02.01.05.1 03.0B.05.1 Figure 42. IBR Position 17 7094 II System and Functional Components 43 Tag Register-TR (Systems 03.05.22.1) The tag register is a three-position trigger register. labeled 18, 19, 20. This register is used to inform the system which index register (XR) or registers are concerned with a particular operation. The outputs of the tag register may be used singly or in combination to specify the index register/registers required. When operating in seven-XR mode, the tag register uses all combinations of these tag bits to specify the particular index register. Therefore, the presence of more than one tag bit does not mean the oR'ing together of index register contents. The 7094 II does provide compatibility with previous systems having only three index registers. The normal power-on status of the 7094 II is the three-xR mode. By using the instructions LMTM and EMTM, the programmer has the ability to place the computer in either seven-XR or three-xR mode, respectively. For a further explanation, refer to the section on "7090/7094/7094 II Compatibility" in Volume 3. MF SB 18 Even SB Even to TR I Time or IA ------~~~J_--~--~~ SB Odd to TR 0 A A (2A) (To Tog Decoding) TR 18 Convert A3(Dl) IA Tgr A (4H) Figure 43. Tag Register Position 18 44 Figure 43 shows condensed logic for TR(18). Inputs to the tag register come from: either the even or odd storage bus during normal instruction loading from memory; the IBR during instruction overlap; or the storage register during manual operations. Outputs go to the tag register decoder circuits which are conditioned by either three-xR or seven-XR mode of operation. Address Register-AR (Systems 03.06.00.1-03.06.01.1) The address register is a 15-position trigger register labeled 3-17. The primary purpose of the register is to provide transfer or data reference addresses to core storage. These references may be: a direct address; an effective address as modified by an index register; an indirect address; or a skip 1/2 address as required by skip-type instructions. The only gated input to the address register (Figure 44) is from the index adders which are used as either an incrementing, modifying, or direct routing path from other counters or registers. Positions AR(S, 16 and 17) can also be forced on as a direct result of control logic resulting from certain trapping operations. The address register is not a counter; any address register incrementing is accomplished through use of the index adders. Outputs from the address register go to the index adders when forming skip 1/2 addresses, or the MAR switch when making references to core storage. The output of AR(17) is sent to the MAR bus selection circuitry (Systems 03.06.28.9). This circuitry, depending on whether AR(17) is a 0/1, controls address register gating to either the even or odd memory during normal mode of operation. When in diagnostic mode, however, AR(17) becomes an integral part of the address; this latter condition, when applicable, forces a bit to be gated to MAR(S) of either the even or odd memory. AR(S or 17) also effects the "AR odd trigger" which in turn controls gating circuitry for either the even or odd storage bus into the CPU. Indicator lights for the address register are located in the CE test panel area of the operator's console. Diagnostic Mode AR 3 Not IBR to AR Odd T r XAD~AR CP Set A Not Diagnostic Mode Gate XAD Latch ~AR 03.08.13.1 R(4E) 02.12.70.2 XAD Latch 17 CP Set D PR Store MAR Bus Selection (03.06.28.9) E Time Early 03.06.02.1 03.06.01.1 Reset AR on Trap Clear or Rst orlntlk Rst Reset AR 03.06.02.1 Gate AR+XAD Figure 44. Address Register Position 17 7094 II System and Functional Components 45 Program Counter-PC (Systems 03.06.30.1-03.06.31.1) The program counter (labeled "Instruction Counter" on the operator's console) is a 15-position trigger register labeled 3-17. The primary purpose of the program counter is to keep account of the next sequential instruction in a running program. The program counter is not a counter; any incrementing or decrementing is accomplished through use of the index adders. During normal sequential instruction execution, the new address associated with the program counter is generated at 6-time of the cycle when a reference is made to core storage. This new address is placed temporarily in either the address register Or!BR until I-time and then returned to the program counter via the index adders. Note that when updating the program counter at I-time, the index adders are used as a routing path only, and do not affect the value being transmitted. The only gated input to the program counter (Figure 45) is from the index adders. As previously explained for the address register, the index adders act as either an incrementing, modifying, or direct routing path from the other counters or registers. PC(S, 16, and 17) can also be forced on directly by control logic resulting . from certain trapping operations. Note that either true or complement outputs can be sent to the index adders. The true output is used dur- ing incrementing, decrementing, or routing of the program counter to other registers. The complement PC output is used only during overlap store operations to check if the store instruction (in the program register) is storing in the next sequential core storage location. If this is the case, overlap is interrupted because the overlapping instruction (now in the !BR) is being modified by the program and therefore invalid as it stands in the !BR. The program counter is also sent to the MAR switch when making references to core storage. The output of PC(17) is sent to the MAR bus selection circuitry (Systems 03.06.29.2) for the same purpose as previously explained for the address register. This circuitry, depending on whether pc(17) is a 0/1, controls program counter gating to either the even or odd memory during normal mode of operation. In diagnostic mode, however, PC(17) becomes an integral part of the address, and when applicable (for example, when PC(17) equals 1) forces a bit to be gated to MAR(S) of either the even or odd memory. PC(17), indicating an even or odd status, also controls gating circuitry from the even or odd storage busses into the program register, storage register and !BR. It must be remembered that the program counter is stepped +1 prior to the time that this gating circuitry is used. Because of this prior stepping, an odd PC indication, for example, gates the even storage bus to the SR and PR, and the odd storage bus to the !BR. Storage bus gating is explained later in the "Master I-Time" section. XAD-+-PC CP Set A Gate Comp PC ~XAD XAD Latch 17 Set PC To XAD 17 Not PC 17 Gate PC .....XAD To XAD 17 Reset PC on Tra Clear or Rst or Int 03.06.32.1 PC 17 MAR Bus Selection (03.06.29.2) Diagnostic Mode PC 3 A a _N__o_t_Di......:ag::....n_os......:ti_c_M~od_e_ _---+~ A PC Not Odd PC Odd (2A) 1-----. 03.06.30.1 FIgure 45. Program Counter Position 17 46 Index Registers-XR (Systems 03.05.00.1-03.05.14.1) Program Register-PR (Systems 03.14.01.1-03.14.06.1) The 7094 II has seven index registers. Each is a 15-position trigger register labeled 3-17. All seven XR'S are identical and used primarily for address modification. The 7094 II can be in either three-xR or seven-XR mode of operation. In three-xR mode (multiple tag mode), multiple tag bits cause an oR'ing of index registers. Input to the index registers (Figure 46) is from the index adders which are used, in this case, as either a modifying or direct routing path from other registers wi thin the CPU. Individual positions of all seven XR'S have a common output which is gated (under control of tag register decoding) to the index adders in either true or complement form. Because of the true XR outputs, it is possible to accomplish true addition in the index adders as is required during TXI operations. True outputs are also used to advantage during PXA, PXD, SXA, and SXD operations when true values are moved to the accumulator or core storage. Tag register decoding gates the complement outputs of the specified index registers to the index adders for a variety of operations, the primary one being address modification. By addition of the complement XR valuc to an address, the address is effectively reduced by the contents of the XR. There are many instructions which operate on the index registers, thereby making these registers also useful programming tools for counting, word alteration, and program loop control. The program register is a ten-position trigger register. The purpose of this register (labeled "Instruction Register" at the operator's console) is to receive and decode the operation portion of the instruction to be executed. Decoder outputs then initiate and control the computer operation until the instruction is completed. The operation code depends on the type of instruction involved and consists of either positions S, 1-2, or S, 3-11 of the instruction word. Positions S, 1-2 are routed to PR(S, 8-9); positions S, 3-11 are routed to PR(S, 1-9). Primary inputs to the program register (Figure 47) come from: either the even or odd storage bus; or the IBR depending on whether or not instruction overlap is involved. Outputs from the PR follow (rise with) the inputs from the storage bus. This feature of the input gating provides the earliest possible decoding of instructions as they are read from core storage. (During memory read-out, the S-bit is strobed before the 35-bit.) Inputs from the IBR or op keys are gated in with a clock pulse and, therefore, differ from SB input gating. Certain trapping operations force specific PR triggers on directly-for example, PR( sand 9 )-without having to rely on storage bus or IBR inputs. During manual operations, the program register can be entered from the console op keys to allow execution of specific instructions set up by the operator or customer engineer. XAD Latch 17 4A . . .:1.:.:.BRc..::2_ _ _...........,~ ° A _1_BR_l_l_ _ _~ ° (4B _O-,-p_K-,eY,-l_l_ _-..J A Gate Op Keys Not IBR 1 Gate Camp XR-XAD Not XR 17 To XAD 17 ,...---!~~.:.L..J"----I~ 03.05.47.1 --'-....:.......t~ (lC) A 0 MF SB 2 Odd A 0 Force Store and Trap XR (17) A (lE) 03.05.14.1 Gate XR-XAD Position 17 PR (9) Not MF SB 1 or 2 Odd MF SB 11 Odd Gate SB Odd-PR A (lD) XRA Not Hold PR MF SB 2 Even A XRF(17) Tag 6 XRG(17) Ta 7 Gate IBR-PR Not MF 5B 1 or 2 Even MF SB 11 Even Gate SB Even-PR A Tag 2 Figure 46. ° ° e 1 --,N.....:o:..:...t-=..0c....K C'-'.L..-_----.! Op Key 2 XRA Set PR 9 (To PR Decoding) .;..:;Re=se,,-,-t-,-,PR,--_~ R(3C) 03.14.06.1 Gate PR to PR in 01 D A ° .:::.G.:::.:.ate:::....:.:::.:IBR~t:=...o-,-,-PR'-'.in:..:...O,,-,-l!::...D_ _~ A 2E) PR 9 to Gate D Figure 47. Program Register Position 9 7094 II System and Functional Components 47 Outputs of the program register feed into the operation decoder circuitry (Figure 48). PR(S, 1-5) provides the primary operation part and PR( 6-9) provides the secondary operation part of the operation decoding. During some instructions, the shift counter becomes an extension of the program register and provides a class and unit address decoding. Note that the 7094 II actually has two program registers: the "master" register, the one which receives inputs directly from the storage bus or IBR, is located in A-gate of CPU-I (7111) and the second (slave) program register is the register that existed in the 7094 system. The "slave" register (Systems 03.04.00.1-03.04.06.1) is located in D-gate of CPU-I and is not really a register as such; converter circuit blocks have replaced the triggers so that these outputs follow the outputs of the master register as gated to the D-gate (Figure 47). Basically, the new program register provides new and faster F-Ievel circuitry necessary in the 7094 II. The old register provides PiN level operation decoding and gating in CPU-2 for instructions that do not have such critical timing requirements. The functions of the two program registers and operation decoders cannot be cleanly separated; therefore, both must be considered when analyzing certain instruction operations. Indicators at the operator's console are powered from the old program register. Shift Counter-SC (Systems 03.14.14.1) The shift counter is an eight-position count-down counter labeled 10-17. Each position consists of a shift cell (double latch) which allows simultaneous read-in and read-out when counting. The shift counter is used to count the number of shifts or indicate the progress of operations such as MPY, DIV, convert, floating-point add, shifting instructions, etc. The shift counter also functions as a class and unit address decoder for operations that have a primary operation of 76 (Figure 48). The primary entrance to the shift counter is by way of the index adders. The time at which the number is gated is controlled by the operation being performed. During POD 76 operations, the sc is set either during 15 time from PR decoding or at the beginning of L time when the IBR is transferred to the PR during overlap. During variable length multiply' or divide operations, the count field is routed from the storage register via the index adders in E time; during convert instructions, the count is routed in the same manner in L time. The shift counter is also used during single and double-precision floating-point add and subtract operations (POD 30). During these operations the shift counter is used in lining up characteristics. The characteristic difference between the numbers in the SR and AC is 48 Primary Operation Decoding (±OO ... ± 76) Class Addr Decoding Secondary Operation Decoding (00 ~ 17) \ (00 .... 36) Unit Addr Decoding (00 -+ 17) v Operation Decoding Figure 48. Operation Decoding computed in the main adders, AD(I-8), and then gated directly to SC(1O-I7) to control the number of shifts necessary to align the fractions. During multiply and divide operations (other than variable length), a constant is forced into the shift counter to control the number of iterations. In the case of a floating-point operation, 33 8 is set into the shift counter; for a fixed-pOint operation, 43 8 is set into the shift counter. Note that the 7094 II actually has two shift counters: the "master" register, the one which receives the inputs just described, is located in A-gate of CPU-I and the second "slave" shift counter is the counter that existed in the 7094 system. This "slave" register (Systems 03.04.14.1-03.04.17.1), located in D-gate of CPU-I, is not really a register as such and has no counting capabilities; converter circuit blocks have replaced the triggers so that these outputs follow the outputs of the master counter as gated to the D-gate. As for the program register described previously, the new shift counter provides fast F-Ievel control and counter circuitry necessary in the 7094 II. The old register, again, provides piN-level controls for instruction operation in CPU-2; the main function is in class and unit decoding and POD 76 controls. The functions of the two counters overlap somewhat; therefore, both must be considered when analyzing certain instruction operations. Indicators at the operator's console are powered from the old shift counter. Stepping the Shift Counter Figure 49 shows condensed logic for positions 10-17 of the shift counter. Remember that the shift counter is a self-contained true binary count-down counter and does not rely on the index adders for decrementing as is necessary with the address register and program counter. ~ ,. (2E) - SC 10 Insert SC 10 ~-r- o r-+--' __ HoidSC10 A 18 Shift Cell ~ No' Lc 10 SC 10 -"" 03.14.14.1 0 ~; -0 (31) '-- ~ .(2D) TO Tl T2 T3 T4 T5 T6 T7 T8 T9 TlO T]] T12 T13 T14 SC (13) SC (14) SC (15) SCj16L SC (17) N F N F N F N F N F N N N N N N N N N N N N Value (10) (11) (12) 33 F F F 23 21 17 15 F F F F F F F F F F F F N N F F F F N N F F N N 13 F F F F N F 11 7 5 F F F F F F F F F F F F N F F F N N 31 27 25 F F F F F F F F F N N NN N N F F F F N N SC sc sc SC 3 F F F F 1 F F F F F F F F 0 F F F F F F F N F F F N N N 1 ~ (10) -- Set SC 10 SC 11 Hold SC 11 0 ~ r-- 0 A 19 ~ r-0 (3H) ~ ,---- - ..... ..... ~ A5 -~ - - HoldSC 12 ~cell (4A) ~ ~ A6 ~~ Y. --'--'(5B) f - I - Set SC 12 SC 13 ~ ~ "- r-=!: == - Hold SC 14 0 A 22 (41) -0 - r-=! ~ ~ , r-- 0 16 t-- ~r-r- ~~(4H) =: Hold SC 15 SC Reset Reset SC SC 15 ~Cell ~- SC 15 Set SC 15 Insert SC 16 .--- HoldSC16 0 A 24 0 , ~ !. 0 (2G ~ I J Shift Cell -- J"':! Not 1 SC 16 SC 16 03.14.14.1 Set SC 16 ~- Step By 1 '--- 33 to SC FAD AD~SC XAD to SC Set Hold Mpy Div 43 to SC 1 03.14.14.1 ~ r-- 2 A (3E) Not -"-- -.--------. ~Stepsc ..... r-- ~ ~OA r-- 23 0 (3D) r--- SC 14 Shift ',--- 03.14.18.1 'Not lsc 14 ... Set SC 14 ~ (lC) Shift Ceil 03.14.14.1 Insert SC 15 (5E) (4D) ... SC 13 SC 14 ,.-- Insert SC 14 0 ~ SC 13 0 (4F) Set SC 13 A (2C) ~ Cell 1 ~03.14.14.1 03.04.15.1 ~ Not Shift 0 A r-- 21 0 &.r-- ~ Insert SC 13 HoldSC13 1 t-- -. A O SC 12 03.14.14.1 1 14 Step By 2 SC 12 ~~ 33 to SC ~ 1 I~- 0 A 20 r:~3G Mpy Div 43 to SC 7 Not Shift .--r-- SC 11 SC 11 SC 12 c=:'0 1 Set SC 11 Insert SC 12 (3B) -- Ir! Not 03.14.14.1 A (4C) .--- Shift Cell '--- T N = On F = Off - Insert SC 11 ~ -(5A) ~017 (4A) ~ 1 SC 17 Hold SC 17 r-- 0 ~~I"""l 0 (2C/2E ~ ro 0 A 25 ... RI to SC Line 1 and Line 2 CP Set C ~ Insert SC 17 o (4F) ~ Shift Ceil Not 1 sc 17 SC 17 03.14.14.1 Set SC 17 Oti4:16.1 Figure 49. Shift Counter Stepping 7094 II System and Functional Components 49 Figure 50 shows a sequence chart of the shift counter stepping to zero from a count of seven. Note that stepping the shift counter is conditioned by either «step by I" or «step by 2" control circuitry. Whether the stepping is by one or by two depends on the particular instruction being executed. Stepping by two is allowed during multiplication, floating-point add/subtract, and shift-type instructions; stepping by one must be performed during divide and convert-type instructions. Each position of the shift counter has control circuitry affecting three inputs to the shift cell ("insert sc 17," «set SC 17," and «hold sc 17"). The «insert" input is activated only when that particular position is to be fumed on. The "set" and "hold" pulses are both CP set pulse inversions of one another, and generated from the same control circuitry (OA blocks) only when the status of that particular position is to be affectedturned off or on. Whether a shift counter position is to be changed or not by a set/hold pulse combination is determined by the next lower position. If the next lower position shift cell is to be set on, the state of the next higher position will be reversed. This rule applies for both single and double stepping. When stepping large numbers, the ripple effect from position to position can become too great for reliable operation. To reduce this ripple time, look-ahead circuitryhas been inserted at the controlling circuitry of SC(IS). Note that SC(16) alternately changes state between and OFF during double stepping and SC(17) alternates state during single stepping. The primary difference between single and double step operations is where the step pulse control comes into the counter; for example, SC(17) for single step and SC(16) for double step. Double stepping is, therefore, the same as single ON 50 stepping except that it occurs one rung higher on the ladder. Assume that a count of 338 has been set into the shift counter and that counting is by 2. The initial status of the counter with 33 8 is: positions SC(17), SC(16), SC(14), and SC(IS) are in the ON position; all remaining positions are OFF. Because the original number is odd, each stepping of the counter produces an odd result. If the original number was even (68 , for example) each double step would, in turn, produce an even result. Note that all during double stepping, SC(17) remains unaffected. The first step reduces the counter from 338 to 31 8 • This reduction is accomplished by reversing the state of SC(16) from ON to OFF. SC(16) being initially on, blocks A7 from activating "insert sc 16." Set and hold pulses, however, are produced at CP set time by OA24 with the overall effect of inserting a zero into SC(16). The second step which reduces the counter from 31 8 , to 278 initially finds positions SC(17), SC(14), and SC(IS) on; and all remaining positions off. SC( 16) being off conditions A7 and 0 16 to change the state of SC(16) to ON: «Insert sc 16" conditions OA 23 to produce set and hold pulses for SC(15), and also test the initial status of SC(15) at A15 • SC(15) being off, activates "insert sc IS" to flip SC(15) on. "Insert sc IS" also conditions SC(14) circuitry to produce only the set and hold pulses thereby turning SC(14) off. At the end of the clock pulse, positions SC(17), SC(16), SC(15), and SC(lS) are on and all remaining positions are off. This pattern of double-stepping continues until all shift counter positions except SC(17) are turned off. As the count is reduced from three to one, «step by I" control circuitry replaces "step by 2." "Step by I" tests the status of SC(17) at A9, but SC(17) being on blocks «insert sc 17." A set and hold pulse is produced at OA 25 , however, with the result of turning SC(17) off. At this point the shift counter equals zero and all stepping controls are dropped. System Page Test Point line Name Level 02. II. 79.1 02A4E13D Shift Gate -N 03.14.20.1 0lA4G26C SC 20r More -F 03.14.20.1 01A4G23F SC Eq 1 -F 03.14.18.1 0lB2B28B SC Zero Slow -F 03.14.18.1 01A4D24D Step by 2 -F 03.14.18.1 01A4G25B Step by I -F 03.14.18.1 01A4A25B Step SC -F 03.04.16.1 01A4F28D CP Set C -F 03.14.14.1 01A4E24B SC 15 -F 03.14.16.1 0lA4F27G Insert SC 15 -F 03.14.16.1 01A4F27F -F Set SC 15 03.14.14.1 01A4F20C SC 16 -F 03.14.16.1 0lA4E27A Insert SC 16 -F 03.04.16.1 01A4F27C Set SC 16 -F 03.14.14.1 01A4E26B Set SC 17 -F 03.14.16. ] 0lA4F25B Insert SC ]7 -F 03.04.16.1 01A4F27E Set SC 17 -F (7) I (5) 2 (3) 3 (I) 4 (0) 5 6 7 :'I I-- . h ~ -. 1-h ~ ~~ ~ I. I--I-- h --- r- h.~ ~ '- Figure 50. Shift Counter Timing Chart 7094 II System and Functional Components 51 Main Adders-AD (Systems 02.02.00.1-02.02.19.1) The adders in the 7094 II are made up of basic building blocks shown in Figure 51. The adders perform arithmetic functions and are involved in many internal data transfers within the computer. Inputs to the basic adder circuit may be from either the accumulator or storage register. Outputs from the adders go to the accumulator, MQ, storage register, and shift counter, depending on the operation involved. The adders are comprised of 39 individual adder units, or bit positions. These positions include AD (Q, P, 1-8, 9Q, 9P, 9-35); AD (Q and P) are used for overflows which might occur during certain arithmetic or logical operations. AD (9Q and 9P) are used during floating-point arithmetic. Individual Main Adder Position Figure 51 shows condensed logic for position 35 of the main adders. The primary objective of an adder position is to determine whether or not there is a sum output. To determine the output, three values must be analyzed; two adder inputs and one carry input. Input A Gate AC-AD AC 35 Gate Cam AC-AD Nat AC 35 Systems 02.02.19.1 A °1 A (4A) Propagate 35 (Either/Both) Input B Gate SR-AD SR 35 Gate Comp SR-AD Not SR 35 A °2 A Generate 35 (Both) (4B) Exclusive OR 35 r Comp Exclusive OR 35 (Both/Neither) Carry In AD Sum 35 Exclusive OR of Adder and Carry Adder Inputs Adder Outputs 0 1 0 1 0 1 0 1 SR 0 0 1 1 0 0 1 1 AC 0 0 0 0 1 1 1 1 Carry - + + + - + + + Propagate 35 - - - + - - - I Generate 35 - + + - - + + - Exclusive OR 35 - + + - + - - I' AD (Sum) 35 + Indicates an active logical state - Indicates an inactive logical stale Figure 51. Main Adder Position 35 ~ 52 True or complement inputs from the storage register and accumulator are gated into the adder at gate circuits AO l and A0 2 • Either/both adder inputs being active produces "propagate 35" at 0 3 ; both inputs being active produces "generate 35" at A4 • Block A05 performs an exclusive-oR test of the adder inputs. An input exclusive-oR condition occurs only when either one but not both of the inputs contains a 1. A0 5 performs the test by looking for a "both" (generate 35) or "neither" condition which is actually a complement exclusive-oR function. The in-phase (complement exclusive-oR) and out-ofphase (exclusive-oR) outputs of A0 5 are tested along with input carry conditions in a second exclusive-oR circuit configuration to determine the actual adder sum output. One adder output condition occurs at Aa with "no carry" and "EX OR 35" (case 1 below); the other adder output condition occurs at 0 7 and Os with "Comp EX OR" and "Carry" conditions, respectively (Case 2). CASE 1 CASE 2 (a) (b) 1 0 0 0 1 0 Input A Input B Input Carry Adder Output 1 1 Adder Output 1 (1)1 Input A Input B Input Carry (a) (b) 0 0 1 1 1 1 Taking the other four possible combinations of carry and adder input conditions will prove that an adder sum output is not produced. All combinations are summarized in the chart on Figure 51. Main Adder Bit Carry Lookahead Each adder position, besides being able to determine a sum condition, must be able to provide necessary carries to adjacent positions. For determining possible carries, three additional lines are generated from each adder position: propagate (P), generate (G), and exclusive-OR (EX oR)-Figure 51. The output of 0 3 , "propagate 35," is active when either one or both of the adder inputs are a 1. With these input conditions, an input carry must logically be passed on, propagated, to adder position 34. This carry requirement is proven below by taking the input combinations of 0-1,1-0, and 1-1 and adding a carry; in each case, a carry results. (Also, see chart on Figure 51.) (a) Input A Input B Input Carry (b) (c) 011 1 0 1 1 1 1 10 10 11 The output of A4 , "generate 35," is active whenever both inputs to the adder position contain a 1. With both inputs active, the adder position has the ability to generate a carry from its inputs alone independent of whether or not an input carry is furnished from an outside source or a lower adder position in the group. Proof of this carry is easily obtained below and summarized in the chart on Figure 51. Note in the chart that "generate 35" is not dependent on the input carry status. (a) (b) Input A l l Input B 1 1 Inpu t Carry 0 1 10 11 The output of A0 5, "EX OR 35," represents an adder input exclusive-oR condition; either one but not both inputs are a 1. With the propagate, generate, and EX OR conditions just explained, individual adder bit carry lookahead can be developed. This lookahead is shown by the solid lines and logic in Figure 52. Using adder 35 as an example, a "carry in 34" will be generated if: adder 35 can generate a carry; or adder 35 can propagate and there is a carry into position 35 (K in). Note that the lookahead circuitry becomes more involved as it progresses from the lowest-order to the highest-order adder position within the adder group. The "carry in 31" circuitry uses exclusive-oR inputs instead of propagate as used in the lower-order positions of the adder group. This change was made because of circuit loading; the basic circuit logic, however, remains the same. Note in Figure 52 that there is no individual adder lookahead circuitry for carry generation into adder 30. Adder 30, which is the low-order position of the next group, receives its carry indication from the group lookahead circuitry. An important point to remember in the carry lookahead concept is that each input carry generation is dependent solely on inputs to individual adder positions; there is no "ripple" effect from one adder position to another as is characteristic of slower type adders. Main Adder Group Carry f.ookahead If the lookahead were continued as explained in the previous section, the circuitry would soon become too large, unwieldly, and expensive. Therefore, to speed up adder operation and facilitate lookahead, the adder has been divided into seven groups as follows: GROUP NUMBER ADDER BIT POSITIONS 1 2 3 4 5 6 7 35,34,33,32,31 30,29,28,27,26 25,24,23,22,21,20 19,18,17,16,15,14 13,12,11, 10,9,9P 9Q,8,7,6,5 4, 3, 2, 1, P, Q Each group contains either four, five, or six adder positions. Adder position 9Q is shown in group six but actually plays no part in the lookahead function. The primary objective of the group lookahead is to determine whether or not carries should be introduced into the low-order adder positions of higher adder groups. Lookahead circuitry for group 1 is shown as dotted lines on Figure 52, lookahead circuitry for group 2 is shown in Figure 53. Overall lookahead for all seven adder groups is shown as solid lines on Figure 54. 7094 II System and Functional Components 53 1011';----- I'~----- Adder Positions 31-35 - - - - - _ .... Bit Corry (02.02.28.1) --------------l,~I~I(----Generote (02.02.36.1) ------;,~I Propagate 31 ----------, Adder 31 I I Generate 31 r-----------I------------- - 02.02.17.1 r AD Sum 31 ~-----u_= t------------ - -i - - - j - - -- - - - - - - - - - ""~ w, : X32 r---X33 A I '"G35 i 32 1 1 I 02.02.17.1 r 1 I G33 I I AD Sum 32 I I Adder 33 Generate 33 ~ AD Sum 33 I In 32 (4D) P33~ G33~ -I (3D) _ _ I I ~ I I A ~ Propogate 35 ~ , (3C) ~- G34~ ·1 (3B) I o Carry (2B) In 33 ~ r----------- t-t- I A ,----------- - I I (4F) ~ Kin (4B) G35 A ...... 0 Carry (3A) In 34 (4A) L1=t-+--r L-t-.- -P34 - - --P3:s ----------------- L-.l_ - - - - - - - - - LI_ "N,"" DO" ~ ~ ~,~ ~- G35 W' J7J i I I I : r--------- - - - 0 L ------- 1-------- - - ---- I I ~ P34 t-Ll- , - - - - - - - - - - - - I I Generate 35 ~ 'd",'," 0'" A I >- Kin I I I t- ' - - - - - - - - - - - - ~III_r / - - - - - - - - - - I I II i II I I I I AD Sum 34 mrr-;El P34 ~A (4C) I ~JJ. Excl usive OR 34 f------- f1, ---- AD Sum 35 02.02.19.1 Figure 52. Group 1 Adder Carry, Generate and Propagate 54 Carry G3<1 (:E) i II i1~ Adder 35 Carry In I II Generate 34 Gate SR . . AD SR 35 Gate Camp SR . .AD Not SR 35 I G35 0 (lD) ~r-iil r-- 02.02.18.1 J 1 A P34 t--!.l r - r - - - - - - - - - - I +---1-1-- - - - - - - - - - I I t-r-- 1 - - - - - - - - - - - ---------- Adder 34 Gate AC+AD AC 35 Gate Camp AC . .AD Not AC 35 1 I (4E) P33 I I I Exclusive OR 33 Propagate 34 r (:G) J P34 P35 r- Kin I I II I I r Carry ~ I\ ------ ~ ''') '"" G32ID (2F) _ II 02.02.18.1 ~ 0 h---- / - - - - - - - - I t---- - - - - - - - - H- -+,--- - - - - - - - - - I Exclusive OR 32 Propagate 33 (Carry In G2) X32~ I I Generate 32 (3G) X32 ' - - X33 :.J A G;L.(J (3F) .-1- L-:::l I Adder 32 1-------- Generate or Propagate G J (3D) X32 , - - A X33 X34 X35 I ~opagate o Group Lookahead Circuitry--- (4G) Lookahead for both groups 1 and 2 is shown because of their difference in logic. The dotted lines in Figure 52 show group 1 lookahead. Six sets of conditions can produce an output from the group: first, AD(31) can generate a bit alone; second, AD(32) can generate a bit and AD(31) can propagate the bit; etc. Note that the logic becomes more involved from top to bottom; the last AND circuit indicates that an input carry (Kin) will produce a group carry if AD(31-35) have the ability to propagate. Group 2, as well as the remaining groups 3-7, produce two outputs; «any generate G2" and «generate or propagate G2." Both conditions must be active to produce a carry into the next higher group. Design of thf' adder lookahead is the result of Boolean algebra, and may not, therefore, seem immediately logical and straightforward. This design, however, performs the required task of adder lookahead in a more efficient G30 " " " G29 ~ 01 G28 ... G27:' (3A) G26 "" (3B) Propagate 26 Generate 26 1 and economical method than other «more straightforward" configurations. The adder group output, "any generate G2" results at 0 1 (Figure 53), from any or all adder pOSitions being able to generate a bit (both adder inputs are a 1). The "generate or propagate G2" output from 0 7 results from the following conditions: first, all adder positions are able to propagate at A6; second, each position, except the low-order adder position, has the ability to generate and have this bit propagated through the remaining higher-order adder positions at A5 , A4 , A3 , and A2 • The ability of the low-order adder position to generate and have its bit propagated by the remaining adder positions is not defined specifically by the function «generate or propagate G2." This latter case is accounted for, however, because if AD(30) can generate, it can also propagate. P30 activates A6 and G30 activates 0 1 ; therefore, both of the group outputs are active and a carry is passed into group 3. Any Generate G2 ~ ID-~ G26... --=-=.:..:.::.:...::.:..=.-:=..::..-----t-t++,....----;l~ A2 (4B) _L-_~---' Propagate 27 Generate 27 P26 ... G27 ~ Propagate 28 Generate 28 P26 _____ P27 ",.I A4 G28 --J (4D) Propagate 29 P26 P27 P28 A3 (4C) ... ,. r-! ~,...-- """ ~ AS -=G_e_ne_ra_te_2_9_ _-+-+++-t-4--G_29---:J~~ (4E) - (3D) Generate or Propagate G2 ... ,. _ ' I - Propagate 30 Generate 30 P26 ... _ P27 P28 P29 P30 ... "" ... "" A6 (4F) i--- 02.02.35.1 Figure 53. Group 2 Generate and Propagate Lookahead 7094 II System and Functional Components 55 Adder Lookahead GJ GJ _, • G~~~:;..", Group Lookahead • I. ___________ :~02~~~ _________ , Iol G~e~t~r~~~e_~? ~[!] .,.02.4011 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ll-fOl ~r--------------------,0 gQ [2] ~--11r-~ rr: I I I 1IIIIr 1IIIII I1I1I1 III GJ I _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ .J 02.02.31. 1 GJ o ~ Carry Q i11111 III 0 Any Generate G6 ~ II I III, ~ L2.J " GJ 1I1I III -GJ --------0---- I: I 02.02 .32 .'1--- rn~mgj~[:l I'II Any Generate G5 'I I I Generateor -0 o A I 0 I :~~~a~,:J._~I_I_-+_I_I-I_I_~~~-=_==_==~~==-=-=-= ~#-~J -----------02:~33:i---- o 0 'II Gen- ~ o I erateor PropagateG4 I I : -~----------o~~7---- C32:J 0 Any Generate G3 I [EJ ~ I I J~ I ~ r----------02~.35.1-----~ 0 :,:~: or L---S---'Propa- I I (5~) (ts) I II : II I I II I II II : I t:t~_o Carry In GR 3 J==rFffiF==~~~==El g~~ - - - - - - - - - - - - - - - - - - - - - , ~t-il.-----L-.....J------+t------------------'fl-TI- (4F) I1II1I I --==~~~~===El 1-::Jt~~Ir~~~ I1II L LL - - ~-L_~Pr..:::!op:::::ag:!:a.:.::te....:::G:..:.1----+-------=C:..:.ar::..!.ry....:.l:..:...nG..:. R. . :._ 2 _---+-- __ =:_ Figure 54. Main Adder Group Lookahead and Q Q Carry I I 1/: / II III 1111/ /IIILy II L~- Generate or ~ I I: II i II JI - I I IIII I II 02.02.36.1 1/ II , Ii II 56 III I ~ G M I 1 I II Any Generate G4 ~0 e r;-:;r " I (4G) Carry Lookahead----- I I I / I __ .J 02.02.40.1 Figure 54 shows all adder group carries. Group 1 having the ability to either generate a bit internally or to propagate a Kin produces a "carry in GR 2." Note that this same carry line also conditions the top OR circuit in each of the higher group lookaheads. Depending on the inputs to the remaining adder positions, this carry from group 1 may be propagated through the rest of the main adder. A carry into group 3 can result from two conditions: group 2 can generate a carry internally, or group 2 can propagate a carry from group 1. Taking the first case, both "any generate G2" and "generate or propagate G2" are active and condition the top leg of 05B and 05.A, respectively. In the second case, group 2 produces only the "generate or propagate G2" line as input to 05.A. However, the carry from group 1 conditions the bottom leg of 05B; A4B is satisfied and a carry is introduced into group 3. Producing a carry into higher-order groups becomes progressively more involved but the basic concept just described remains the same. Note that each high-order group is conditioned by the outputs of each lowerorder group. In this manner simultaneous decisions can be made at each group lookahead; no "ripple" is involved from one group to another. Q Carry Lookahead The carry from the main adder, Q carry, is used in testing and logical operations as well as during arithmetic instructions. The speed at which a Q carry can be obtained becomes important in many operations. Therefore, adder group outputs are also used in Q carry lookahead circuitry (shown as dotted lines in Figure 54) to obtain the earliest possible Q carry indication. This lookahead follows the same philosophy described in the preceding sections. 7094 II System and Functional Components 57 Index Adders-XAD (Systems 03.05.40.1-03.05.47.1) The index adders (XAD ) in the 7094 IT perform a variety of functions such as: 1. Address modification on the instruction in either the storage register or IBR. In either case modification is from an index register as specified by the tag portion of the instruction concerned. 2. Tests on or modification of index registers as required by TIX, TNX, TXH, TXL, TXI (class A) instructions. 3. Incrementing the program counter, address register or address portion of the IBR. 4. Providing an address reference to MAR during overlap operations. 5. Producing the required data reference addresses during convert instructions. 6. Providing direct data flow paths to or from the program counter, address register, index register, storage register (positions 3-17 or 21-35), instruction backup register (positions 3-17 or 21-35) and shift counter. Two data input gates are provided to the index adders; input A and input B (Figure 55). These inputs are divided such that one half of the input data always arrives at input A and the other half arrives at input B. During address modification, for example, the address arrives at input A and the modifying index register value (complement) arrives at input B; during incrementing, the PC, AR, or IBR arrive at input A and a "hot 1" is inserted at input B of XAD( 17). It is possible to insert "hot 1's" into either all positions of the index adder or just XAD( 17); the first case causes decrementing of the input value (-1); the latter case causes incrementing of the input value (+1). When the index adders are being used as a data path, outputs go directly to SR(S-17), SR(21-S5) or the shift counter. During instruction overlap, the index adders supply a second memory address to the MAR even/odd switch. Note that during address modification, if no tag (tag of 0) is specified, 1's are gated into all index adder positions along with a carry into XAD(17). These ones effectively add a zero to the incoming value; therefore, the output value is the same as that supplied by the storage register or IBR. An index adder latch (XAD LTH) is used in conjunction with the index adders. The latch acts as a common delay (second half of a shift cell) in the circuit for modifying the index registers, program counter, address register and instruction backup register. These registers consist of triggers (not shift cells); the delay therefore, allows proper operation on successive clock pulses. The XAD latch is "free running" and always copies the contents of the index adders at the end of each 58 clock pulse. The receiving register receives the new information immediately thereafter which is slightly into the following clock period. When gating from the index adder latch to the XR, PC, AR, or IBR, four additional triggers are required as a short-time memory device to remember which register is to receive the output from the XAD latch. This remembering is necessary because the gating line to the index adders (XAD to AR, for example) will have already dropped before the new output can be gated from the XAD latch to the address register. Figure 44 shows condensed logic of the gating sequence for data flow from the index adder latch to the address register. Index Adder Position Figure 56 shows condensed logic for index adder positions 3, 4, and 17. As can be seen, the index adder is almost identical to the main adders discussed previously. The main difference between the XAD and AD is the number of inputs and outputs used. Also, because of the type of circuit cards used, some logic blocks appear as micro blocks whereas others appear as macro blocks. The basic function remains the same-that of determining whether or not there is an adder sum output. SR 3-17 SR 21-35 IBR3-17 AR True PC Compl PC ' - - - - - - + T o SR 21-35 To MAR Switch (3-16) Figure 55. Index Adder Routing The three basic inputs to the index adder are: input A, input B, and an input carry. As already discussed, the XAD inputs are divided so that the two values to be added arrive at inputs A and B, respectively. Logically, there never should be more than one input active at either input A or B at any given time. Using XAD ( 17) in Figure 56 as an example, there are two A-input gates (AO l l and A0 12 ) tied together to form a 6-way AND-OR function. Two B-input gates also exist (A0 13 and A0 14 ) but are not oR'ed together directly. The difference between the A- and B-inputs is due to circuit card usage and not because of logic. Any data input being active (Ad A2/BdB 2) produces "propagate 17" at 0 19 . Either AdA2 being active at 0 15 , and either BdB2 being active at 0 16 produces "generate 17" at A20 . These propagate and generate lines are used in the adder lookahead circuitry to be discussed later. The circuit combination of A17 , A18 , and 0 21 performs an exclusive OR test of the two data inputs. The in-phase (active) output of 0 21 indicates a "not exclusive OR" state (not A"'V'B); the A and B inputs are either both O's or both 1's. The out-of-phase (inactive) output of 0 21 indicate an "exclusive OR" state (A"'tB); either A or B is a 1, but not both. In the main adders, loading conditions required that this exclusive OR function be used in adder lookahead. In the index adders, however, loading is not as great and the exclusive OR function is not required outside of the immediate adder positions. The carry into XAD(17) affects only the sum output. The input carry is, therefore, combined with the output conditions of 0 21 in a second exclusive OR network consisting of A22 and OA 23 . Taking various examples of A, B, and carry inputs should quickly prove the circuit validity. Index Adeler Group Carry Lookahead The 15 index adder positions are divided into three groups of 5 positions each as follows: GROUP NUMBEB XAD POSITIONS 1 2 17, 16, 15, 14, 13 12, 11, 10, 9, 8 7,6,5,4,3 3 The object of dividing the index adders into three groups is to speed up adder operation and simplify lookahead circuitry. Group lookahead circuitry, as shown by solid lines in Figure 58, determines whether or not carries should be introduced into the low-order positions of groups 2 and 3. The lookahead circuitry for group 1 is also shown by dashed lines in Figure 57. Lookahead circuitry for groups 2 and 3 are different than group 1 but identical to that of the main adder as shown in Figure 53. Because of the similarity, see "Main Adder Group Carry Lookahead" section for a more detailed explanation. Index Adder 3 Carry A carry from the index adder, "XAD 3 carry," is used primarily as a test for successful transfers when executing class A instructions (TIX, TNX, etc.) from either the program register or IBR. Direct outputs from all three index adder groups are used in detecting an XAD ( 3) carry. The circuitry is shown as dashed lines in Figure 58 and is simply an extension of the lookahead circuitry used for groups 1 and 2. No XAD( 3) carry trigger is necessary; the circuit output is active long enough to allow all tests to be made. Again, carry lookahead is dependent solely on inputs to the individual adder positions; no individual adder or group carries are involved. Index Adder Compatibility Index Adder Bit Carry Lookahead As was explained with the main adders, each index adder position must be able to provide necessary carries into adjacent positions. For determining possible carries, two lines are produced from each adder position: propagate (P) and generate (G). The output of 0 19 , "propagate 17" (Figure 56), is active if any A/B input is a 1. The output of A20 , "generate 17," is active if both A and B inputs are 1's. With the propagate and generate lines, individual adder bit carry lookahead is developed as shown by the solid lines in Figure 57. This lookahead circuitry, again, becomes progressively more involved as it progresses from the lowest-order to the highest-order adder position within the adder group. No adder bit carry lookahead circuitry is provided for XAD(12). XAD(12), which is the low-order position of the next group, receives its carry indication from the group lookahead circuitry to be discussed next. When executing 704 or 709 programs under compatibility mode, XAD ( 3) and XAD ( 4) must be effectively bypassed. This bypassing can be accomplished by forcing the XAD position into a propagate condition (Figure 56). When the memory nullify trigger is on, "memory null" is applied through 0 6 to activate "propagate 3." Any resulting carry up through XAD ( 4) will, therefore, logically be "passed through" XAD ( 3 ) and give an XAD( 3) carry. Because XAD( 3) does not theoretically exist at this point, any possible sum output must be blocked. This blocking is accomplished at the output of OA24 by dot AND'ing with a "not memory null" condition. If the 16K/24K switch on the CE panel is in the 24K position, 24K of memory is available to the I/O compatibility program and 8K available to the executing program. Under these conditions, XAD( 4) must also be bypassed and the XAD ( 4) sum blocked. The methods used are the same as those described above for XAD(s). 7094 II System and Functional Components 59 Input AI AR 3 IBR 3 r---------~~~<3;r-- XAD 3 (03.0S.40. I) .-----------+l~ ~1~BR~2~1~_____~_H~====~~~~-~~~(4A),~_+~------~~r_~~_, I ~ '-'-:Input A2 r-- f----4 ~S~R~3_ _ _ _ _-+~~~-----+l'~ 04 0 (3A) ~_ _ _ _ _ _ _ _ _ _ _ _P_r_op_a_g_at_e_3_~ r-r--_-_-_-_-_-_-_-_-_--'--:--,-::J r -_ _ _~r-'- ~SR21 ~~ ..!.P~C~3~_ _ _ _+++++-+_ _~ A (4B) ' ~ 0 (2B) A ~ (IB) ,-I,;,- n ut B3 r-ill 06 (4D) ~ Generate 3 ~-- .~~J Input A2 ~~-+-l-+-~',.....:--:-ATOa r---- r- I 0 (3A) ~-------------P-ro-p-a-ga-t-e-4-~ e-+--I-I-I----',~t_; .:.P-=C_4~_ _ _ _l_+++!-·+--_++++++_-'~~t-; (4B)I--+.e-t-t--:~r---oI;;;~ Not PC 4 r-r-'-r-- - NotXR4 ~7 (2B).--. ~r==o=rI (~) 1~~.~----------G-e-ne-r-at-e-4-~~ . ~ (3B) ~ T ..:..=---'---'------+-4++-I-+--+-++_'~ A 09 "':":'~"':':"'-~--+-4++~I----'+-+4--+lfXR4 ~ A (4C)fOnes to XAD _24_K_M_o_de--'.(S_w_it_c_h)_ _ _-.t, J;;pu'ts2 ~i) f--=2_4K:..:...:.N~..:.U~II~~~~~~:~~~~~:::::;:~:~::::'l_'~"'4'::~oL) I~ ~ (2C) ~~~ (3C) j -_ _----' ->. ' - - 1~ 02.12.76. I ~ ~-+---===---~r--r- (IC) ~ ~NO Carry _I Carry In 4 lOA (:D) ... -.C~ ~:D XAD 4 Sum ~'-'-- In 4 I Not 24K Memory Null '-------------~(IE) 1'-- _ I . . . . . Dot AND ----- --- -----H-++-++.-+-H'--. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - XAD 17 (03. OS. 47. J) ~A~R~17___~-~~!_!_!_+_-++4~ A 0 1 ~?~B~~t~r7~IB~R~3~-~1~7--.~X~A~D~++I-+--+++-1~~ ~G~a~t~e~I~BR~2~1~-~3S~-.~~X~A~D~~I__+_+4_~~ AI/~BI/B2 ...:1.::::BR~3S=---_______-+-IH--+++~ A 4F) I-A__ I .....t-------~"' '--';:;-:-:::1 '--~ r------~~J ~G~a~t~e~S~R~3--1~7-~~X~A~D~__~--!_++_~r-'-ln-~p~~ -"S~R~I!....7..,.,....__-::-:::---,.,...,..:::____f-+-_+++~ A 012 A2 ~~~R~;~~~S~R~2~1~-3~5~-.~X~A~D~~~-+~---:~~~ Gate PC_XAD A C13 BI NotXR 17 A D O l " " 20 (3)i') ~ A ~ ~G~a~t~e~X~R~-.~X7A~D~-----------+-1~~ XR 17 I Figure 56. Index Adder Positions 60 Generate 17 (G,;nerate) T (2H) _ Ne~B A¥ B ~ --------------~ A L.-~ AI8 MQ5 (41) B2 (3H) L..- " - ' - I ~ LA~; 4H) BI u;.E Not .:.A~C::.....:.3S:"""=-=-=__""""'-=-_______--+l-"~ C1 ~2· Gate MQ S-5 .... XAD Propagate 17 ~~ ..:G::.:a::.;.t=:-e.,.::C;.:::o;:,;m~pX:;.:R--.=....:.X.::.A~D~------'_+-~ ... A ..:G~a~to:.e.!:A~C'-'3~O'_-~3S~~....::...~X.:...:A""D_ _ _ _ _ _~ ,In I AI/A2 -:-ca-t-~-C-o-m-p-p-C--.-X-A-D-----+++---'l~*; Not PC 17 019 ~ Carry In 17 0 21 (IH) ........."., T I NotA¥B ~ Not Carry, In 17 A 22 (21) r--r0 ~ A 23 r4' ~~ XAD 17 Sum XAD Positions 13-17 -----~.....I~f---- Generate Bit Carry (03.05.52.1) (03.05.57.1) --1 ....--X-A-D-I-3--- Propagate 13 Generate 13 I - - -- -, - 03.05.45.1 - r ..... - - - - XAD Sum 13 ~ -- - - - - - - o -- -- I ~------------- ..!13 ~--~-------------I , I I P14 P15' P16" P17'" I I I I I I I P14 P15 ~ I P16 G17 P14 P r o p o - I II ...- - - - - - - . gate XAD 14 14' I I _ .J!3.:..2? .46.1- _ r _ = A (2G)_ ==t:== (2F) '------- I--- -=----=--=-_--=- -:~ ~ ~-....--, -:1+- _=- .:- - _ ~ f-I .... (lG) I PI?: Kin" (41) PI5"7'= P16 ~ A ,Q..!Z.;: (4H) I I ,. ~ ~ Carry In (3H) ~ '--"--'- G15~ ~ I I : III ~ (4F) H :n=I"_~ ~ -=--=-"-- _-=--=--=--= -=--~: XAD Sum 15 ... ....>0.. 1""' . . ._ _ _ _ _.... .......-0 Ur I I I Generate 15 ~_--.J P15P16 --: A ___ Pr-'opo'-----+-1--+--1-'" gate 15 I P15~ ~ __+--'--I-+--+--4--I,--G_l_6~ ~ (4G) f------- G_15 -11- --' - - - - - - - - - - I I I I I I 03.05.46.1 Carry In 13 G14 ~ 1 7 ' (2 E) \ - - _ - - J II XAD15 ~- 0 -- A ....... P14? G15 I I S~m XAD _ _ _ _ _ _.. 14 --, .. = (2H) P15~ I I A G16 I Gen- I erate i. 14 r--- ,.---+ I ' I I -f----- - III I I I I I I I I' i I Generate 16 I I ~ P16 G17~ I I III I XAD Sum I ________~16 ~ ___________PI ------------~ 1AI I: I I I Kin I G17 - ~ - - ",.--;t:;I II TI I I T I I (4A) l I ~~~.~1__ XAD Sum 17 ~ .. A I H--f----- _______ ~5 -J (ts) l I 15 + ___ - _ - _ - _ - __ -.fl I Generate 17 ~ ~arry ~ (~) ~ A ~+tl ! - f - - I I , I Propagate 17 I I : I P17 ~ ,- (4D) (4E) -~ I II~ ....-X-A-D-I-7--.. -'- tJ ~ G16",-1 ~l _ ..2.3.:..22 ..£:!.... ~~ I I I P16_ P17 -~ A Kin -;. (4E) A PIS ---- -- ... --tTT- - f- - - - - - - __ - __G ' Propagate ....----~·16 I XAD 16 ,.----l Generate Gl ~~(,_2_1.).,t------l~-""'A_, I I (3D) --- ---- *. ... - - - - - - - - - - - XAD Group Lookahead and XAD 3 Carry U---I -----------------::.~I Any Gen"a!e G3 ~-----tr-~ 0 (3A-3B) r~-' II o (3D) Ge~ate ~ 4I ~o Propagate G3 i (4E) I : I I !~ro-ll~XAD3 L+t~~~a~_~ III 03.05.70. 1 ------------1-1-1 ------~ , III ~ III III Any Generate G2 r-t::,(3A-3B) II II --II II '~ ~ (4C) II ..----. , 0 II A ... ~J (40) II ,... 03 .05.55. 1 .-------. P8 1 8 I G8 P9 I 9 I G9 I ~ L:: P10 ~G10 - ~ ~ ..... ~P11 11 I G11 .!" (4E) (30) Generate or Propagate G2 .-------. P12 112 I , G12 I --'PT3 - - ~G13 ....---... I - A ... ~ o (5 1 I 16 --- A (3C) l -- Carry In Gr 3 03.05.70.1 ~-- ---- ----- ~ ... P14 , G15 (4C) A (40) ~~ '----" r-:lI (4B) (4B) ;;::l I A I .-------. P15 15 ~ 03.05.56.1 ------ -- -------------- - ~G14 Q. TO (~) l I I 1 ~ r-- ~ , I -.:. A ~() ... .... ~ ~ (30) Generate G1 Carry In Gr 2 ,.... P16 G16 .... ~~r-- :;;:, P17 .... : 1 NG17 (4G) -- ---- -- -- --- -- ---- Carry In 17 03.05.57.1 - - ---------- -- - - - XAO 3 Carry Circuitry------ Figure 58. Index Adder Group Lookahead and XAD(3) Carry 62 SR Zero Check (Systems 02.12.47.1 ) Outputs from storage register input gating (1-35) are fed to a zero check circuit as shown in Figure 59. Note that the test is made from data introduced at the storage register input gating and not from data presently in the storage register. In most cases, the data to be tested will not actually be set into the storage register. Many instructions gate the contents of the various CPU registers and adders to this zero check circuit as a check for either zero or equal conditions. Arithmetic operations make extensive use of this check circuit for detecting initial zero quantities (i.e., multiplier or multiplicand) or zero final results. When tests are made on the fraction of floating-point quantities, only positions 9:"35 are gated from the register concerned to the check circuit. Not Not Not Not Not Not Not SR Input 1 SR Input 2 SR Input 3 SR Input 4 SR Input 5 SR Input 6 SRTnout 7 .--- ~ A ~ ~ ~ is generated; when computer circuitry requires a reference to the odd memory, a «memory select odd" pulse is generated-if both references are required, both select pulses are generated. In Figure 60, 0 1 and O2 receive the circuit indications as to which memory/memories are required. These gating pulses occur at the end of the machine cycle, 6(D2) time, and also cause gating of MAR(S-16) addressing lines to the 7302-3. Just how these even or odd gating lines are developed will be discussed in the following section. The memory select pulses can occur at two different times: CPU operations at 7(D1) time; and channel operations at A2(D1) time. Data channel operations select memory ata later time because of the longer 12-point cycle. The memory, operating at a faster rate, provides the necessary data at the proper channel time. The reverse also applies when storing data from the channel. Note from A0 3 that channel operations assume priority-cpu memory selection is blocked until all channels are completed with either B or E time. ~ ? (4A) SR 1-7 = Zero '---- Not Not Not Not Not Not Not SR Input 8 SR Input 9 SR Input 10 SR Input 11 SR Input 12 SR 1nput 13 SR 1riPut T4 Not Not Not Not Not Not Not SR SR SR SR SR SR SR Input Input Input Input Input Input Input 15 16 17 18 19 20 21 Not Not Not Not Not Not Not SR SR SR SR SR SR SR Input Input Input Input Input Input Inout 22 23 24 25 26 27 28 Not Not Not Not Not Not Not SR SR SR SR SR SR SR Input Input Input Input Input Input Inout 29 30 31 32 33 34 35 Figure 59. ~~ "'7' , ""7' ~ (4B) ~ , SR 9-14 = Zero ~ A ~ -: ~-;;- ~ , (4D) ? SR 15-21 = Zero ~ ~ .---- ~ , (2D) ~~ A SR 1-35lnput = Zero ~ A7(Dl)S ... _ -,. ""7' (4E) ? SR 22-28 = Zero Gate Gate Gate Gate Gate ~r---;;~ ,.. PC MAR Even AR MAR Even IBR MAR Even BAR MAR Even XAD MAR Even ,. r--- 01 - (4A) I~ A5 -] ~ , SR (4G) SR 29-35 = Zero Input Zero Check Gate Gate Gate Gate Gate PC MAR Odd AR MAR Odd IBR MAR Odd BAR MAR Odd XAD MAR Odd ~ (3B) l -.,., -,. 1 A6 (3E) I,Mem Select (Odd) r02 (4E) L....--- Memory Selection Circuits (Systems 03.06.29.4) The 7302-3 core storage is divided into two logically independent units of 16,384 positions each. To initiate these two units, separate «memory select" pulses are required. When computer circuitry requires a reference to the even memory, a «memory select even" pulse ~41 (3A) Mem Select (Even) ,. 1 I--,--:I ,....-- Not Chan A 03 1Tr;;"e- Channel E Time MF Go Tgr 1J A7 (3F) 03.0"6.2'9.4 A (3C) Chan Time B Time '-- Channel A2(Dl) Dlyd Figure 60. Memory Select Circuitry 7094 II System and Functional Components 63 overall logic for all five controlling sources; the program counter controls as shown in Figure 61 are repeated at the top of Figure 62. Note the similarity of the controls and also that they control only MAR(4-16). Controls for MAR(S) are handled separately because of the MAR(S and 17) "switching" for normal and diagnostic modes of operation. Circuit tests are made to determine whether or not a bit should be sent on the MAR(S) bus. Four combinations of address bits (3 and 17) can exist: MAR Switching and Address Controls Only 14 of the 15 address bits are required to select anyone of the 16,384 even or odd memory positions. The remaining 15th address bit determines selection of the appropriate memory. Memory selection and MAR gating is effected by whether the computer is operating in the normal mode or diagnostic mode. In normal mode, the even memory contains all of the even addresses (0,2,4, 6, 10 through 77,7768); the odd memory contains all of the odd addresses (1,3, 5, 7, 11 through 77,7778). With this normal case, memory selection is under control of address bit (17)-if bit (17) = 0, select even memory; if bit (17) = 1, select odd memory. This selection circuitry for program counter gating is shown in Figure 61 at A2 and A4 , respectively. In diagnostic mode, the "even" memory contains the lower half of all memory addresses (0 through 37,777s); the "odd" memory contains the upper half of all memory addresses (40,000s through 77,777s). With this diagnostic case, memory selection is under control of address bit (3)-if bit (3) = 0, select even memory; if bit (3) = 1, select odd memory. This selection circuitry is shown at A3 and A5 • MAR selection and gating is under control of five sources: the program counter as just described, the address register, IBR, buffer address register for channel operations, and index adders. Figure 62 shows the Select even memory Select odd memory Select even memory Select odd memory PC-'MAR PC'" MAR Line 1 PC .... MAR Li ne 2 PC-.MAR Line 3 Not Diagnostic Mode Not PC 17 17 0 0 1 1 NORMAL MODE Select even memory Select even memory Select odd memory Select odd memory Selection of MAR( S)-even memory requires item 2 for normal mode and item 3 for diagnostic mode. These two cases are shown for the program counter at A6 and A7 in Figures 61 and 62. Selection of MAR(s)-odd memory requires only item 4. Note that this item is independent of either normal or diagnostic mode, and is shown at As in Figures 61 and 62. When operating in diagnostic mode, two sequential instructions cannot be fetched from the same memory on the same cycle. Because of the memory arrangement, "gate XAD to MAR" circuitry has no logic, and is, therefore, blocked at A4B , Figure 62. Not Block PC and XAD to MAR on Channel Trap Gate Gate Gate Gate 3 0 1 0 1 DIAGNOSTIC MODE 1. 2. 3. 4. 03.06.29.2 ,.... ~ Dot ~ :.,. /AND - (4A) Diagnostic Mode Not PC 3 PC 17 ~ .... (3A) ~ A3 I _ _ J g~ I Gate PC--.MAR (Even) ~ J, - . . - -1 Gate PC --. MAR (Odd) I~ (3B) ...-.- ~ A4 , (3E) 1 .... _ _ _ . I~ PC 3 >-- ... ~ 7 ~_ === (3C) 64 , l (3D) ----=tJllil ~ Figure 61. Program Counter MAR Bus Selection (2 F) A7 '.... 010 - A6 ~ ... ..... ___ (3F) (3G) (2D) PC MAR 3 (Even) PC MAR 3 (Odd) .... ... ,. Not Gate Gate Gate Gate Block PC and XAD to MAR on Channel Trap PC-MAR PC MAR line 1 .1 0 1 PC - MAR line 21 PC-MARline 3.1 (4A) 03 06 29 2 Dot AND ,/ ~ Not Diagnostic Mode Not PC 17 (3A) J g:) l ~ Diagnostic Mode Not PC 3 (31l) PC 17 ..... (3:) -~ J 010 I""'dC-MAR (Odd) ~--~ PC 3 ..... : : A6 ,...~ (3C) '- f-~ 011 (2 D) PC MAR 3 (Even) ~ PC MAR 3 (Odd) (3D) ~ ~~) - Gate Gate Gate Gate Gate PC-MAR (Even) ;fA - -- -- -- -- -- - - --- - - - - - - - - -r- AR-MAR AR-MAR line 1 AR- MAR Line 2 AR- MAR Line 3 03.06.28.9 0 (4A) ~~ Not AR 17 ...-r-- f---,;:f-- (3B) Not AR 3 ~ r- ~r-- I -=:! -------J J r- (~) I L:I -(~) I ~ Gate AR-MAR (Odd) PC 4 (3C) ~ -=;;-" I--r----' ~AR 3 (Even) (3D) A A (3A) A (3D) MAR 4 (Even) XAD 4 ~~ AR 3 (Odd) - - -- ---- -- - - - ~A 03.06.29.1 ~ r~ ~ (tA) r--~ (3~ L D ~ A (3C) A (3D) MAR 4 (Odd) (~B) ~'-- I 03.06.28.1 Gate IBR-MAR (Even) ·~I -=:I '-C:; (~F) ;::::::= ..... t;: '-r-- - .... r- .::; (~) Gate IBR-MAR (Odd) ~ I A ~ J (2~) - - --- - 17=1,~~~ .~ Hot One to BAR Bus 17 ..... ~BAR 17 .... - ~ (4A) h :B AQO (4B) (3B) = r;! (~E) Not BAR 3 - MAR .... BAR 3-MAR - -- (4D) --~ (4G) - - - - - ~t Diagnosti~e NotBlockPCandxADp':j" A to MARon Channel Trap (4B) I PC 16 IBR 34 (~D) I ~ BAR MAR 3 (Even) BAR 16-MAR XAD16 I I ..... 'A""'O ~ "A I-- ~ (3A) A A (3B) MAR 16 (Even) ~~ BAR MAR 3 (Odd) - - - - - --- --- --- - - -03.06.28.8 ~~~ ~ I-- a a A Dot AND A / (4A) -- ~ ~ XAD 17 L~ XAD 3 I MAR I I A Gate XAD*-MAR Gate XAD*-MAR Line 1 Gate XA[)'MAR Line.1.. Gate XAD--MAR Line i Gate XAD*-MAR Line 4.. I AR 16 D I I I I lGate BAR-MAR (Even) ~ 8F) IGate BAR-MAR (Odd) I I I ~ ~r-;;- I I I I I h ~ I I IBR MAR 3 (Odd) !'"""---A - - - - - - - - - ---03.06.29.3 17 I I I IIBR MAR 3 (Even) A (3G) I I I I I I I A (3C) = I I ~ ~~ (~E) ~ IBR 21 GaleB~M~ 0 --A A (4A) IBR 35 Figure 62. A IBR 22 BAR 4-MAR 0 Not IBR 21 Not XAD 17 --'A'Cl AR 4 ~ Not IBR 35 BAR 17 ..... MAR (BAR (Odd) 03.06.29.3 Gate AR-MAR (Even) A Gate IBR-MAR Gate IBR-MAR Gate IBR-MAR Dot OR /MAR3 A dm-- t:~ ~- 1" 0 (31) I 0 (2B) ~--,;:- ~r-- MAR 3 (Even)' ~ ~-=:---= AR 3 0 (3H) f ~==:=:= AR 17 ~~ }jrl ~ (3D) MAR 16 (Odd) (3Q A Gate XAD*-MAR (Even) ~'-- 03.06.28.7 Gate XAD-MAR (Odd) XAD MAR 3 (Even) XAD MAR 3 (Odd) Bus Selection and Switching 7094 II System and Functional Components 65 Timing Master Clocks and Pulses All of the computer functions are directly related to two master clocks shown in Figure 63. These clocks located in the multiplexor, provide the basic pulses necessary for CPU and channel operation. The heart of these clocks is a 5.71 megacycle oscillator which produces a complete output cycle once every .175 microseconds (175 nanoseconds). Each positive and negative oscillator output pulse is, then, approximately 87 nanoseconds long (Figure 64). The CPU clock is an 8 cycle-point clock composed of eight triggers (cpu clock triggers 0-7); the channel clock is a 12 cycle-point clock composed of 12 triggers (channel clock triggers 0-11). In either case, the clock forms a closed ring-each trigger is turned on in sequence and provides a gated output of 175 nanoseconds. The combined outputs produce the basic cycle times as follows: = CPU Cycle: 8 x .175 lAO microseconds Channel Cycle: 12 x .175 2.10 microseconds = Clock drive pulses needed to sequentially step the clock triggers are produced by the clock drive trigger (Figure 63). Oscillator outputs gate two controlling AND circuits to the clock drive trigger; the top AND circuit is used to set, and the lower AND circuit is used to reset the trigger. Delayed outputs from the clock drive trigger condition the input AND circuits such that the trigger changes state with each positive oscillator pulse. Outputs from the clock drive trigger ("even clock drive~' and "odd clock drive"), Figure 65, are a series of pulses 175 nanoseconds in duration. The clock drive trigger has acted to halve the output rate of the ·oscillator. Both the CPU and channel master clocks are reset under power-on conditions or by depression of the console clear key (Figure 63). At the end of the reset pulse, the DLY-AND circuit combination produces a 100 nanosecond "start clock" pulse which turns on both the CPU clock 0 and channel clock 0 triggers. Using the CPU clock as an example in Figure 63, the next "even clock drive" gates the CPU clock 0 trigger output to produce an AO(Dl) pulse of 175 nanoseconds duration. This same AO(Dl) pulse, besides being gated to other circuitry, also turns on the CPU clock 1 trigger. 66 Note that even though the CPU clock 1 trigger is turned on at O-time, the output of its AND circuit does not become active until "odd clock drive," 175 nanoseconds later. The main logical output of the trigger is, therefore, Al(Dl). As a result of the preceding logic, the Al(Dl) pulse turns on CPU clock 2 trigger. This trigger output when gated with "even clock drive" produces an A2(Dl) pulse. Referring to Figure 63, CPU clock 2 trigger turning on at I-time acts as a reset to CPU clock 0 trigger, and the A2(Dl) gated output acts as a set pulse for CPU clock 3 trigger. This sequence continues through CPU clock 7 trigger. Rise of the A7(Dl) pulse turns the CPU clock 0 trigger back on and the clock continues to run in a closed ring. I/O operations on the 7094 II are bascially the same as for the 7090/7094. Timing conditions, however, are such that the 1.4 microseconds cycle is too fast to support channel operations without considerable rework within the channel itself. Therefore, to accommodate the channels and simplify the 7094 to 7094 II conversions, the 12 cycle-point clock has been retained strictly for use by the channels. Looking at the channel clock in Figure 63 shows that operation is identical to the CPU clock except that the cycle duration is extended beyond clock 7 time to include clock 11 time. Note that even though the CPU and channel clocks are separate circuits, both are reset, started, and under control of common clock logic. From Figure 66 it can be seen that two channel cycles occur for every three CPU cycles. Therefore, at every third CPU cycle both clocks align themselves with respect to O-time. More information concerning channel operation and timings will be found later in this section and in volume 3. Figure 67 shows outputs and controls significant to each stage of the CPU and channel clock rings. Figures 68 and 69 shows sequence charts for the CPU and channel clocks respectively. Note that each clock trigger is on for 350 nanoseconds, twice the time duration of an individual clock pulse. A particular clock trigger is turned on one clock pulse early and only the last half of the output is gated by the clock drive pulse. This slower SWitching of the clock triggers provides increased reliability in the clock operation as well as additional pulses usable in the computer. Set Pulse Drive r--5.71 MC Osc ---(5A) .------~,.. ~J J A (4C) Clock Drive L~ Even Clock Drive ~---------------, T - ,. Odd Clock Drive Clock Drive (Off) 1rDi;r l.J (2C) lClock Drive (On) I Power On Reset 100* I ~ &lJ Clock Reset on Clear I... (:H) Clock Start Clock Reset __ ~OO.44.~~ _ _ _ _ _ _ _ _ 1 II Ch~O l+ T(OR) ~ _ f----',.~I R(4C) r C=0 ~ ~ __ l---~----------------08.00.40.1-08.00.43.1 All (D2) 08.00.43.2-08.00.43.3 T(OR) ... _ MPXR A7(D2)... ~ ~ A - (3B) t7' _Chan AO(Dl) I r-----4-+-~f-------~ AO(D2) J----I (:E) IChan Clk 2 - t7 ' .. R(35) I Al(D2) ---- I .-f----:..~R(4i) ,.~ (:H) 1 . r------r~+_++-----~ L7' _ ... MPXR A3(D2) MPXR Al (D2) ~MPXRA'(Dl)- ~ .~ ..B..illL ...Chan A2 (D 1>.... 1 ~MPXRA1(Dl) ... r~ _Chan Al (D1)... ~r_A_3~(D_2~)_~~_.~~~ MPXR AO(D2) ... ,. ,. I !chan Clk 3 ~rA2(D2) ... A4(D2)e-+---l~~~ _A_4~(_D2~)_ _~+-_... ~RtKt ,. ---:........;.. , MPXR A2(D2) ~ MPXR e-1---i""''''R(4Cf --'----'-+-+---l~~~ ,~ Turn~ T , Chan Clk 4~ ' , I Chan Clk 11 MPXR Al0(Dl) ~ ------~-~, T A6(Dl) .. - ---'----'-+-+-~~ _ ~) I Channel Clock CPU Clk 7 I Al0(D2) T MPXR A6(D2) .. _~ .ChaoAl1(Dl) ~ T J (:G) I~XR ________~J A7(Dl) .. ,. CPU Clock Figure 63. Master Clock Logic Timing 67 CPU Clock Tgr. Turned On by A7 AO AI A2 A3 A4 A5 A6 0 I 2 3 4 5 6 7 Figure 64. 5.71 MC Oscillator Output (01) (01) (OJ) (OJ) (01) (OJ) (01) (OJ) Channel Clock Tgr. Turned On by 0 I 2 3 4 5 6 7 8 9 10 II AII(OI) AO (01) AI (OJ) A2 (01) A3 (OJ) A4 (OJ) A5 (OJ) A6 (01) A7 (OJ) A8 (OJ) A9 (01) AIO (01) Figurc 67. CPU Turned Off by Clock Clock Clock Clock Clock Clock Clock Clock 2 3 4 5 6 7 a I Tgr Tgr Tgr Tgr Tgr Tgr Tgr Tgr Turned Off by Clock Clock Clock Clock Clock Clock Clock Clock Clock Clock Clock Clock 2 Tgr 3 Tgr 4 Tgr 5 Tgr 6 Tgr 7 Tgr 8 Tgr 9 Tgr 10 Tgr I I Tgr 0 Tgr I Tgr Odd Ring Drive Start Clack Pulse CPU Clack 0 Trigger MPXR AO(Dl) Clock Pulse Figure 65. Oscillator Output and Even Clock Drive CPU Clock 1 Trigger MPXR Al(Dl) Clock Pulse CPU Clock 2 Trigger MPXR A2(Dl) Clock Pulse CPU Clock 3 Trigger MPXR A3(Dl) Clock Pulse CPU Clock 4 Trigger MPXR A4(Dl) Clock Pulse CPU Clock 5 Trigger MPXR A5(Dl) Clock Pulse CPU Clock 6 Trigger MPXR A6(Dl) Clock Pulse CPU Clock 7 Trigger MPXR A7(Dl) Clock Pulse 68 CPU and Channel Cycle Relationship Figure 68. CPU A7 AO AI A2 A3 A4 A5 A6 (02) (02) (02) (02) (02) (02) (02) (02) Clock Sequence Chart Gated Output AO AI A2 A3 A4 A5 A6 A7 (01) (OJ) (01) (01) (01) (01) (01) (01) Tgr Ouration and Output Gated Output A II (02) AO (02) AI (02) A2 (02) A3 (02) A4 (02) A5 (02) A6 (02) A7 (02) A8 (02) A9 (02) Ala (02) AO (01) AI (01) A2 (01) A3 (OJ) A4 (01) A5 (OJ) A6 (01) A7 (01) A8 (OJ) A9 (01) Ala (01) A II (01) and Channel Clock Output and Controls Even Ring Drive Figure 66. Tgr Ouration and Output Even Ring Drive Odd Ring Drive Start Clock Pulse I I LlJ Ll.J Channel Clock 0 Trigger' -W AO(D1) Clock Pulse ~~ Channel Clock 1 Trigger -W-------iLJ ____~n~____~n~___ I I I I LJ ::::::IC~::k:I:;gg" ----'~r-------+i-...u A2(D1) Clock Pulse Channel Clock 3 Trigger A3(D1) Clock Pulse Channel Clock 4 Trigger A4(D1) Clock Pulse Channel Clock 5 Trigger A5(D1) Clock Pulse Channel Clock 6 Trigger A6(D1) Clock Pulse ::;::~IC~~:k:I~~;gg" -----'~ _ _--.~ Channel Clock 9 Trigger A9(D1) Clock Pulse Channel Clock 10 Trigger A1O(D1) Clock Pulse Channel Clock 11 Trigger A11(Dl) Clock Pulse ! rLL---- W I I W ~ I ~ I W I L..J I u______~~~____~I--~n~----~I--~~ W I ~----~I--~ I l...J I W-- ----~~~--~I--~~~~I--~~ W -~~ 2J I W..-----.I-----.W- I ~r----L....-I-~ 2J..--.. . ----,~ I I! _ _ _ I Channel Clock 8 Trigger A8(D1) Clock Pulse ~ i L...J :Lj'------ _ _ _ _ _----...In I r1 I ----W I"""---+-I ...... I ------~~~~I----~~ W I I I I Wr-...!-[----- ----------~n~i~------~n i w Wr------_______ I ~~ I n~ ____ Figure 69. Channel Clock Sequence Chart Timing 69 CPU Clock Pulse Distribution The CPU clock is located in the multiplexor and the output pulses are distributed to both CPU-I and CPU-2 for usage. Because of inherent delays in logic blocks and cable transmission lines, clock pulses arrive at the various CPU gates after finite delays. Figure 70 shows the typical distribution pattern of a CPU pulse. This MPXR A3(Dl) pulse is generated by the CPU clock 3 trigger (Figure 63) and sent to both CPU-I and CPU-2 for usage. By the time the MPXR A3(Dl) pulse arrives at CPU-2, however, the MPXR A4(Dl) pulse is being formed. Therefore, to provide alignment of pulses between the multiplexor and CPU-2, the original MPXR A3(Dl) pulse is relabeled as MF A4(Dl) upon entering CPU-2. As the original MPXR A3(Dl) pulse continues on its way to CPU-I, additional delays are encountered similar to the ones just described. With the help of a DLY block, the incoming pulse to CPU-I is delayed a small amount and again relabeled as a MF A A5(Dl). There is, therefore, a skew of one clock pulse between the CPU-2 and multiplexor, and a skew of two clock pulses between gates A and B of CPU-I and the multiplexor. A5(Dl) pulses throughout the computer, for example, are the result of three different CPU clock trigger outputs in the multiplexor. Except for troubleshooting purposes, these various outputs need not concern the reader. The computer circuitry doesn't care where the pulses come from as long as they represent the correct timing or logic. Although the AO(Dl) through A7(Dl) pulses are the prime outputs of the CPU clock, the D2 pulses are also distributed and used throughout the computer. Pulses of other durations such as A3(D3), AO(D6), etc., are produced throughout the systems as needed by using either triggers or combinations of AND and OR circuits. 70 In the computer, clock pulses are gated during cyclic operation and then labeled I6(Dl) or E5(Dl), depending on the particular cycle of operation. Whenever an A pulse is encountered in studying the computer, for example A6(Dl), it means this pulse is used directly from the clock and is independent of the computer's cycle of operation. This A pulse ·always occurs at 6 time and is always available. CPU 1 A and B Gate Clock Pulse Designation Timing pulse alignment becomes increasingly more important as the internal computer operations are compressed into smaller and smaller intervals of time. In the A and B gates of CPU-I where a major portion of the computer operatiol)s are performed, the clock pulses have been systematically delayed, aligned and labeled accordingly. In Figure 70 the clock pulse coming into the A gate of CPU-I is directed into two similar groups of circuitry. The top group of pulses labeled LN B (Line B) are sent to the B gate for usage; the bottom group (LN A) remains in the A gate for usage. Pulses used in both A and B gates of CPU-I also have an additional letter and number designationA5(Dl) G3, for example. In Figure 70 "-F A5(Dl) G5 LNA" is the earliest A5(Dl) pulse used in the A gate. This pulse delayed through an inverter produces a "+ F A5(Dl) G4." Additional inverter delays produce "-F A5(Dl) G3," "+F A5(Dl) G2," and "±F A5(Dl) R» pulses. Note that the higher the G number, the earlier the pulse. By picking the appropriate pulse, compensation can be made for the number of levels of logic delay preceding a trigger or register gating circuitry. Multiplexor (Figure 63) CPU 2 CPU Clk 3 CPU 1 (D Gate) CPU 1 (A Gate) 02.15.70.6 MPXR A3(D1) R MF A4(D1) 02.15.08.1 Odd Clock Drive -F A5(D 1) G5 LNB +FA5(D1)G4 LNB : -F A5(D1)G3 LNB B Gate Usage +F A5(Dl)G2 LNB I ' I - F A5(D1)R LNB LMJ -FA5(D1)G5 LNA +FA5(D1) G4 LNA -FA5(D1)G3 LNA +FA5(D1)G2 LNA Figure 70. CPU Clock Pulse Distribution A Gate Usage Figure 71 shows four typical examples of pulse usage. At circuit (a), the A5(D1) G4 pulse is used because of the four levels of logic in front of the register gate. At circuits (b) and (c) the G3 and G2 pulses are used because of the three and two levels of logic in front of the register gating. The R (register) pulse, circuit (d), is the latest pulse and is usually used as a direct input to a trigger. There are, of course, exceptions to the pattern of usage just discussed. Note that the G-pulse levels alternate between +F and -F at each step because of the natural inversion from each DIF logic block. The majority of the triggers and register inputs use + level logic (i.e. +G or +AOI). CP Set Pulse Generation and Distribution Computer set pulses (cp set pulses) are developed directly from the master oscillator. These pulses are pre- r cisely generated and timed to set triggers and registers, and control much of the gating and shifting of data within the computer. Width and timing of the Cp set pulses, as related to the clock pulses, are extremely important to successful machine operation. The input set pulse drive pulses are received directly from the oscillator. A variable delay is used to pOSition the cp set pulse when aligning it with the clock pulses in the computer. Actual CP set pulse settings are made at the A and B gates in CPU-I. Figure 72 shows condensed logic of how the CP set pulses are shaped, delayed, and distributed for specific usage in the computer. The inverters in Figure 72 have been included, not because they perform logic, but because they contribute to the timing delays. 2 Levels (a) +F A5(Dl}G4 r-1 Level ----:i~ -F +F -AI Gate +G (b) -F A5(Dl)G3 (c) +F A5(Dl}G2 * Refer to Figure 70 Figure 71. CPU Clock Pulse Designations and Usage 72 (d) +F A5(D])R A Gate (02.15.61.1) CP Set (SCS) CP Set CP Set A CP Set B CP Set C CP Set D CAS CP Set ---------------------B Gate (02.15.61.2) .... Simi lar to A Gate Above ....-"l-----~~~J--- Figure 72. CP Set Pulse Distribution Machine Timing Cycles 7904 II computer operation consists of several types of machine cycles concerned with both CPU and channeloperations. CPU CYCLES I II E L Instruction IBR Instruction Execution Logical CHANNEL CYCLES B Chan I Chan E Chan L Buffer Channel Instruction Channel Execution Channel Logical The CPU cycles are directly concerned with CPU operations and the 8 cycle-point 1.4 microsecond clock. The channel cycles are directly concerned with channel operation and the 12 cycle-point 2.1 microsecond clock. The cyclic sequence of a computer instruction is fixed and, depending on overlap conditions, begins with either an I or II cycle. The total number and types of machine cycles required for each instruction is determined by the number of steps to be performed before the operation is completed. This number varies depending on the conditions set forth by the particular instruction. The various cycles essentially perform the following functions. Each will be discussed in greater detail in later sections. I Cycle: The I cycle occurs because of a break in the overlap sequence. References are made to core storage and two sequential instructions are received into the program register and IBR for execution. II Cycle: The II cycle occurs simultaneously with an E or L cycle and is used to perform functions concerned with the overlapping instruction in the IBR. E Cycle: The E cycle is used for data or IA (indirect address) cycle references to core storage. This E cycle can be initiated by an instruction in either the program register or IBR depending on the conditions of overlap. L Cycle: During the L cycle the computer performs logical or arithmetic functions without reference to core storage. This L cycle can be initiated by an instruction in either the program register or IBR depending on the conditions of overlap. B Cycle: During the B cycle, a data channel uses core storage for either accepting or delivering data (or a data channel command word) in connection with an input-output operation. This B cycle occurs simultaneously with L cycles but never occurs simultaneously with I, E, or II cycles. Timing 73 Chan I Cycle: The channel I cycle supplies 1.05 microsecond I time gatings to channel banks 1 and 2 and is used primarily for reset functions. Chan E Cycle: The channel E cycle supplies 2.1 microsecond E times gatings to the CPU circuits as well as channel banks 1 and 2. This gating is used primarily during channel operations requiring references to core storage (i.e., POD'S 54/64 or ENB). Chan L Cycle: The channel L cycle supplies 2.1 microsecond L time gatings to banks 1 and 2 for channel circuit controls. The CPU I and E cycle timings are such that they each gate their own instruction and data memory references. To accomplish this, I or E time for addressing comes up early (6 time of the previous cycle). There is, therefore, considerable overlapping of cycle times as shown in Figure 73. I I 6 Time Gating I I ~ I 6 E Time Gating I i i I 7 L Time Gating " Time Gating i I Figure 73. Cycle Time Relationship 74 I iI 6 I U I 6 .... i --~L I I Master I Time I time is not the steadily recurring type of cycle that existed on previous systems; it results only because of a break in the overlap sequence of instructions. For example, this break in sequence might occur because of a successful transfer or skip condition resulting from the instruction in the program register or IBR. Figure 74 shows condensed logic of I time. Note that even though the I time trigger might be turned on, its outputs may be blocked until such time that channel B cycle demands are satisfied. I time outputs are also gated by the master stop trigger or machine cycle gate (manual operations). The basic objectives of the I cycle are to: 1. Select the proper memory/memories and gate the corresponding address( es) out to MAR( s). 2. Update the program counter to the proper address. 3. Gate the proper instruction(s) from the even/odd storage bus into the program register, storage register, tag register and IBR. 4. Perform address modification when applicable. 5. Determine and initiate the next type of machine cycle. Machine Cycle Gate Master Stop Trigger Off o Manual I Time Control (IG) 04.20.10.1 Not B Interrupt A5(Dl)R CP Set Not IBR Go to E L FAD E Time End-O End-Op Condition Overlap Conflict Condition I Time Trigger Skip Trigger Pre IA Trigger I Time Early (Addressing) Test I Late Dlyd I Late Delayed Al(Dl) 08.00.18.3 E Time Trigger Program { Register Controls E Time Early (Addressing) E Time (for IA) Pre IA Trigger Not Indirect Addressing 08.00.19.2 (2G) Gen Reset Not B Interrupt For L Time IBR Controls A6(Dl) R CP Set B (To CPU 2) L Time Early L Time Late Figure 74. I, E, and L Time Cycle Logic Timing 75 Address Gating Objective 1 is initiated in the upper area of Figure 75 (sheet 1). This area shows the decisions required for initiating an I cycle. The skip trigger coming on as the result of an instruction in the program register nullifies overlap and forces an I cycle because the instruction in the IBR is going to be bypassed. If the skip trigger does not come on but there is an overlap conflict condition, overlap will, again, be inhibited. In this latter case, the computer waits until the instruction in the program register has ended operation before initiating an I cycle. Special FAD E time end-op circuitry has been added because of timing considerations. If the skip trigger is off and there is no overlap conflict condition (Figure 75), a test is made to see if the instruction being executed from the IBR will send the computer into either an E or L cycle. If this ElL test is successful, the computer proceeds to ElL time to complete execution of the overlapping instruction. If the test fails (the IBR contains a I-cycle instruction, for example) the computer waits until the instruction in the program register completes operation and then initiates an I cycle. Note that the I time trigger is turned on at 5 time (three clock pulses before the end) of the previous cycle. Turning the trigger on early allows its outputs to provide addressing gates starting at 6 time. Machine cycle design is such that the upcoming cycle provides its own memory address gating. With the exception of MAR addressing controls, the I time does not logically start until the following 0 time. Any B times requested by a data channel will be serviced before the I cycle is allowed to continue. This blocking of I time is accomplished by degating the output of the I time trigger with "B interrupt." I time outputs can be grouped as follows: I time early for addressing: These outputs are available starting at 6 time of the preceding cycle and are used primarily for gating the proper instruction addresses to memory. I time early: These outputs are delayed outputs of the I time trigger and are available primarily for gating functions at the beginning or during early portions of the I cycle. I time late: These outputs are delayed longer than the early pulses and are, therefore, available for gating functions during the latter portion of the I cycle. The I time trigger is turned off at I5(DI) time so it is the delayed outputs that maintain the required gating until the end of the cycle. I time late delayed: This output is used to decrement the program counter under HTR control so that the counter will indicate the actual address of the HTR instruction. 76 Going into I time is like getting a fresh start in the program sequence and because of this, two sequential instructions will be referenced from core storage and brought into the CPU. Except for certain operations, the address of one of the next two instructions to be executed will always be found in either the program counter, address register, or the address portion of the IBR. The address of the other instruction can be obtained by incrementing or decrementing the first address. Both of these addresses when gated to the even and odd memories will retrieve the next two desired sequential instructions. Note that the addressing blocks in the lower area of Figure 75 are divided from left to right into distinct groups according to function. Group 1-Address Gating to MAR: One of these three gates will be active to provide one instruction address to MAR. This same address is gated to the index adders for either incrementing or decrementing. Group 2-XAD Gating to j\1 AR: This block provides the other instruction address to MAR. In the majority of cases this second address is an incremented (+ 1) value of the first address. In some cases, however, this second address is a decremented (-1) value of the first address. Note that the XAD to MAR path is only active when the computer is in overlap mode and not diagnostic mode. Group 3-1's to XAD (3-17): These two blocks cause incrementing or decrementing in the index adders. Gating a 1 to XAD(17) is an unconditional I6(D2) func~ tion and causes address incrementing (+ 1). Gating 1's to XAD(3-16) along with the 1 to XAD(17) causes address decrementing (-1) by performing 1's complement addition. Group 4-Address Updating Paths: This fourth functional group takes the incremented (or decremented) XAD address output and routes it to either the address register or IBR for temporary storage. This address will update the program counter at Il(DI) time. Group 5-BAR Address Gating to MAR: This last functional group is used during channel trap operations when addressing the specific data channel trap location. When progressing through the series of I time addressing decisions, a channel trap is concerned with group 5 and directs the routing of the buffer address register (BAR) to MAR. For all other I time addressing decisions, groups 1 through 4 are used according to the following rules (Figure 75): l. One block (and only one block) is always used from group l. Either the PC, AR, or IBR will contain the correct address of one of the next two sequential instructions. This same address is also gated to the index adders. 2. An entry will always be made into group 2. Whether or not the index adders are gated to MAR depends on both overlap mode and diagnostic mode. 3. In group 3, a 1 is always gated to XAD(17) for incrementing purposes. During the two cases where decrementing is involved, the 1's to XAD(3-16) are also used. 4. One block (and only one block) is always used from group 4. These updating paths route the index adder value to either the address register or IBR as temporary storage until the program counter can be updated at the next Il(Dl) time. Note that the value set into the address register or IBR at this time normally corresponds to the second instruction selected from memory (i.e., the instruction which will be set into the IBR at 14 time of the I cycle). U sing the above rules and making a systematic progression through the maze of addressing decisions should prove the logic of the I time addressing. Program Counter Update Objective 2 updates the program counter to the proper address. The program counter is always one step ahead of the current instruction being executed; therefore, the address to be set into the program counter will correspond to the instruction destined for the IBR. In the previous section, addresses were sent to MAR during I6(D2). At the same time, the incremented (or decremented) value from the index adders was also routed and set into either the address register or IBR (Figure 75, sheet 1). These group 4 routing paths place the updating program counter address in "temporary storage." During Il(Dl) time, this address is routed unmodified to the program counter via the index adders (Figure 75, sheet 2). Note that the pattern of decision making for the updating circuitry is similar to that for address gating discussed previously. If the address register or IBR had received an incremented (+ 1) address from the index adders, that corresponding register contains the correct updating address for the program counter. If, however, the address register or IBR had received a decremented (-1) address from the index adders, the register originating the address contains the correct updating address for the program counter. In some cases the program counter already contains the correct address. When these cases occur, the updating paths are blocked and the program counter value remains unchanged. One case, a successful class A transfer in trap mode, requires a carry to XAD(17) during update to put the value 00001 into the program counter. Storage Sus Gating The third objective is to gate the storage buses into the computer. The instructions which were addressed at the previous 6 time will be available by 4 time on the even and odd storage buses. In Figure 75 (sheet 2) an external trapping condition will nullify the two new instructions arriving from core storage and, instead, force an STR operation by turning on PR (s, 9). Excluding the trap condition, five routing and setting functions are available for the even and odd storage buses: SB(S, SB(S, SB(S, SB(S, SB(S, 1-35) ~ SR(S, 1-35) and TR (18-20) 1-2) ~ PR(S, 8-9) 3-11) ~ PR(S, 1-9) 1-35) ~ IBR(S, 1-35) 14(Dl) Set Pulse 1-35) ~ IBR(S, 1-35) 14(D2) Set Pulse Excluding the XEC instruction and channel traps for the present, the program counter is the deciding factor as to which storage bus is routed into the program register (PR) and IBR. Remember that the program counter was already updated (incremented) at II time and is +1 ahead of the current instruction. Because of this, if the program counter is at an odd value the even storage bus is gated to the program register and the odd bus to the IBR. If the program counter is at an even value the odd storage bus is gated to the program register and the even bus to the IBR. A class A instruction coming into the program register is "pre-sensed" from the storage bus. If the instruction is a one cycle class A instruction (TIX, TNX, TXH, TXL, TXI) circuits are immediately set up to also route this same instruction into the IBR with an I4(D2) pulse. Note that this longer I4(D2) pulse overrides the instruction placed in the IBR with the normal I4(Dl) pulse. The reason for this special gating into the IBR is because of addressing considerations which will be discussed in Volume 3. Tracing through the flow chart (Figure 75-sheet 2) will also show the storage bus gatings for channel trap and XEC conditions. In some of these latter conditions the program counter value (even/odd) may cause some unnecessary gatings into the IBR. These gatings cause no problems, however, because overlapping is prevented at this time. Figures 75 (sheet 2) and 76 provide a summary of the various SB gating conditions. After the program register has been set with the new instruction, a check is made to determine if overlapping is possible. If so, the IBR loaded trigger is turned on. Timing 77 Address Modification Objective number 4 is address modification. This modification is performed during the I5(Dl) clock pulse so that the modified address (if an index register was specified) will be available for MAR gating during the next two clock pulses; i.e., E6(D2) early. Class A instructions (TIX, TNX, TXH, TXL, TXI) block address modification. Instead, the I4(D2) period of time is used to perform the specific test or operation on the index register (Figure 75, sheet 2). If the conditions of the test are successful, a PR condition-met-trigger is turned on; if the test fails, the trigger remains off. If the instruction in the program register is indexable, SR(21-35) and the complement of the specified index register(s) are gated to the index adders with a carry to XAD(17). This 2's complement addition effectively performs a subtraction of the index register value from the address portion of the instruction. If no index register is specified, all 1's are gated from the index register along with the carry to XAD(17) to effectively subtract zero from the address in SR(21-35). The modified address resulting in the index adders is gated to the address register for MAR gating at SB Gating E6(D2) time. If the instruction is a POD 76, the index adders are also routed to the shift counter for further decoding. Next Machine Cycle At the same time that the address modification or class A testing is being performed, other circuitry is determining the next machine cycle. If the PR instruction contains bits in positions 12 and 13, SR( 12, 13), and meets the other conditions shown in Figure 75 (sheet 2), the pre-IA trigger is turned on and an E cycle follows. If no E(IA) cycle is called for, a test is made to see if a I-cycle instruction is in the program register. If so, the end-op trigger is turned on and another I cycle follows. If the instruction in the PR does not call for ending opera tion, tests are made to see if an E cycle should follow. If the test is successful, the master E time trigger is turned on and an E cycle follows. If, however, neither an I or E cycle is requested, the master L time trigger is turned on and an L cycle follows. SB Gating Oecision Blocks XEC Operation EVEN SB SS.... SR and TR SB(S,l-2)-'PR(S,8-9) SB(S,3-11)-'PR(S,l-9) SB ....,BR 14(01) SB~IBR 14(02) OOD SB SB....SR and TR SB(S, 1-2)~PR(S, 8-9) SB{S,3-11)+PR(S,l-9) SB .... ,BR 14(01) SB~IBR 14(02) 78 SB Normal ~ I~1i s{ef- a.ddlOa 1 f.-' r es fr1~"/--,--",,""\ $ I;s't~~sc~on ~ PC The instructions coming into the computer / ore blocked. Instead, the computer is forced into on STR operation by setting PR(S,9). Chon Trap Off ~PC Od:j!~" ;;:~' 'h, 'BR 00"'" W(~C";;:~ Od:j!~" ~''" ~~~~~d;(~~}- ! \03.06.30.1(2AY . I 5B(S,I-2)-PR(S,8-9) 03.0S.13.1 (11) 03.14.00.1 (IA) ! D \o3.06"30.1( 2A SB-IBR IIII SB -IBR 03.0B.13.1(3C) 03.0B.13.l(3C) -t ________ Instructions Such as PAX, PAC, PCA,PCD, etc. Details of their execution found in volume 3. ~~~ SB-SR, TR II 02.12.52.2(3E) G H T I ~ ~~~ /4(:'; 5B(S,I-2)-PR(S,S-9) 03.0S.13.1(IH) 03.14.00.1(lC) SB(s,3-11)-PR(S,I-9) 03.0S.13.1(IH) 03.14.00.1(lC) ~v=-l~d _ _ _ _ _ _ _ _ _ _ 14([)1) SB-IBR 14(D2) Set PR(S,9) 14(D2) SB-IBR + _______ 03.0B.13.1(3F) 03.0B.13.1(3F) I 03.14.00.1(2A) ~dd~ f f{ 03.08.16.2 No Class A Instructi on s F I ~(';.~ [~D2) SB(S,3-ll)-PR(S;I-9) 03.08.13.1 (II) 03.14.00.1 (IA) J1~3.16.13.1(4BV In:!~~:t~on ~ 'llis \03.06.13.1(4BV E Odd V~ ~ .... C ~e~ _ _ _ _ ~ _ _ _ _ _ ~C PpdC-E-ve-n--'PC-O-d· 'd/Even J \03.06.30.1(2AL~ B 02.12.52.2(3A) / Yes :d~3.06.13.1(4Bi In;tlr~s;t~n ~ vd I~~;~~~an ~ vd In~:~~~t~on ~ ~ ';~t~~Sc~on ~ Y~ In~!~~:t~on ~.d I~s~~~c~on ~3.06.13.1(4B)/~3.06.13.1(4BV ~3.06.13.1(4By ~3.06.13.1(4Bf·r II ~ /4~ I On PC Even A SB -SR, TR 03. 14.00. 1(2B) Manual operations are covered in a separate section in volumn 3. This is the I time preceding the E time of the STR operation. The channel trap MDBO latch prevents the IBR loaded trigger from being set during the I time following the STR operation. - - - - - - - On 03.08. 17.2(4A) AR Odd Tgr 03.06.30. I (2A) 03.06.13.1 (4B); Yes Off 03.0S.13.1(3G) },PCCpr<' '1 PC Odd/Even Odd D3R XEC Trigger n t FP Trap or Interrupt (Chon Trap Dem) 3.14.00.1 (2B) No ..,~ &I /J I I ' L I . C I"........ /7 ,"-:;,V"uC.T'OVl ctV~ "Tv-ctPlstev 1/ ;hSi-Y"Wrt-;c' .L6J ; If (Note 1) Manual Ctrls No PRPODto Inhibit Ovlp No I5(Dl) Tum On IBR Loaded Tgr 03.0S.16.2(3F) t .?= Yes (TIX, TNX, TXH, TXL, TXI, STR) 03.06.13.1 (4B) PR Indexable (PR 6 & 7) 02.10.65.2(5A) No l I .J 15 Dl) )II( 15(Dl) SR(21-35j-XAD Carry-XAD(I7) 03.06.03.2(4H) ,03.06.13.2(3E)1 03.06.07.1 (2C) ,03.05.52.1 (5A) 15\Dl) Camp XR-XAD 03.06.03. 2(5F) ,03.06.07.1 (2C) ~ Yes 14(D2) Comp XR-XAb 03,06.02.3(5 F) , 03.06.07.1 (2C I I t o 02.12.76.1 Note: A POD 76 instruction 15 CP Set _will have address modification XAD-SC performed if togged. 03.06.03.2(41) ,03.06.11.1 (3F) Yes 5 ( ) RI2,13=I's_ 02.10.65.2 (5E) / ~N I 15(ep) Set Turn On End-Op Tgr 08.00.09.2(3B) I Time Sheet I --- jI 15 CP Turn on Pre Set IA Tgr 02.10.65.2(3B) 0 I L Time Figure 75. I Time Flow Chart (Sheet 2 of 2) J6(Dl) Reset AR 03.06. 14.2(3D) the first I cycle of a transfer instruction ·in trap mode. See IA Flow Chart forDetails 14_ _ _ _ _ _---'N~o_< pe':m'i~ed?}-YJ.-"e""s'----..." , . - - - - - -Yes - - { Cycle Inst I Turn on PR Cond Met Tgr 03.06.19.1 (3D) T ~____________.l______________~J / ,l4---------------------------------------------I This reset will also occur for ~ N I N~TraClassA Cond Meth e s "'------------+, 15 CP Set XAD-AR 02.12.70.1 (4D) ,03.06.02.1 (5A) 14(D2) SR(3-17) - XAD 03.06.13.1(2B) I( 14(D2) 1-XAD(l7) 03.06.07.1 (2C) ,03.05.52.1 (5A) SR(21-35). If no XR. is specified, all l's gated from XR output circuitry. (IH) Yes 03.06.13.1(4C t XR value subtracted from instruction address, Na~OD76 '--------{03.01.01.1 SIR No es 16(DI) Turn On Master E Tgr 08.00.19.2(2B) E Time SB Gating XEC Operation 58 Gating Decision Blocks Chon TrOD I~ ~ 1KY1 f01 EVEN SB V SB-5R and TR SB(s,I-2) __ PR(S 8-9) V sB(S ,3-11) __ PR(S, 1-9) IBR J4(Dl) 5B L-SB -IBR J 4(D2) ODD 5B sB-5R and TR SB(S, 1-2) -PR(S,S-9) SB(S 3-11) - PR(S, 1-9) V 5B-JBR I4(01) SB IBR 14(D2 V V V Norm,,1 KVl I~ ~ ~ V V V V V V V V V V V V V V V V V V V- V V V V V V V V V V V V V V V- V V V- V V- V V V V V V V V- V V Master E Time Figure 74 shows condensed logic for the E cycle. Note that this E cycle can be initiated by an instruction in either the program register or lBR depending on overlap conditions. In either case, the E time trigger is turned on by 6- time of the preceding cycle to allow memory addressing. In this manner, E time provides memory addressing for its own data reference. Outputs from the E time trigger may be blocked until channel B time requests are satisfied. This same situation exists for I time, which was described previously. After all of the B times are satisfied, the blocking is removed and the computer proceeds with the instruction. The master E time trigger is reset at E4 time; however, by use of delays, E time gating circuits are extended until the end of the cycle. E time outputs include: 1. E time early for addressing (not deconditioned bYlA) 2. E time early 3. E time early (for lA) 4. E time late 5. E time late (for lA) Note that, unless specifically labeled «for lA," E time gatings (with exception of E time for addressing) are blocked during an lA (indirect address) cycle. In this manner the instruction execution is blocked until the normal E cycle (not lA cycle) occurs. If an instruction requires several consecutive E cycles (CVR, for exampIe) the master E time trigger is set and reset for each cycle. IA E Cycle Indirect addressing requires making a second reference to core storage. This second reference either obtains the actual data word in cases of data handling instructions (CLA, ADD, etc.), or the new transfer address in cases of transfer instructions (TRA, TQO, etc.). During the first memory reference (IA E cycle), a new word is brought into the storage register from a location specified by positions 21-35 of the original instruction word. Positions 21-35 of the lA word contain, in effect, a new memory reference. This new address when sent to core storage retrieves the desired data (or transfer address) . Several examples of indirect addressing are shown in an earlier section of this manual, «Instruction Addressing." The basic objectives of an lA cycle are: l. Make proper reference to memory for the IA word. 2. Gate the storage bus even/odd into the storage register. 82 3. Perform address modification according to the tag specified in the lA word. 4. Determine and initiate the next type of machine cycle. 5. Select the proper memory(s) and gate the corresponding address(es) to MAR(S). An lA cycle can be initiated by an instruction in either the program register (I time) or the lBR (II time). See Figure 77. Test conditions for the IA cycle are similar in either case. Indirect addressing has the following restrictions: 1. The instruction must be indexable. Non-indexable instructions such as the class A type (TlX, TNX, etc.) make use of the entire decrement portion of the instruction word for indexing purposes. Therefore,. positions 12 and 13 lose their meaning for indirect addressing. Other non-indexable instructions such as PAX, PXA, SXA, SXD, etc., actually operate on the index register and are, therefore, not subject to address modification. 2. Except for transfer instructions, indirect addressing is only possible on instructions of 2X, 3X, 4X, 5X, and 6X operation codes. Codes below 2X and above 6X require L cycles for their execution. Without references to core storage, indirect addressing has no purpose. 3. Transfer instructions can normally be indirectly addressed. However, this feature is nullified (except for TTR/ESNT) if the machine is in the trap mode of operation. When in the trap mode, the address of the transfer instruction is stored in location 000008 and if the transfer conditions are met, the computer traps to location 00001 8 • Because a transfer is never actually made to the transfer address, indirect addressing in the trap mode accomplishes no useful purpose. The programmer's trap subroutine must test to determine if indirect addressing existed in the transfer instruction. If the lA cycle is initiated by the program register instruction, a pre lA trigger is turned on. If the IA cycle is initiated by the IBR instruction a pre IIA trigger is turned on. In either case, a common lA trigger is turned on at the next EO time. II times are blocked from occurring simultaneously with the E( lA) cycle because of a conflict in usage of the index adders at 5 time. Note that during an II cycle the instruction in the program register may detect a skip condition which will bypass the overlapping instruction. A trap condition may also demand recognition ahead of the overlapping instruction. In either of these two cases, overlapping is nullified and the E( lA) cycle is blocked from occurring. If the lA cycle is requested by the overlapping instruction (i.e., II time), the lA cycle must wait for an end-op signal from the instruction in the program register. This condition would occur, for example, if the overlapping instruction is preceded by a multiply instruction. See example on Figure 77. IA Memory Reference: The address to be gated to MAR depends on whether the IA cycle was initiated by a preceding I or II cycle. This fact is determined by the OFF and ON states of the end-op trigger, respectively, and causes gating of either the address register or IBR to MAR (Figure 77). Note that only one address is sent and only one memory is selected. A check is made to determine if the value in the address register or IBR is even! odd. If odd, the AR odd trigger is turned on for later references. The even!odd conditions are also dependent on whether the computer is in normal or diagnostic mode of operation. Storage Bus Gating: The data word arriving at E4 time is gated into the storage register. The proper bus is determined by the status of the AR odd trigger (Figure 77). Address Modification: At E ( IA) 5 (D 1) time, Figure 77, SR( 21-35) are routed to the index adders together with the complement of the specified index register. A carry to XAD( 17) causes 2's complement addition and the index register value is effectively subtracted from SR( 21-35). This new modified memory address is set into the address register for gating to MAR at E6( D2) time. This modified address is also gated to IBR( 21-35) if the IBR loaded trigger is off. If the IBR loaded trigger is off, the IA cycle was initiated from an II cycle. Under these conditions, an II cycle may follow the IA cycle. Routing the modified address to the IBR allows use of memory conflict detection circuits on Systems 08.00.22.2 (5F). Next Machine Cycle: One cycle transfer instructions end-op during the IA cycle; therefore, the next cycle will be an I cycle. Two cycle transfer instructions require an L cycle to complete operation. Because of this, an L cycle will follow the IA cycle. If a I-cycle or 2-cycle transfer condition does not exist, the next cycle must be an E cycle (except for possible B cycle interrupts). If conditions allow, an II cycle may also occur simultaneously with this next E cycle. Note that the path which turns on the II time trigger (Figure 77) also turns on the E time trigger. The reverse condition, however, is not true. MAR Gating and Selection: A I-cycle E ( IA) transfer end-op causes two sequential addresses to be sent to MAR from the address register and index adders. This condition was covered in a previous section on I time and in Figure 75. All other conditions force a data E cycle with the modified data address sent to MAR from the address register. If an II cycle is also allowed to occur, the program counter is sent to MAR to fetch a new overlapping instruction. There is one condition on the PC~MAR gating, however. The IBR loaded trigger being on indicates an II time without a reference to memory because the IBR had already been loaded with an instruction during the previous I time. Timing 83 Basic IA Cycle Objecti yes Ves 1. Address memory for IA data word - E6(D2) 2. GateSBtoSR- E4(Dl) 3. Perform Address Modification - E5(Dl) 4. Initiate next machine cycle - E6(Dl) 5. Address Memory -l/E/lI6(D2) t - - - - - - - - - - - - - - - - - 1 - . _ Indirect Addressint not Indirect Address not possible - proceed normally. The Pre IIA trigger is normally reset at possible - proceed normally. the next A2(Dl) time. ~ The Pre IA trigger is reset at the An overlap conflict or skip condition will reset the Pre lIA trigger immediately at 6-time. next A2(DI) time. II time is not allowed to occur simultaneously with an E(lA} cy~ 1. Reset II Time Tgr 1. Reset I Time Tgr 2. Block II~II 08.00.22.2 (2C) 08.00.22.2 (5E) 2. ResetIlTimeTgr 08.00.18.2 (3B) 08.00.22.2 (2C) On Indicates that the ~ instruction in the PR is skipping around the overlapping Inst. Ves A function of the U Indicates that a condition (such as a channel? trap) is nullifying the overlap operation. The f time trigger can not be turned on until the imtruction / in the PR has completed operation. Example: 0 CLA n 1 MPV II 2 ADD' I The end-op trigger being off ot this time indicates an IA cycle initiated by the PR instruction during an I cycle E(lA) The end-op trigger being on at this time indicates on IA cycle initiated by the IBR instruction during an II cycle. The end-op trigger is turned off at E4(D I) time of the I A cycle. Systems 08.00.19.2(41 ) Reset ot E7 CP set time of the IA cycle E4(Dl) SB-SR 02.12.52.2 (3A) ----.r-- E4(Dl) SB(18-20) ........TR 02.12.52. 2(3A), 0J.05.22.1 (SA) The IBR loaded trigger being on ot this time indi cates entrance from an I cycle with on overlapping instruction in the IBR .,..------.-::~..( The IBR loaded trigger being OFF ot this time indicates entrance from an II cycle. }---:::::.:....------, The modified address is also placed into the ~ IBR to make use of the PC-IBR memory conFlict checking circuitry on Systems OB.OO.22.2(5F}. On Ve, This is really a memory conflict check of the AR and PC. The AR ====~ value was also set into the lBR at the previous E(IA)5(Dl) Time ....... Turned on because of no request for 1 or E time Both paths must be used if the 11 time trigger was turned ON. E(IA)S CP Set Turn On I Time Tgr 08.00.18.2 (3C) If the IBR loaded trigger is ON at th;s time, the tBR had already been loaded ----....r--..On with a second instruction during the preceding I time. See the I time flow chart (Figure 76) for complete detoi Is of address gating. / 16(D2) AR-MAR AR+l-MAR 03.08.15.1 ~ Figure 77. IA Cycle Flow Chart "'" ~~---A76~C~P~S-et-------' Turn On L Time Trigger 08.00.20.2 (3C) Master L Time HTR L time provides a logic cycle of operation during which the computer performs functions not related to core storage. Control circuitry (Figure 74) is similar in some respects to E time because it can be initiated by an instruction in either the program register or IBR depending on overlap conditions (only IBR shift instructions can force an L cycle). The turn-on logic is such that an L cycle occurs if the instruction does not specifically call for an I or E cycle. The I or E turn-on is at A5 CP set; if these I or E controls are not present, the master L time trigger is turned on one clock pulse later at A6 CP set. As long as the input PR or IBR controls are active, the L time trigger is turned on again each cycle. In this manner, the L time trigger remains on as long as needed. Use of delay circuitry produces both "L time early" and "L time late" outputs for circuit controls in CPU-l and CPU-2. When the master stop trigger is turned on, the computer remains at whatever cycle time it was proceeding to when the master stop trigger came on. An inhibit L trigger is turned on by the master stop trigger (Figure 78) to block both the master L time and II time outputs (Figure 74). The inhibit L trigger is also turned on by the HTR and HPR instructions. L time of these two halt instructions is blocked until the start key is depressed. At this time, the inhibit L trigger is turned off and the computer either transfers or proceeds to the next instruction. B Interrupt For L Time MST Off Not I Time Late 6 CP Set (Blocks L & II Times) 08.00.21.2 Figure 78. Inhibit L Trigger Logic During channel operations, the CPU remains in L time until the next instruction is begun. Master II Time II time (IBR I time) is the instruction cycle for the overlapping instruction in the IBR. During this cycle address modification and preliminary tests are performed by the overlapping instruction. When not prohibited, II time occurs simultaneously with either an E or L cycle of the preceding instruction. Figure 79 shows basic condensed logic of the II time trigger circuitry. Note that II time can be initiated by instruction controls from either the program register or IBR. The II time trigger is turned on at 6 time of the preceding cycle to allow memory addressing. In this manner, II time performs memory addressing for its own data reference. Details of II time together with an II time How chart are found in Volume 3. II Time Tgr E Time Tgr CP Set Conditioning and Blocking Circuits II Time Early Gen Reset Pre IIA Tgr II Time Late 08.00.22.2 Figure 79. II Time Condensed Logic 86 Master B Time B (buffer) time is the cycle during which one of the data channels makes reference to core storage. Most I/O devices move at a fixed rate and therefore request or transmit data at specific intervals. When data word requests are not serviced in time, information is lost and I/O checks result. Because of this timing requirement, channel B cycle demands are serviced during the cycle immediately following the request. If the next CPU cycle is to be an I, E or II cycle, the output of the master trigger is blocked. If the next CPU cycle is to be a normal L cycle, both the Band L cycles occur simultaneously. "B cycle demand" is sent out by the channel early in the cycle such that the signal arrives at the multiplexor before channel 9 time. Provided the computer is not servicing a channel trap or a POD 64 instruction, the master B time trigger is turned on with the next channel A9(Dl) pulse (Figure 80). The POD 64 (store channel instruction) restriction prevents a possible alteration in the channel registers which would cause false error indications while running CE diagnostic programs. B time requests are blocked during channel trap operations to prevent destroying the trap address in the buffer address register. The B time request occurs at channel 7 time. Channels requesting additional B cycles cause the B time trigger to be turned on again in time to prevent the CPU from regaining cycle control. "Retain priority" conditions (TCH or IA) recognized by the multiplexor lookahead circuitry bypasses the normal turn-on controls. The B interrupt trigger (Figure 80) blocks outputs from the master I, E, and II time triggers either during B times or when the master stop trigger is on. Note that the B time trigger is under control of channel clock pulses whereas the B interrupt trigger is under control of CPU clock pulses. The B time trigger turning on causes the B interrupt trigger to turn on at CPU-5 time which is early enough to block the next cyclic output from the master I, E, and II triggers. MST On AS(Dl) B Interrupt MST Off AS CP Set Ch A9(Dl) Not POD 64 Not Trap to Channel B Cycle Demand (Blocks I, E and II times) A Retain Priority (3G) Ch A7(Dl) B Time (B Time Gating) Figure 80. B Time Condensed Logic Timing 87 interrupt trigger only blocks one CPU I, E or II cycle. Case 2 shows the channel 9-CPU-5 timing relationship where both the B time and B interrupt triggers turn on during the same clock pulse. In this later case, however, the timings are such that the B interrupt trigger remains on for one extra cycle and two CPU cycles are blocked. Because of timing relationships between the channel and CPU clocks, the B interrupt trigger can be turned on and off for two cases (Figure 81). Case 1 shows the channel 9-CPU-I timing relationship where the B time trigger turns on at channel 9 time and the B interrupt trigger turns on four clock pulses later at CPU-5 time. Considering only one channel B cycle request, the B Item B Cycle Demand B B Time Trigger I C B Interrupt Trigger ! I I CPU I, E & " times blocked due to B time B Cycle Demand CPU I, E & " times blocked due to B time 01234567012 B-7 3~5670 I I I I I i 1 I I I I B-CPU5 : B-CPU~ I I l i t I I I Ie ..... -------1 I I I I II CI I __ ~I----_I I I I I I II I I I I I I I I 3 I I A-9 r B-CPU5 I I I 2 l~ ___________________: I B-7 I I s-cPu51 ,~__________~"______~______ I ______~i""-I""""""""-I~~ I I C1 . . . . . . . . . . . . . . ._ _ _ _.. '~------------~,--------l~------~:--------------~ Figure 81. B Time Sequence Chart 88 I I I I, I 'I II D 7 I~----------r___----,~--~r I B Interrupt Trigger 6 I A-9' I B Time Trigger C ~567012345 3 I CASE 2 A 2 I I I : D 0 0 0 1 2 3 4 5 6 7 8 9 10 11 1 2 3 4 5 6 7 8 9 10 11 1 2 3 4 ~ 2 ~~----.I-------I--~I 3 A 10 11 I C B Cycle The main system objectives associated with B cycles are to: 1. Accept B time requests from the data channel 2. Set BAR to the proper address 3. Select the proper even/odd memory 4. Route data and commands to the proper system areas 5. Test and perfonn TCH and IA functions 6. Initiate subsequent B cycles where applicable B cycle requests are initiated by circuitry from within the attached data channels and are based on the immediate channel demands. In most cases, the "B cycle demand" is honored during the cycle following receipt into the cpu. Refer to Figure 82. A "B cycle demand" arriving during the execution of a channel trap operation or POD 64 (store channel) instruction is delayed until after their completion. These delays are necessary because of conflicts in usage with the buffer address register, and for diagnostic programming reasons, respectively. "B cycle demands" are initiated at approximately channel 3 time and arrive at the cpu circuitry in time to tum on the B time trigger with a channel A9(Dl) pulse. The B time interrupt trigger is turned on at cpu 5 time to block the master I, E, and II time outputs. This blocking remains in effect until cpu A5 time following the turn off of the B time trigger (Figure 81). Instruction overlap is suspended during all B cycles. Only one memory (even/odd) is addressed, and only one memory select pulse is generated. Therefore, only one data word is placed on the storage bus. SDW Cycle The first function perfonned (Figure 82) is setting the buffer address register (BAR) to the proper memory data reference as specified by the channel address switches (CAS). BAR was reset by a previouschannelAl1 pulse and is now set at BO(D2) time. For a read BDW cycle, the channel store trigger is turned on to provide the necessary memory "read-out" and "store" controls. Selection of the proper memory and BAR gating is based on BAR(17). The memory select pulse is initiated at approximately B2 time. For a write BDW cycle, the data word is available on the storage bus at approximately B7 time for sampling into the channel's data register. For a read BDW cycle, the channel has the data word on the storage bus at the proper time to be sampled into the memory buffer register (Figure 35). SCW Cycle Consider a nonnal BCW cycle with no indirect addressing or TCH command (Figure 82). Accepting the "B cycle demand," turning on the B time trigger and selecting memory are as explained previously. The channel determines the memory location by routing an address from its location counter/command counter into the multiplexor buffer address register. Note that all BCW cycles have the characteristics of a write operation where the data word is taken from memory and sent to the channel. The data command read from memory is placed on the storage bus and sampled into the channel's operation register, word counter and address register. The channel's location counter (7607) or command counter (7909) is stepped +1 to indicate the next sequential command's location. If the channel is performing a write operation at the time of the BCW cycle, a BDW cycle is immediately requested. This BDW cycle may not be initiated in time, however, to prevent the cpu from regaining program control. This channel must also seek priority with other channels requesting B cycles on the system. Indirectly Addressed SCW Cycle Consider IA commands other than TCH (Figure 82). Accepting the "B cycle demand," turning on the B time trigger, and selecting memory are the same as explained previously. The initial pass through the flow chart finds the IA address control trigger off. Position 18 of both the even and odd storage bus are oR'ed together, MB(18), as a test for indirect addressing. oR'ing both storage buses is valid because only one of the buses will contain information. Ignore for the moment the special case of a proceed type command with a word count equal zero. The command word on the storage bus is sampled into the channel's operation register, word counter, and address register. The operation register and word counter contain valid data; the address register contents will be replaced with a new value during the next cycle. Because of SB(18), an IA address control trigger is turned on which generates and sends a "retain priority" signal to all channels (Figure 82). In the multiplexor, "retain priority" produces an immediate "B cycle demand" and prevents the B time trigger from being reset. In this manner a second B cycle is initiated and the cpu is prevented from regaining control. The IA address control trigger being on allows an IND 18 trigger to be turned on and its "IND 18" Signal sent out on banks 1 and 2 to all data channels. The address portion of the initial BCW data word represents a new memory reference. Because of this, multiplexor storage bus positions 21-35 are routed directly into the buffer address register (Figure 35). The data word resulting from the second BCW cycle is placed on the storage bus as before. This time, however, the data channel only accepts the address portion (21-35) into its address register; the original operation coding and word count remain unaltered. Following this second Timing 89 cycle, the channel's location counter/command counter is stepped +1. The IA address control trigger is turned off at channel A3(D2) time and the IND 18 trigger is turned off at the following channel AI0(D2). Timing is such that only one level of indirect addressing is permitted. A bit in SB( 18) of the second BCW word is ignored and the IND 18 trigger is not turned on again. A special case of a proceed type command with a word count equal zero was ignored in the previous discussion. An IOCP, IORP, or IOSP command with a word count equal zero must be bypassed. Indirect addressing, therefore, performs no logic and is ignored (Figure 82). The command is routed to the channel and set into the operation register, word counter and address register as usual. Channel circuitry recognizes the zero word count condition and initiates another "B cycle demand." Multiplexor lookahead circuitry does not retain priority in this case. BCW TCH Command Consider a TCH command without indirect addressing. The TCH acts as a channel transfer instruction to alter the sequence of I/O commands being executed. The transfer is accomplished by altering the value in the channel's location counter (7607) or command counter (7909). The channel determines the memory location by routing an address from its location counter/command counter into the multiplexor's buffer address register (BAR). Accepting the initial "B cycle demand," turning on the B time trigger and selecting memory are as explained previously. The TCH command is detected (Figure 82) by testing the storage bus for not S, not 1, 2. The 7607 data channel's operation register, word counter and address register are set as usual; the 7909 sets only the address register and a TCH trigger. A TCH address control trigger is turned on in the multiplexor and a second B cycle is immediately initiated by "retain priority." The address portion of the TCH command contains the transfer-to address. This address is immediately set 90 into the buffer address register for the next memory reference. The address is also set into the channel location counter/command counter to indicate the new I/O command sequence. The second BCW cycle places the new I/O command on the storage bus and the operation continues as explained preViously. This new command (Figure 82) may be a TCH command, another I/O command, or an indirectly addressed I/O command. Indirectly Addressed TCH Command An IA TCH command operates similar in.most respects to the normal TCH (Figure 82). During the first BCW cycle, the storage bus indicates not S, not 1, 2, and 18. As a result, both the "TCH address control" and "IA address control" triggers are turned on. "Retain priority" circuitry in the multiplexor immediately initiates a second BCW cycle and turns on the IND 18 trigger. The new (IA) memory reference is routed from SB(21-35) and set into the buffer address register. Note that "IND 18" control circuitry in the channel prevents setting of the location counter/command counter during the first BCW cycle. At channel A3 time, the IA address control trigger is turned off but the reset to the TCH address control trigger is blocked. "Retain priority" is maintained because of the TCH address control trigger and the multiplexor immediately requests the third BCW cycle. During the second BCW cycle, the address portion of this IA data word contains the actual TCH transfer-to address. This address is immediately set into the buffer address register for the third memory reference, and also set into the channel location counter/command counter to indicate the new I/O command sequence. The IND 18 trigger is turned off during the second BCW cycle which, in turn, allows the address control trigger to be turned off at channel A3 time. The third BCW cycle places the new I/O command on the storage bus and the operation continues as explained previously. From Sheet 2 1. 2. 3. 4. 5. 6. Wait for channel B cycle demand Accept B time Requests from channel Set BAR to proper address Select the proper even/odd memory Route data and commands to proper system areas Test and perform TCH and IA functions Initiate subsequent B cycles where applicable Wait unti I completion of either the channe I trap or POD 64 /V'" operation ~ The B Time Interrupt trigger is turned ON at CPU A5(Dl) time to block the Master I, E, and II time outputs. This blocking remains until CPU A5 time following the turn OFF of the B time trigger. From Sheet 2 (Start Next Sequential BCW Cycle) On Reset at Chan A 10 time following turn off of the IA address control trigger B11 (Dl) / Turn On Ind 18 Trigger 06.10.01.l(2C) The master B time trigger reset occurs at every Channel A7(Dl) 7 The I A Address Control trigger --z.-and TCH Address Control trigger will both be in the OFF status on ~:":" :";:":";':" :" :' ':' ' ';' :-Jfu' "'oo,h >h. fI~ ,hort. 06.10.01.1(2G,2H) Off Data Channel BAR/ 1. Loc Cntr (7607) 2. Cmmd Cntr (7909) B8 -A3 SB -BAR / 06. 10.00. 1(3C) ~ This gating occurs after BAR has been set from the SB Ch All(Dl) Reset BAR 03.06.27.1(3H) The new address is routed directly from multiplexor circuitry into BAR. The All(Dl) BAR reset pulse does not occur at this time because of "retain priority" being active. ChAll(Dl) Reset Chan Store Tgr 01.00.00.1 (3C) BO(D2) CAS --BAR 06.10.00.1 (2B) BDW Cycles - CAS set from Channel Address Counter. BCW Cycles - CAS set from Channel Location Counter (7607) or Command Counter (7909). ...-z---::: BCW Even Memory RO Odd Memory RO Store prefix Core Storage Controls ~ Store Decrement Store Tag Store Address 01.00.00.1 (2D) (2E) (4D) (2F) (4E) (4F) B2(Dl) Dlyd Select Memory 03.06.29.4(3F) B2(Dl) Dlyd Select Memory 03.06.29.4(3B) To Sheet 2 Figure 82. B Time Flow Chart (Sheet 1 of 2) Timing 91 From Sheet 1 On (Data Word Cycle) BOW ! This trigger will always be OFF on the first pass through the flow chart. The trigger being turned ON prevents a second IA cycle. (Only one level of indirect addressing is allowed}. Off ~Also LlPT command for 7909 , . - _......_'" ~ data channel No Yes No Yes .¥'I--J~~P Channel Commands 10RP IOSP / .( Storage Bus (3-17) = 0 Yes Data Channel Set (7607) 1. Opn Reg. 2. Word Cntr 3. Addr Cntr Data Channel ~ 1. Opn Reg 2. Word Cntr 3. Addr Cntr Write Channel initiates another request for B (BOW) Cycle and retains priority. The multiplexor retains priority because the TCH address control tri gger was not turned off during the previous cycle. To Sheet 1 (Start Next Sequential BCW Cycle) To Sheet 1 Figure 82. B Time Flow Chart (Sheet 2 of 2) 92 Data Channel Set 1. Opn Reg 2. Word Cntr 3. Addr Cntr Step 1. TciCCntr/ 2 Cmmd Cntr Data Channel Set 1. Op,;-Reg 2. Word Cntr 3. Addr Cntr 1. 2 Channel Cycle Times Channel L Time Three channel cycle time triggers (Figure 83) have been added in CPU-2 to supply 2.1 microsecond cycle times for data channel usage. While these three cycle time triggers are controlling the channel, the CPU waits in CPU L time until the channel operation is completed. The end of the channel operation is indicated by the "MF go" trigger which turns on the end-op trigger and sends the CPU into its next cycle time. The channel L time trigger (Figure 83) provides 2.1 microsecond cycles necessary for channel operations. These operations include all of the I/O select, sense, and test instructions. POD 54 instructions (RCH/LCH/ RSC/STC) also require L cycles prior to the time that the channel may signal a "proceed to E." Note that the channel L time trigger is turned on at channel 11 time which may be aligned with either a CPU 3 or 7 pulse. Synchronism is not performed as with the preceding channel E time trigger. Channell Time The channel I time trigger (Figure 83) is not the first channel cycle trigger to be turned on as might be thought. Instead, it is the last channel cycle time available (I time next) and performs mostly housekeeping functions (I time resets) in the channel. Channel E Time The channel E time trigger (Figure 83) provides a 2.1 microsecond cycle for channel operations requiring a reference to core storage. These operations would include: enable, store channel, POD 54(RCH/LCH/RSC/STC), and channel traps. Circuit controls and timings are such that the channel EO time is aligned with the CPU 4 time. This alignment is necessary for POD 64 and enable instructions. The channel E time indicates the end of a channel operation. Because of this, the MF go trigger is turned on late in the channel E cycle to force an end-op condition and allow the CPU program to continue with the next instruction. Channel-CPU Cycle Time Controls Two triggers are used to control cycle times. The channel "L-E end" trigger controls channel cycle times, and the "MF go" trigger controls the CPU time. The channel L-E end trigger coming on signals the end of the channel operation. In some cases, the channel I time trigger is then turned on to accomplish housekeeping resets to the channel. The MF go trigger is normally on. At the start of the channel operation, the MF go trigger is turned off which, in turn, turns the "L end-op sync" trigger on (Figure 83). When the channel operation is completed, the MY go trigger is turned on again and the output of tlie "L end-op sync" trigger is gated to produce an "E or L end-op." This latter signal turns on the end-op trigger and allows the CPU to continue with the next instruction. Timing 93 SOD 02 (Read) SOD 06 (Write) I (~) '"AO Not Tape Class AdcIress Chan End Op Chan A 1O(D 1) Any I/o Test gj;]IhO" ~ or Sense , (4C) , Channel E Time ,. } To Channel Banks 1 & 2 Channel L Time , 08.00.27.1 -: -AO ~, ~ ~ (4B) "" ,..--- ... ~ AO 4.- I-~ POD 54 Proceed to E (4D) H ~ AO ~~ Chan E Time ~ L-I-~ Channel Trap ~ L Any Select or Test or POD 54 A (3F) J---------... Not POD 54 ~ .-1- (4E) ~ ~ A (4F) ~ ~ t---<~ I-r:; Elf R 3F AO 08.00.28.1 .... ~ ~ Not MST I Time Early (4F) ... ~ = - AO Chan L Time ~ POD 54 Reset Ctrl CTearTtri Interlock Reset On Load Channel I Time I R(3D) '~ I (~) POD 64 IA Ctrl Tgr Off' Not A3(D2) (CPU Pulse) Ch All (D1) Not B Time Not BCW Not Proceed to E T Ch A6(D1) Enable r Hm. ~ T (4G) J (5~) I ~ I -====. ~r> L ~J-------..... (:H) ~ "'"'R'('3R) ~R(3R) 0 (4B) 04.20.12.1 '--'---'- 08.00.29.1 Reset CT Tgrs A (4E) ~ T(OR) AO , (4F) AO ~~ TLJ Not Manual Stop Not Channel Trap 13 D1) Tape Class Addr or POD 54 ~1mHr ~ >-- ~~ 08.00.27.1 ,....... (41) L End-Op Sync r--- L-- A (4S) L Time Late n I'I' ~ U ~ r=-= ~ A ~o(Off~] P"'" 0 I-- A (4D) '---- Ch AO(D1) L Time A4(D1) POD 54 POD 64 (3C) ~ -. AO (3D) I'W J'--~ AO ~~ ~ ~ (3F) ~ '-+ T(OR) r--- (4E) T 94 _:::,~,JOff) l .,. ;:-= ,. ~ Figure 83. Channel Cycle Time Logic Chan L-E r---- Any I/o Sense or Test Ch A8(D 1) Delayed Chan End-Op Chan A10(D1) (On) ~~ 08.00.27.2 A (3S) l~ T ~~ ~ R 0"8:"00.02. 2 A E or L End-Op L Time B Interrupt I Time E Time II Time Late Not L Time Late Not E Time Late A2(Dl) AO (3H) A1(Dl) A5(Dl) Mult Time Mult Time Error Lite On CE Panel , .: Chan A9(Dl) CP Set 08.00.22.2 \ I 1;- -:. Turn On MST Tgr ~ln.:..:..te::..:.r~lo=ck::......:R=e=5e-,-t_ _ _~ R(2E) 08.00.17.1 Figure 84. Multiple Time Check Circuitry Multiple Cycle Time Error Detection Eight different cycle times and two separate clocks exist in the 7094 II. Only certain cycles can occur simultaneously without causing machine malfunctions. Illegal cycle combinations turn on a multiple time trigger which immediately stops the computer and lights a mult time light on the console CE panel. Test circuitry (Figure 84) monitors the various CPU cycle times once every cycle. The top group of circuits checks for the following illegal combinations: I and E; L and I; Land E; B and I; and Band E. Note that Band L times are allowed to occur simultaneously. II time is allowed to occur Simultaneously with an E or L cycle. II time occurring without either an E or L time is detected at A0 3H • Clock alignment is checked at A0 3I • Every channel A9(DI) pulse should occur Simultaneously with either a CPU AI(DI) or A5(DI) pulse. An absence of these conditions turns on the multiple time trigger. There is no check on multiple channel cycle times. Waveforms and Variable Delay Adjustments The oscilloscope being used must contain probes of the same length. Test the oscilloscope by placing both probes on the same point and noting if there is any difference in the time relationship between the A and B sweeps. The following adjustments should be made in the order presented because in some cases a later adjustment is based on the proper earlier setting. It must be remembered, however, that many of these settings are nominal (starting point) settings and may vary slightly with later adjustments or from system to system. Initial Delay Settings All variable delays shall be set to the nominal values specified on Systems 00.92.01.0 (sheets "1 and 2) before proceeding with the following adjustments. Timings for F lines should be measured at the 1.5 volt level. Timing 95 Odd and Even Clock Drive Pulses Odd and even clock drive pulses should meet all requirements of the waveform shown in Figures 85 and 86 when observed at 03A4C15F (Systems 08.00.44.1, IF). Ref -6v CP Set Pulses set pulses should meet all requirements ()f the waveform shown in Figure 87 when observed at 01AIC12C (Systems 02.15.6l.1,4D) and 01BIC20B (Systems 02.15.61.2,5E). t1 = t2 within 7ns CP Figure 85. Clock Drive Pulse CPU-l CP Set Pulse Width Adjustment Connect the scope probe to the test points indicated below and adjust the CP set pulse width by means of the corresponding V7 delay card. The first two adjustments determine the CP set pulse widths at the A and B gates, respectively. The last width adjustment is for setting the FACT triggers during arithmetic operations and is more critical than most other set pulses. TEST POINT LEVEL 01A1E13E 01B1E24B 01B1F21C -F -F +F WIDTH DLY LOCATION 50 ± 3 ns 01A1E14 (A-H) 60 ± 3 ns 01B1E25 (B-G) 40 ± 3ns 01B1E12 (D-E) SYSTEMS 02.15.61.1,4B 02.15.61.2,4A 02.15.61.2,2B Figure 86. Even and Odd Clock Drive Pulses CP Set Pulse-Clock Pulse Alignment In this section the multiplexor CP set pulse is adjusted with respect to the A7(Dl) CPU-l pulse. This A7(Dl) pulse is already properly aligned to the CPU-2 pulse because of the VB delay 01D2E21 (Systems 02.15.42.1, 3H). 1. Synchronize the scope on "-F A6(D2)" at 01AlB14B (Systems 02.15.71.1, 4H). 2. Connect scope probe A to "-F A7(Dl)" at 01BlC05E (Systems 02.15.70.8, 41). 3. Connect scope probe B to "+ F CP set" at 01BlE24C (Systems 02.15.61.2, 5A). 4. Adjust the variable delay control at 03B3D03 (Systems 08.00.47.1, 4C) so that the fall of the set pulse occurs 10 nanoseconds before the fall of the "-F A7(Dl)" pulse. This 10 nanosecond timing is measured at the point where both pulses cross the F reference level (Figure 88). The optimum multiplexor delay line operation point is determined by running 9M8l. The setting is the midpoint of the error free operating range of 9M81 running at normal voltage as the delay line is varied. After determining the optimum operating point and without disturbing the delay line adjustment, remove and reinstall the delay line knob at 03B3D03 to read zero. 96 1 .5v Ref t1 = t2 within4ns Figure 87. CP Set Pulse Figure 88. CP Set Pulse--Clock Pulse Alignment CPU and Channel Memory Select Alignment Memory selection occurs with different clock pulses (and from different clocks) when initiated during either CPU or channel operations. These two memory selects must be aligned to insure proper operation when memory is being alternately used by both the CPU and channel. 1. Synchronize the scope on "+F A6(Dl)" at 01AlB05G (Systems 02.15.70.7, 41). 2. Connect scope probe A to "+ F A2(Dl)" at 01A4J24E (Systems 03.06.29.4, 3B). 3. Connect scope probe B to "+F A7(Dl)" at 01A4J24G (Systems 03.06.29.4, 3A). 4. Adjust the V7 delay card at 03A4J17 (A-H) (Systems 08.00.40.1, IG) so that the rise of the channel A2(Dl) dlyd pulse crosses the reference line at the same time as the A7(Dl)S pulse, ±3 nanoseconds as shown in Figure 89. Figure 89. Memory Select Pulse Alignment Memory Select and MAR Bus Alignment This section checks that the MAR bus pulses precede the MAR set pulse by at least 20 nanoseconds. This check is made at both the even and odd memories. Figure 90 shows a representative pulse. Note that all of the test points are at the 7302-3 panels. 1. Execute either a TRA (+0020) 77776, 0 or TRA 77777, 0 as indicated below in continuous enter instruction. 2. Synchronize the scope on "-FE time" at 01A2E12P (Systems 08.00.19.3, 3A). This test point is in the CPU. 3. Connect scope probe A to the MAR set point indicated below. 4. Connect scope probe B to the MAR bus points indica ted below. 5. All of the MAR bus lines should precede their respective MAR set pulses by at least 20 nanoseconds. CONDITIONS Instruction MAR set MAR4 MAR 8 MAR 12 MAR 16 EVEN MEMORY TRA 77776,0 01B3C16D (01.11.01.1 -3B) Probe A 01B3C16E (01.11.01.1 -3B) Probe B 01B3C18E (01.11.01.1 -3F) Probe B 01B3C20V (01.11.02.1 -3C) Probe B 01B3C22V (01.11.02.1 -3G) Probe B Figure 90. Memory Select-MAR Bus Alignment ODD MEMORY TRA 77777,0 01C3C16D (01.41.01.1 -3B) Probe A 01C3C16E (01.41.01.1 -3B) Probe B 01C3C18E (01.41.01.1 -3F) Probe B 01C3C20V (01.41.02.1 -3C) Probe B 01C3C22V (01.41.02.1 -3G) Probe B Timing 97 Appendix A: Octal-Decimal Integer Conversion Table 0000 0000 to 0777 to 0511 (Octal) (Decimal) Octal Decimal 10000- 4096 20000 - 8192 30000 - 12288 40000 - 16384 50000 - 20480· 60000 - 24576 70000 - 28672 1000 0512 to 1777 (Octal) to 1023 98 (Decimal) 0 1 2 3 4 5 6 7 0000 0010 0020 0030 0040 0050 0060 0070 0000 0008 0016 0024 0032 0040 0048 0056 0001 0009 0017 0025 0033 0041 0049 0057 0002 0010 0018 0026 0034 0042 0050 0058 0003 0011 0019 0027 0035 0043 0051 0059 0004 0012 0020 0028 0036 0044 0052 0060 0005 0013 0021 0029 0037 0045 0053 0061 0006 0014 0022 0030 0038 0046 0054 0062 0007 0015 0023 0031 0039 0047 0055 0063 0100 0110 0120 0130 0140 0150 0160 0170 0064 0072 0080 0088 0096 0104 0112 0120 0065 0073 0081 0089 0066 0074 0082 0090 00~7 0098 0105 0106 0113 0114 0121 0122 0067 0075 0083 0091 0099 0107 0115 0123 0068 0076 0084 0092 0100 0108 0116 0124 0069 0077 0085 0093 0101 0109 0117 0125 0070 0078 0086 0094 0102 0110 0118 0126 0200 0210 0220 0230 0240 0250 0260 0270 0128 0136 0144 0152 0160 0168 0176 0184 0129 0137 0145 0153 0161 0169 0177 0185 0130 0138 0146 0154 0162 0170 0178 0186 0131 0139 0147 0155 0163 0171 0179 0187 0132 0140 0148 0156 0164 0172 0180 0188 0133 0141 0149 0157 0165 0173 0181 0189 0300 0310 0320 0330 0340 0350 0360 0370 0192 0200 0208 0216 0224 0232 0240 0248 0193 0201 0209 0217 0225 0233 0241 0249 0194 0202 0210 0218 0226 0234 0242 0250 0195 0203 0211 0219 0227 0235 0243 0251 0196 0204 0212 0220 0228 0236 0244 0252 0 1 2 3 1000 1010 1020 1030 1040 1050 1060 1070 0512 0520 0528 0536 0544 0552 0560 0568 0513 0521 0529 0537 0545 05.53 0561 0569 0514 0522 0530 0538 0546 0554 0562 0570 1100 1110 1120 1130 1140 1150 1160 1170 0576 0584 0592 0600 0608 0616 0624 0632 0577 0585 0593 0601 0609 0617 0625 0633 1200 1210 1220 1230 1240 1250 1260 1270 0640 0648 0656 0664 0672 0680 0688 0696 1300 1310 1320 1330 1340 1350 1360 1370 0704 0712 0720 0728 0736 0744 0752 0760 0 1 2 3 4 5 6 7 0400 0410 0420 0430 0440 0450 0460 0470 0256 0264 0272 0280 0288 0296 0304 0312 0257 0265 0273 0281 0289 0297 0305 0313 0258 0266 0274 0282 0290 0298 0306 0314 0259 0267 0275 0283 0291 0299 0307 0315 0260 0268 0276 0284 0292 0300 0308 0316 0261 0269 0277 0285 0293 0301 0309 0317 0262 0270 0278 0286 0294 0302 0310 0318 0263 0271 0279 0287 0295 0303 0311 0319 0071 0079 0087 0095 0103 0111 0119 0127 0500 0510 0520 0530 0540 0550 0560 0570 0320 0328 0336 0344 0352 0360 0368 0376 0321 0329 0337 0345 0353 0361 0369 0377 0322 0330 0338 0346 0354 0362 0370 0378 0323 0331 0339 0347 0355 0363 0371 0379 0324 0332 0340 0348 0356 0364 0372 0380 0325 0333 0341 0349 0357 0365 0373 0381 0326 0334 0342 0350 0358 0366 0374 0382 0327 0335 0343 0351 0359 0367 0375 0383 0134 0142 0150 0158 0166 0174 0182 0190 0135 0143 0151 0159 0167 0175 0183 0191 0600 0610 0620 0630 0640 0650 0660 0670 0384 0392 0400 0408 0416 0424 0432 0440 0385 0393 0401 0409 0417 0425 0433 0441 0386 0394 0402 0410 0418 0426 0434 0442 0387 0395 0403 0411 0419 0427 0435 0443 0388 0396 0404 0412 0420 0428 0436 0444 0389 0397 0405 0413 0421 0429 0437 0445 0390 0398 0406 0414 0422 0430 0438 0446 0391 0399 0407 0415 0423 0431 0439 0447 0197 0205 0213 0221 0229 0237 0245 0253 0198 0206 0214 0222 0230 0238 0246 0254 0199 0207 0215 0223 0231 0239 0247 0255 0700 0710 0720 0730 0740 0750 0760 0770 0448 0456 0464 0472 0480 0488 0496 0504 0449 0457 0465 0473 0481 0489 0497 0505 0450 0458 0466 0474 0482 0490 0498 0506 0451 0459 0467 0475 0483 0491 0499 0507 0452 0460 0468 0476 0484 0492 0500 0508 0453 0461 0469 0477 0485 0493 0501 0509 0454 0462 0470 0478 0486 0494 0502 0510 0455 0463 0471 0479 0487 0495 0503 0511 4 5 6 7 0 1 2 3 4 5 6 7 0515 0523 0531 0539 0547 0555 0563 0571 0516 0524 0532 0540 0548 0556 0564 0572 0517 0525 0533 0541 0549 0557 0565 0573 0518 0526 0534 0542 0550 0558 0566 0574 0519 0527 0535 0543 0551 0559 0567 0575 1400 1410 1420 1430 1440 1450 1460 1470 0768 0776 0784 0792 0800 0808 0816 0824 0769 0777 0785 0793 0801 0809 0817 0825 0770 0778 0786 0794 0802 0810 0818 0826 0771 0779 0787 0795 0803 0811 0819 0827 077~ 0780 0788 0796 0804 0812 0820 0828 0773 0781 0789 0797 0805 0813 0821 0829 0774 0782 0790 0798 0806 0814 0822 0830 0775 0783 0791 0799 0807 0815 0823 0831 0578 0586 0594 0602 0610 0618 0626 0634 0579 0587 0595 0603 0611 0619 0627 0635 0580 0588 0596 0604 0612 0620 0628 0636 0581 0589 0597 0605 0613 0621 0629 0637 0582 0590 0598 0606 0614 0622 0630 0638 0583 0591 0599 0607 0615 0623 0631 0639 1500 1510 1520 1530 1540 1550 1560 1570 0832 0840 0848 0856 0864 0872 0880 0888 0833 0841 0849 0857 0865 0873 0881 0889 0834 0842 0850 0858 0866 0874 0882 0890 0835 0843 0851 0859 0867 0875 0883 0891 0836 0844 0852 0860 0868 0876 0884 0892 0837 0845 0853 0861 0869 0877 0885 0893 0838 0846 0854 0862 0870 0878 0886 0894 0839 0847 0855 0863 0871 0879 0887 0895 0641 0649 0657 0665 0673 0681 0689 0697 0642 0650 0658 0666 0674 0682 0690 0698 0643 0651 0659 0667 0675 0683 0691 0699 0644 0652 0660 0668 0676 0684 0692 0700 0645 0653 0661 0669 0677 0685 0693 0701 0646 0654 0662 0670 0678 0686 0694 0702 0647 0655 0663 0671 0679 0687 0695 0703 1600 1610 1620 1630 1640 1650 1660 1670 0896 0904 0912 0920 0928 0936 0944 0952 0897 0905 0913 0921 0929 0937 0945 0953 0898 0906 0914 0922 0930 0938 0946 0954 0899 0907 0915 0923 0931 0939 0947 0955 0900 0908 0916 0924 0932 0940 0948 0956 0901 0909 0917 0925 0933 0941 0949 0957 0902 0910 0918 0926 0934 0942 0950 0958 0903 0911 0919 0927 0935 0943 0951 0959 0705 0713 0721 0729 0737 0745 0753 0761 0706 0714 0722 0730 0738 0746 0754 0762 0707 0715 0723 0731 0739 0747 0755 0763 0708 0716 0724 0732 0740 0748 0756 0764 0709 0717 0725 0733 0741 0749 0757 0765 0710 0718 0726 0734 0742 0750 0758 0766 0711 0719 0727 0735 0743 0751 0759 0767 1700 1710 1720 1730 1740 1750 1760 1770 0960 0968 0976 0984 0992 1000 1008 1016 0961 0969 0977 0985 0993 1001 1009 1017 0962 0970 0978 0986 0994 1002 1010 1018 0963 0971 0979 0987 0995 1003 1011 1019 0964 0972 0980 0988 0996 1004 1012 1020 0965 0973 0981 0989 0997 1005 1013 1021 0966 0974 0982 0990 0998 1006 1014 1022 0967 0975 0983 0991 0999 1007 1015 1023 Octal-Decimal Integer Conversion Table (Continued) 2000 2010 2020 2030 2040 2050 2060 2070 0 1 2 3 4 5 6 7 1024 1032 1040 1048 1056 1064 1072 1080 1025 1033 1041 1049 1057 1065 1073 1081 1026 1034 1042 1050 1058 1066 1074 1082 1027 1035 1043 1051 1059 1067 1075 1083 1028 1036 1044 1052 1060 1068 1076 1084 1029 1037 1045 1053 1061 1069 1077 1085 1030 1038 1046 1054 1062 1070 1078 1086 1031 1039 1047 1055 1063 1071 1079 1087 I 2400 2410 2420 2430 2440 2450 2460 2470 0 1 2 3 4 5 6 7 1280 1288 1296 1304 1312 1320 1328 1336 1281 1289 1297 1305 1313 1321 1329 1337 1282 1290 1298 1306 1314 1322 1330 1338 1283 1291 1299 1307 1315 1323 1331 1339 1284 1292 1300 1308 1316 1324 1332 1340 1285 1293 1301 1309 1317 1325 1333 1341 1286 1294 1302 1310 1318 1326 1334 1342 1287 1295 1303 1311 1319 1327 1335 1343 1349 1357 1365 1373 1381 1389 1397 1405 1350 1358· 1366 1374 1382 1390 1398 1406 1351 1359 1367 1375 1383 1391 1399 1407 2100 2110 2120 2130 2140 2150 2160 2170 1088 1096 1104 1112 1120 1128 1136 1144 1089 1097 1105 1113 1121 1129 1137 1145 1090 1098 1106 1114 1122 1130 1138 1146 1091 1099 1107 1115 1123 1131 1139 1147 1092 1100 1108 1116 1124 1132 1140 1148 1093 1101 1109 1117 1125 1133 1141 1149 1094 1102 1110 1118 1126 1134 1142 1150 1095 1103 1111 1119 1127 1135 1143 li51 2500 2510 2520 2530 2540 2550 2560 2570 1344 1352 1360 1368 1376 1384 1392 1400 1345 1353 1361 1369 1377 1385 1393 1401 1346 1354 1362 1370 1378 1386 1394 1402 1347 1355 1363 1371 1379 1387 1395 1403 1348 1356 1364 1372 1380 1388 1396 1404 2200 2210 2220 2230 2240 2250 2260 2270 1152 1160 1168 1176 1184 1192 1200 1208 1153 1161 1169 1177 1185 1193 1201 1209 1154 1162 1170 1178 1186 1194 1202 1210 1155 1163 1171 1179 1187 1195 1203 1211 1156 1164 1172 1180 1188 1196 1204 1212 1157 1165 1173 1181 1189 1197 1205 1213 1158 1166 1174 1182 1190 1198 1206 1214 1159 1167 1175 1183 1191 1199 1207 1215 2600 2610 2620 2630 2640 2650 2660 2670 1408 1416 1424 1432 1440 1448 1456 1464 1409 1417 1425 1433 1441 1449 1457 1465 1410 1418 1426 1434 1442 1450 1458 1466 1411 1419 1427 1435 1443 1451 1459 1467 1412 1420 1428 1436 1444 1452 1460 1468 1413 1421 1429 1437 1445 1453 1461 1469 1414 1422 1430 1438 1446 1454 1462 1470 1415 1423 1431 1439 1447 1455 1463 1471 2300 2310 2320 2330 2340 2350 2360 2370 1216 1224 1232 1240 1248 1256 1264 1272 1217 1225 1233 1241 1249 1257 1265 1273 1218 1226 1234 1242 1250 1258 1266 1274 1219 1227 1235 1243 1251 1259 1'267 1275 1220 1228 1236 1244 1252 1260 1268 1276 1221 1229 1237 1245 1253 1261 1269 1277 1222 1230 1238 1246 1254 1262 1270 1278 1223 1231 1239 1247 1255 1263 1271 1279 2700 2710 2720 2730 2740 2750 2760 2770 1472 1480 1488 1496 1504 1512 1520 1528 1473 1481 1489 1497 1505 1513 1521 1529 1474 1482 1490 1498 1506 1514 1522 1530 1475 1483 1491 1499 1507 1515 1523 1531 1476 1484 1492 1500 1508 1516 1524 1532 1477 1485 1493 1501 1509 1517 1525 1533 1478 1486 1494 1502 1510 1518 1526 1534 1479 1487 1495 1503 1511 1519 1527 1535 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 3000 3010 3020 3030 3040 3050 3060 3070 1536 1544 1552 1560 1568 1576 1584 1592 1537 1545 1553 1561 1569 1577 1585 1593 1538 1546 1554 1562 1570 1578 1586 1594 1539 1547 1555 1563 1571 1579 1587 1595 1540 1548 1556 1564 1572 1580 1588 1596 1541 1549 1557 1565 1573 1581 1589 1597 1542 1550 1558 1566 1574 1582 1590 1598 1543 1551 1559 1567 1575 1583 1591 1599 3400 3410 3420 3430 3440 3450 3460 3470 1792 1800 1808 1816 1824 1832 1840 1848 1793 1801 1809 1817 1825 1833 1841 1849 1794 1802 1810 1818 1826 1834 1842 1850 1795 1803 1811 1819 1827 1835 1843 1851 1796 1"804 1812 1820 1828 1836 1844 1852 1797 1805 1813 1821 1829 1837 1845 1853 1798 1806 1814 1822 1830 1838 1846 1854 1799 1807 1815 1823 1831 1839 1847 1855 3100 3110 3120 3130 3140 3150 3160 3170 1600 1608 1616 1624 1632 1640 1648 1656 1601 1609 1617 1625 1633 1641 1649 1657 1602 1610 1618 1626 1634 1642 1650 1658 1603 1611 1619 1627 1635 1643 1651 1659 1604 1612 1620 1628 1636 1644 1652 1660 1605 1613 1621 1629 1637 1645 1653 1661 1606 1614 1622 1630 1638 1646 1654 1662 1607 1615 1623 1631 1639 1647 1655 1663 3500 3510 3520 3530 3540 3550 3560 3570 1856 1864 1872 1880 1888 1896 1904 1912 1857 1865 1873 1881 1889 1897 1905 1913 1858 1866 1874 1882 1890 1898 1906 1914 1859 1867 1875 1883 1891 1899 1907 1915 1860 1868 1876 1884 1892 1900 1908 1916 1861 1869 1877 1885 1893 1901 1909 1917 1862 1870 1878 1886 1894 1902 1910 1918 1863 1871 1879 1887 1895 1903 1911 1919 3200 3210 3220 3230 3240 3250 3260 3270 1664 1672 1680 1688 1696 1704 1712 1720 1665 1673 1681 1689 1697 1705 1713 1721 1666 1674 1682 1690 1698 1706 1714 1722 1667 1675 1683 1691 1699 1707 1715 1723 1668 1676 1684 1692 1700 1708 1716 1724 1669 1677 1685 1693 1701 1709 1717 1725 1670 1678 1686 1694 1702 1710 1718 1726 1671 1679 1687 1695 1703 1711 1719 1727 3600 3610 3620 3630 3640 3650 3660 3670 1920 1928 1936 1944 1952 1960 1968 1976 1921 1929 1937 1945 1953 1961 1969 1977 1922 1930 1938 1946 1954 1962 1970 1978 1923 1931 1939 1947 1955 1963 1971 1979 1924 1932 1940 1948 1956 1964 1972 1980 1925 1933 1941 1949 1957 1965 1973 1981 1926 1934 1942 1950 1958 1966 1974 1982 1927 1935 1943 1951 1959 1967 1975 1983 3300 3310 3320 3330 3340 3350 3360 3370 1728 1736 1744 1752 1760 1768 1776 1784 1729 1737 1745 1753 1761 1769 1777 1785 1730 1738 1746 1754 1762 1770 1778 1786 1731 1739 1747 1755 1763 1771 1779 1787 1732 1740 1748 1756 1764 1772 1780 1788 1733 1741 1749 1757 1765 1773 1781 1789 1734 1742 1750 1758 1766 1774 1782 1790 1735 1743 1751 1759 1767 1775 1783 1791 3700 3710 3720 3730 3740 3750 3760 3770 1984 1992 2000 2008 2016 2024 2032 2040 1985 1993 2001 2009 2017 2025 2033 2041 1986 1994 2002 2010 2018 2026 2034 2042 1987 1995 2003 2011 2019 2027 2035 2043 1988 1996 2004 2012 2020 2028 2036 2044 1989 1997 2005 2013 2021 2029 2037 2045 1990 1998 2006 2014 2022 2030 2038 2046 1991 1999 2007 2015 2023 2031 2039 2047 2000 1024 to to 2777 1535 (Octal) (Decimal) Octal Decimal 10000 - .4096 20000 - 8192 30000 - 12288 .40000 - 16384 50000 - 20480 60000 - 24576 70000 - 28672 3000 1536 to to 3777 2047 (Octal) (Decimal) Octal-Decimal Integer Conversion 99 Octal-Decimal Integer Conversion Table (Continued) .(000 to ,(777 20.(8 to (Odal) (Decimal) 2559 Octal Decimal 10000 - 4096 20000 - 8192 30000 - 12288 40000 - 16384 50000 - 20480 60000 - 24576 70000 - 28672 2560 to 5777 to 3071 (Odal) (Decimal) 100 5 6 7 2305 2306 2307 2313 2314 2315 2~21 2322 2323 2329 2330 2331 2337 2338 2339 2345 2346 2347 2353 2354 2355 2361 2362 2363 2308 2316 2324 2332 2340 2348 2356 2364 2309 2317 2325 2333 2341 2349 2357 2365 2310 2318 2326 2334 2342 2350 2358 2366 2311 2319 2327 2335 2343 2351 2359 2367 2369 2377 2385 2393 2401 2409 2417 2425 2370 2378 2386 2394 2402 2410 2418 2426 2371 2379 2387 2395 2403 2411 2419 2427 2372 2380 2388 2396 2404 2412 2420 2428 2373 2381 2389 2397 2405 2413 2421 2429 2374 2382 2390 2398 2406 2414 2422 2430 2375 2383 2391 2399 2407 2415 2423 2431 2432 2440 2448 2456 2464 2472 2480 2488 2433 2441 2449 2457 2465 2473 2481 2489 2434 2442 2450 2458 2466 2474 2482 2490 2435 2443 2451 2459 2467 2475 2483 2491 2436 2444 2452 2460 2468 2476 2484 2492 2437 2445 2453 2461 2469 2477 2485 2493 2438 2446 2454 2462 2470 2478 2486 2494 2439 2447 2455 2463 2471 2479 2487 2495 2496 2504 2512 2520 2528 2536 2544 2552 2497 2505 2513 2521 2529 2537 2545 2553 2498 2506 2514 2522 2530 2538 2546 2554 2499 2507 2515 2523 2531 2539 2547 2555 2500 2508 2516 2524 2532 2540 2548 2556 2501 2509 2517 .2525 2533 2541 2549 2557 2502 2510 2518 2526 2534 2542 2550 2558 2503 2511 2519 2527 2535 2543 2551 2559 0 1 2 3 4 5 6 7 1 2 3 4 5 6 7 4000 4010 4020 4030 4040 4050 4060 4070 2048 2056 2064 2072 2080 2088 2096 2104 2049 2057 2065 2073 2081 2089 2097 2105 2050 2058 2066 2074 2082 2090 2098 2106 2051 2059 2067 2075 2083 2091 2099 2107 2052 2060 2068 2076 2084 2092 2100 2108 2053 2061 2069 2077 2085 2093 2101 2109 2054 2062 2070 2078 2086 2094 2102 2110 2055 2063 2071 2079 2087 2095 2103 2111 4400 4410 4420 4430 4440 4450 4460 4470 2304 2312 2320 2328 2336 2344 2352 2360 4100 4110 4120 4130 4140 4150 4160 4170 2112 2120 2128 2136 2144 2152 2160 2168 2113 2121 2129 2137 2145 2153 2161 2169 2114 2122 2130 2138 2146 2154 2162 2170 2115 2123 2131 2139 2147 2155 2163 2171 2116 2124 2132 2140 2148 2156 2164 2172 2117 2125 2133 2141 2149 2157 2165 2173 2118 2126 2134 2142 2150 2158 2166 2174 2119 2127 2135 2143 2151 2159 2167 2175 4500 4510 4520 4530 4540 4550 4560 4570 2368 2376 2384 2392 2400 2408 2416 2424 4200 4210 4220 4230 4240 4250 4260 4270 2176 2184 2192 2200 2208 2216 2224 2232 2177 2185 2193 2201 2209 2217 2225 2233 2178 2186 2194 2202 2210 2218 2226 2234 2179 2187 219.5 2203 2211 2219 2227 2235 2180 2188 2196 2204 2212 2220 2228 2236 2181 2189 2197 2205 2213 2221 2229 2237 2182 2190 2198 2206 2214 2222 2230 2238 2183 2191 2199 2207 2215 2223 2231 2239 4600 4610 4620 4630 4640 4650 4660 4670 4300 4310 4320 4330 4340 4350 4360 4370 2240 2248 2256 2264 2272 2280 2288 2296 2241 2249 2257 2265 2273 2281 2289 2297 2242 2250 2258 2266 2274 2282 2290 2298 2243 2251 2259 2267 2275 2283 2291 2299 2244 2252 2260 2268 2276 2284 2292 2300 2245 2253 2261 2269 2277 2285 2293 2301 2246 2254 2262 2270 2278 2286 2294 2302 2247 2255 2263 2271 2279 2287 2295 2303 4700 4710 4720 4730 4740 4750 4760 4770 0 5000 4 0 0 1 2 3 4 5 6 7 -_. 1 2 3 5000 5010 5020 5030 5040 5050 5060 5070 2560 2568 2576 2584 2592 2600 2608 2616 2561 2569 2577 2585 2593 2601 2609 2617 2562 2570 2578 2586 2594 2602 2610 2618 2563 2571 2579 2587 2595 2603 2611 2619 2564 2572 2580 2588 2596 2604 2612 2620 2565 2573 2581 2589 2597 2605 2613 2621 2566 2574 2582 2590 2598 2606 2614 2622 2567 2575 2583 2591 2599 2607 2615 2623 5400 5410 5420 5430 5440 5450 5460 5470 2816 2824 2832 2840 2848 2856 2864 2872 2817 2825 2833 2841 2849 2857 2865 2873 2818 2826 2834 2842 2850 2858 2866 2874 2819 2827 2835 2843 2851 2859 2867 2875 2820 2828 2836 2844 2852 2860 2868 2876 2821 2829 2837 2845 2853 2861 2869 2877 2822 2830 2838 2846 2854 2862 2870 2878 2823 2831 2839 2847 2855 2863 2871 2879 5100 5110 5120 5130 5140 5150 5160 5170 2624 2632 2640 2648 2656 2664 2672 2680 2625 2633 2641 2649 2657 2665 2673 2681 2626 2634 2642 2650 2658 2666 2674 2682 2627 2635 2643 2651 2659 2667 2675 2683 2628 2636 2644 2652 2660 2668 2676 2684 2629 2637 2645 2653 26tH 2669 2677 2685 2630 2638 2646 2654 2662 2670 2678 2686 2631 2639 2647 2655 2663 2671 2679 2687 5500 5510 5520 5530 5540 5550 5560 5570 2880 2888 2896 2904 2912 2920 2928 2936 2881 2889 2897 2905 2913 2921 2929 2937 2882 2890 2898 2906 2914 2922 2930 2938 2883 2891 2899 2907 2915 2923 2931 2939 2884 2892 2900 2908 2916 2924 2932 2940 2885 2893 2901 2909 2917 2925 2933 2941 2886 2894 2902 2910 2918 2926 2934 2942 2887 2895 2903 2911 2919 2927 2935 2943 5200 5210 5220 5230 5240 5250 5260 5270 2688 2696 2704 2712 2720 2728 2736 2744 2689 2697 2705 2713 2721 2729 2737 2745 2690 2698 2706 2714 2722 2730 2738 27'6 2691 2699 2707 2715 2723 2731 2739 2747 2692 2700 2708 2716 2724 2732 2740 2748 2693 2701 2709 2717 2725 2733 2741 2749 2694 2702 2710 2718 2726 2734 2742 2750 2695 2703 2711 2719 2727 2735 2743 2751 5600 5610 5620 5630 5640 5650 5660 5670 2944 2952 2960 2968 2976 2984 2992 3000 2945 2953 296.1 2969 2977 2985 2993 3001 2946 2954 2962 2970 2978 2986 2994 3002 2947 2955 2963 2971 2979 2987 2995 3003 2948 2956 2964 2972 2980 2988 2996 3004 2949 2957 2965 2973 2981 2989 2997 3005 2950 2958 2966 2974 2982 2990 2998 3006 2951 2959 2967 2975 2983 2991 2999 3007 5300 5310 5320 5330 5340 5350 5360 5370 2752 2760 2768 2776 2':84 2792 2800 2808 2753 2761 2769 2777 2785 2793 2801 2809 2754 2762 2770 2778 2786 2794 2802 2810 2755 2763 2771 2779 2787 2795 2803 2811 2756 2764 2772 2780 2788 2796 2804 2812 2757 2765 2773 2781 2789 2797 2805 2813 2758 2766 2774 2782 2790 2798 2806 2814 2759 2767 2775 2783 2791 2799 2807 2815 5700 5710 5720 5730 5740 5750 5760 5770 3008 3016 3024 3032 3040 3048 3056 3064 3009 3017 3025 3033 3041 3049 3057 3065 3010 3018 3026 3034 3042 3050 3058 3066 3011 3019 3027 3035 3043 3051 3059 3067 3012 3020 3028 3036 3044 3052 3060 '3068 3013 3021 3029 3037 3045 3053 3061 3069 3014 3022 3030 3038 3046 3054 3062 3070 3015 3023 3031 3039 3047 3055 3063 3071 Octal-Decimal Integer Conversion Table (Continued) 4 5 6 7 3331 3339 3347 3355 3363 3371 3379 3387 3332 3340 3348 3356 3364 3372 3380 3388 3333 3341 3349 3357 3365 3373 3381 3389 3334 3342 3350 3358 3366 3374 3382 3390 3335 3343 3351 3359 3367 3375 3383 3391 3394 3402 3410 3418 3426 3434 3442 3450 3395 3403 3411 3419 3427 3435 3443 3451 3396 3404 3412 3420 3428 3436 3444 3452 3397 3405 3413 3421 3429 3437 3445 3453 3398 3406 3414 3422 3430 3438 3446 3454 3399 3407 3415 3423 3431 3439 3447 3455 3457 3465 3473 3481 3489 3497 3505 3513 3458 3466 3474 3482 3490 3498 3506 3514 3459 3467 3475 3483 3491 3499 3507 3515 3460 3468 3476 3484 3492 3500 3508 3516 3461 3469 3477 3485 3493 3501 3509 3517 3462 3470 3478 3486 3494 3502 3510 3518 3463 3471 3479 3487 3495 3503 3511 3519 3520 3528 3536 3544 3552 3560 3568 3576 3521 3529 3537 3545 3553 3561 3569 3577 3522 3530 3538 3546 3554 3562 3570 3578 3523 3531 3539 3547 3555 3563 3571 3579 3524 3532 3540 3548 3556 3564 3572 3580 3525 3533 3541 3549 3557 3565 3573 3581 3526 3534 3542 3550 3558 3566 3574 3582 3527 3535 3543 3551 3559 3567 3575 3583 0 1 2 3 4 5 6 7 3840 3848 3856 3864 3872 3880 3888 3896 3841 3849 3857 3865 3873 3881 3889 3897 3842 3850 3858 3866 3874 3882 3890 3898 3843 3851 3859 3867 3875 3883 3891 3899 3844 3852 3860 3868 3876 3884 3892 3900 3845 3853' 3861 3869 3877 3885 3893 3901 3846 3854 3862 3870 3878 3886 3894 3902 3847 3855 3863 3871 3879 3887 3895 3903 0 1 2 0 1 2 3 4 5 6 7 6000 6010 6020 6030 6040 6050 6060 6070 3072 3080 3088 3096 3104 3112 3j20 3128 3073 3081 3089 3097 3105 3113 3121 3129 3074 3082 3090 3098 3106 3114 3122 3130 3075 3083 3091 3099 3107 3115 3123 3131 3076 3084 3092 3100 3108 3116 3124 3132 3077 3085 3093 3101 3109 3117 3125 3133 3078 3086 3094 3102 3110 3118 3126 3134 3079 3087 3095 3103 3111 3119 3127 3135 6400 6410 6420 6430 6440 6450 6460 6470 3328 3336 3344 3352 3360 3368 3376 3384 3329 3337 3345 3353 3361 3369 3377 3385 3330 3338 3346 3354 3362 3370 3378 3386 6100 6110 6120 6130 6140 6150 6160 6170 3136 3144 3152 3160 3168 3176 3184 3192 3137 3145 3153 3161 3169 3177 3185 3193 3138 3146 3154 3162 3170 3178 3186 3194 3139 3147 3155 3163 3171 3179 3187 3195 3140 3148 3156 3164 3172 3180 3188 3196 3141 3149 3157 3165 3173 3181 3189 3197 3142 3150 3158 3166 3174 3182 3190 3198 3143 3151 3159 3167 3175 3183 3191 3199 6500 6510 6520 6530 6540 6550 6560 6570 3392 3400 3408 3416 3424 3432 3440 3448 3393 3401 3409 3417 3425 3433 3441 3449 6200 6210 6220 6230 6240 6250 6260 6270 3200 3208 3216 3224 3232 3240 3248 3256 3201 3209 3217 3225 3233 3241 3249 3257 3202 3210 3218 3226 3234 3242 3250 3258 3203 3211 3219 3227 3235 3243 3251 3259 3204 3212 3220 3228 3236 3244 3252 3260 3205 3213 3221 3229 3237 3245 3253 3261 3206 3214 3222 3230 3238 3246 3254 326~ 3207 3215 3223 3231 3239 3247 3255 3263 6600 6610 6620 6630 6640 6650 6660 6670 3456 3464 3472 3480 3488 3496 3504 3512 6300 6310 6320 6330 6340 6350 6360 6370 3264 3272 3280 3288 3296 3304 3312 3320 3265 3273 3281 3289 3297 3305 3313 3321 3266 3274 3282 3290 3298 3306 3314 3322 3267 3275 3283 3291 3299 3307 3315 3323 3268 3276 3284 3292 3300 3308 3316 3324 3269 3277 3285 3293 3301 3309 3317 3325 3270 3278 3286 3294 3302 3310 3318 3326 3271 3279 3287 3295 3303 3311 3319 3327 6700 6710 6720 6730 6740 6750 6760 6770 0 1 2 3 4 5 6 7 7000 3584 3585 3586 3587 3588 3589 3590 3591 3 7'020 7030 7040 7050 7060 7070 3600 3608 3616 3624 3632 3640 3601 3609 3617 3625 3633 3641 3602 3610 3618 3626 3634 3642 3603 3611 3619 3627 3635 3643 3604 3612 3620 3628 3636 3644 3605 3613 3621 3629 3637 3645 3606 3614 3622 3630 3638 3646 3607 3615 3623 3631 3639 3647 7400 7410 7420 7430 7440 7450 7460 7470 7100 7110 7120 7130 7140 7150 7160 7170 3648 3656 3664 3672 3680 3688 3696 3704 3649 3657 3665 3673 3681 3689 3697 3705 3650 3658 3666 3674 3682 3690 3698 3706 3651 3659 3667 3675 3683 3691 3699 3707 3652 3660 3668 3676 3684 3692 3700 3708 3653 3661 3669 3677 3685 3693 3701 3709 3654 3662 3670 3678 3686 3694 3702 3710 3655 3663 3671 3679 3687 3695 3703 3711 7500 7510 7520 7530 7540 7550 7560 7570 3904 3912 3920 3928 3936 3944 3952 3960 3905 3913 3921 3929 3937 3945 3953 3961 3906 3914 3922 3930 3938 3946 3954 3962 3907 3915 3923 3931 3939 3947 3955 3963 3908 3916 3924 3932 3940 3948 3956 3964 3909 3917 3925 3933 3941 3949 3957 3965 3910 3918 3926 3934 3942 3950 3958 3966 3911 3919 3927 3935 3943 3951 3959 3967 7200 7210 7220 7230 7240 7250 7260 7270 3712 3720 3728 3736 3744 3752 3760 3768 3713 3721 3729 3737 3745 3753 3761 3769 3714 3722 3730 3738 3746 3754 3762 3770 3715 3723 3731 3739 3747 3755 3763 3771 3716 3724 3732 3740 3748 3756 3764 3772 3717 3725 3733 3741 3749 3757 3765 3773 3718 3726 3734 3742 3750 3758 3766 3774 3719 3727 3735 3743 3751 3759 3767 3775 7600 7610 7620 7630 7640 7650 7660 7670 3968 3976 3984 3992 4000 4008 4016 4024 3969 3977 3985 3993 4001 4009 4017 4025 3970 3978 3986 3994 4002 4010 4018 4026 3971 3979 3987 3995 4003 4011 4019 4027 3972 3980 3988 3996 4004 4012 4020 4028 3973 398r 3989 3997 4005 4013 4021 4029 3974 3982 3990 3998 4006 4014 4022 4030 3975 3983 3991 3999 4007 4015 4023 4031 7300 7310 7320 7330 7340 7350 7360 7370 3776 3784 3792 3800 38{)8 3816 3824 3832 3777 37-85 3793 3801 3809 3817 3825 3833 3778 3786 3794 3802 3810 3818 3826 3834 3779 3787 3795 3803 3811 3819 3827 3835 3780 3788 3796 3804 3812 3820 3828 3836 3781 3789 3797 3805 3813 3821 3829 3837 3782 3790 3798 3806 3814 3822 3830 3838 3783 3791 3799 3807 3815 3823 3831 3839 7700 7710 7720 7730 7740 7750 7760 7770 4032 4040 4048 4056 4064 4072 4080 4088 4033 4041 4049 4057 4065 4073 4081 4089 4034 4042 4050 4058 4066 4074 4082 4090 4035 4043 4051 4059 4067 4075 4083 4091 4036 4044 4052 4060 4068 4076 4084 4092 4037 4045 4053 4061 4069 4077 4085 4093 4038 4046 4054 4062 4070 4078 4086 4094 4039 4047 4055 4063 4071 4079 4087 1.010 3592 3593 3594 3595 3596 3597 3598 3599 6000 3072 to to 6777 3583 (Octal) (Decimal) Octa I Decima I 10000 - 4096 20000 - 8192 30000 - 12288 40000 - 16384 50000 - 20480 60000 - 24576 70000 - 28672 7000 3584 to to 7777 "095 (Odal) (Decimal) 40~5 Octal-Decimal Integer Conversion 101 Appendix B: Octal-Decimal Fraction Conversion Table OCTAL DEC. OCTAL DEC. OCTAL DEC. OCTAL DEC. .000 .001 .002 .003 .004 .005 .006 .007 .010 .011 .012 .013 .014 .015 .016 .017 .020 .021 .022 .023 .024 .025 .026 .027 .030 .031 .032 .033 .034 .\)35 .036 .037 .040 .041 .042 .043 .044 .045 .046 .047 .050 .051 .052 .053 .054 .055 .056 .057 .060 .061 .062 .063 .064 .065 .066 .067 .070 .071 .072 .073 .074 .075 .076 .077 .000000 .001953 .003906 .005859 .007812 .009765 .011718 .013671 .015625 .017578 .019531 .021484 .023437 .025390 .027343 .029296 .031250 .033203 .035156 .037109 .039062 .041015 .042968 .044921 .046875 .048828 .050781 .052734 .054687 .056640 .058593 .060546 .062500 .064453 .066406 .068359 .070312 .072265 .074218 .076171 .078125 .080078 .082031 .083984 .085937 .087890 .089843 .091796 .093750 .095703 .097656 .099609 .101562 .103515 .105468 .107421 .109375 .111328 .113281 .115234 .117187 .119140 .121093 .123046 .100 .101 .102 .103 .104 .105 .106 .107 .110 .111 .112 .113 .114 .115 .116 .117 .120 .121 .122 .123 .124 .125 .126 .127 .130 .131 .132 .133 .134 .135 .136 .137 .140 .141 .142 .143 .144 .145 .146 .147 .150 .151 .152 .153 .154 .155 .156 .157 .160 .161 .162 .163 .164 .165 .166 .167 .170 .171 .172 .173 .174 .175 .176 .177 .125000 .126953 .128906 .130859 .132812 .134765 .136718 .138671 .140625 .142578 .144531 .146484 .148437 .150390 .152343 .154296 .156250 .158203 .160156 .162109 .164062 .166015 .167968 .169921 .171875 .173828 .175781 .177734 .179687 .181640 .183593 .185546 .187500 .189453 .191406 .193359 .195312 .197265 .199218 .201171 .203125 .205078 .207031 .208984 .210937 .212890 .214843 .216796 .218750 .220703 .222656 .224609 .226562 .228515 .230468 .232421 .234375 .236328 .238281 .240234 .242187 .244140 .246093 .248046 .200 .201 .202 .203 .204 .205 .206 .207 .210 .211 .212 .213 .214 .215 .216 .217 .220 .221 .222 .223 .224 .225 .226 .227 .230 .231 .232 .233 .234 .235 .236 .237 .240 .241 .242 .243 .244 .245 .246 .247 .250 .251 .252 .253 .254 .255 .256 .257 .260 .261 .262 .263 .264 .265 .266 .267 .270 .271 .272 .273 .274 .275 .276 .277 .250000 .251953 .253906 .255859 .257812 .259765 .261718 .263671 .265625 .267578 .269531 .271484 .273437 .275390 .277343 .279296 .281250 .283203 .285156 .287109 .289062 .291015 .292968 .294921 .296875 .298828 .300781 .302734 .304687 .306640 .308593 .310546 .312500 .314453 .316406 .318359 .320312 .322265 .324218 .326171 .328125 .330078 .332031 .333984 .335937 .337890 .339843 .341796 .343750 .345703 .300 .301 .302 .303 .304 .305 .306 .307 .310 .311 .312 .313 .314 .315 .316 .317 .320 .321 .322 .323 .324 .325 .326 .327 .330 .331 .332 .333 .334 .335 .336 .337 .340 .341 .342 .343 .344 .345 .346 .347 .350 .351 .352 .353 .354 .355 .356 .357 .360 .361 .362 .363 .364 .365 .366 .367 .370 .371 .372 .373 .374 .375 .376 .377 .3.75000 .376953 .378906 .380859 .382812 .384765 .386718 .388671 .390625 .392578 .394531 .396484 .398437 .400390 .402343 .404296 .406250 .408203 .410156 .412109 .414062 .416015 .417968 .419921 .421875 .423828 .426781 .427734 .429687 .431640 .433593 .435546 .437500 .439453 .441406 .443359 .445312 .447265 .449218 .451171 .453125 .455078 .457031 .458984 .460937 .462890 .464843 .466796 .468750 .-470703 .472656 .474609 .476562 .478515 .480468 .482421 .484375 .486328 .488281 .490234 .492187 .494140 .496093 .498046 102 ~347656 .349609 .351562 .353515 .355468 .357421 .359375 .361328 .363281 .365234 .367187 .369140 .371093 .373046 Octal-Decimal Fraction Conversion Table (Continued) OCTAL DEC. OCTAL DEC. OCTAL DEC. OCTAL DEC. .000000 .000001 .000002 .000003 .000004 .000005 .000006 .000007 .000010 .000011 .000012 .000013 .000014 .000015 .000016 .000017 .000020 .000021 .000022 .000023 .000024 .000025 .000026 .000027 .000030 .000031 .000032 .000033 .000034 .000035 .000036 .000037 .000040 .000041 .000042 .000043 .000044 .000045 .000046 .000047 .000050 .000051 .000052 .000053 .000054 .000055 .000056 .000057 .000060 .000061 .000062 .000063 .000064 .000065 .000066 .000067 .000070 • 000071 .000072 .000073 .000074 .000075 .000076 .000077 .000000 .000003 .000007 .000011 .000015 .000019 .000022 .000026 .000030 .000034 .000038 .000041 .000045 .000049 .000053 .000057 .000061 .000064 .000068 .000072 .000076 .000080 .000083 .000087 .000091 .000095 .000099 .000102 .000106 .000110 .000114 .000118 .000122 .000125 .000129 .000133 .000137 .000141 .000144 .000148 .000152 .000156 .000160 .000164 .000167 .000171 .000175 .000179 .000183 .000186 .000190 .000194 .000198 .000202 .000205 .000209 .000213 .000217 .000221 .000225 .000228 .000232 .000236 .000240 .000100 .000101 .000102 .000103 .000104 .000105 .000106 .000107 .000110 .000111 .000112 .000113 .000114 .000115 .000116 .000117 .000120 .000121 .000122 .000123 .000124 .000125 .000126 .000127 .000130 .000131 .000132 .000133 .000134 .000135 .000136 .000137 .000140 .000141 .000142 .000143 .000144 .000145 .000146 .000147 .000150 .000151 .000152 .000153 .000154 .000155 .000156 .000157 .000160 .000161 .000162 .000163 .000164 .000165 .000166 .000167 .000170 .000171 .000172 .000173 .000174 .000175 .000176 .000177 .000244 .000247 .000251 .000255 .000259 .000263 .000267 .000270 .000274 .000278 .000282 .000286 .000289 .000293 .000297 .000301 .000305 .000308 .000312 .000316 .000320 .000324 .000328 .000331 .000335 .000339 .000343 .000347 .000350 .000354 .000358 .000362 .000366 .000370 .000373 .000377 .000381 .000385 .000389 .000392 .000396 .000400 .000404 .000408 .000U1 .000415 .000419 .000423 .000427 .000431 .000434 .000438 .000442 .000446 .000450 .000453 .000457 .000461 .000465 .000469 .000473 .000476 .000480 .000484 .000200 .000201 .000202 .000203 .000204 .000205 .000206 .000207 .000210 .000211 .000212 .000213 .000214 .000215 .000216 .000217 .000220 .000221 .000222 .000223 .000224 .000225 .000226 .000227 .000230 .000231 .000232 .000233 .000234 .000235 .000236 .000237 .000240 .000241 .000242 .000243 .000244 .000245 .000246 .000247 .000250 .000251 .000252 .000253 .000254 .000255 .000256 .000257 .000260 .000261 .000262 .000263 .000264 .000265 .000266 • ()00267 .000270 . .0002'71 .000272 .000273 .000274 .000275 .000276 .000277 .000488 .000492 .000495 .000499 .000503 .000507 .000511 .000514 .000518 .000522 .000526 .000530 .000534 .000537 .000541 .000545 .000549 .000553 .000556 .000560 .000564 .000568 .000572 .000300 .000301 .000302 .000303 .000304 .000305 .000306 .000307 .000310 .000311 .000312 .000313 .000314 .000315 .000316 .000317 .000320 .000321 .000322 .000323 .000324 .000325 .000326 .000732 .000736 .000740 .000743 .000747 .000751 .000755 .000759 .000762 .000766 .000770 .000774 .000778 .000782 .000785 .000789 .000793 .000797 .000801 .000805 .000808 .000812 .000816 .000820 .000823 .000827 .000831 .000835 .000839 .000843 .000846 .000850 .000854 .000858 .000862 .000865 .000869 .000873 .000877 .000881 .000885 .000888 .000892 .000896 .000900 .000904 .000907 .000911 • 000911~ .000919 .000923 .000926 .000930 .000934 .000938 .000942 .000946 .000949 .000953 .000957 .000961 .000965 .000968 .000972 .000576 .000327 .000579 .000583 .000587 .000591 .000595 .000598 .00'0602 .000606 .000610 .000614 .000617 .000621 .000625 .000629 .000633 .000637 .000640 .000644 .000648 .000652 .000656 .000659 .000663 .000667 .0006'71 .000675 .000679 .000682 .000686 .000690 .000694 .000698 .000701 • Q00705 .000709 .000713 .000717 .000720 .000724 .000728 .000330 .000331 .000332 .000333 .000334 .000335 .000336 .000337 .000340 .000341 .000342 .000343 .000344 .000345 .000346 .000347 .000350 .000351 .000352 .000353 .000354 .000355 .000356 .000357 .000360 .000361 .000362 .000363 .000364 .000365 .000366 .000367 .000370 .000371 .000372 .000373 .000374 .000375 .000376 .000377 Octal-Decimal Fraction Conversion 103 Octal-Decimal Fraction Conversion Table (Continued) OCTAL DEC. OCTAL DEC. OCTAL DEC. OCTAL DEC, ,000400 ,000401 .000402 .000403 .000404 .000405 .000406 .000407 .000410 .000411 ,000412 ,000413 .000414 ,000415 .000416 ,000417 .000420 ,000421 ,000422 .000423 .000424 .000425 .000426 ,000427 .000430 ,000431 .000432 .000433 .000434 .000435 .000436 .000437 ,000440 .000441 ,000442 ,000443 ,000444 .000445 .000446 .000447 .000450 .000451 .000452 ,000453 .000454 .000455 .000456 ,000457 .000460 ,000461 ,000462 ,000463 .000464 ,000465 .000466 ,000467 ,0004'10 ,000471 ,000472 .000473 ,000474 .000475 ,000476 ,000477 ,000976 .000980 .000984 ,000988 .000991 .000995 .000999 .001003 .001007 .001010 .001014 .001018 .001022 .001026 ,001029 .001033 ,001037 .001041 .001045 ,001049 .001052 .001056 ,001060 .001064 .001068 ,001071 .001075 ,001079 .001083 .001087 .001091 ,001094 ,001098 ,001102 .001106 ,001110 ,001113 ,001117 .001121 .001125 ,001129 .001132 .001136 .001140 .001144 .001148 .001152 .001155 ,001159 ,001163 .001167 ...001171 .001174 .001178 ,001182 ,001186 ,001190 .001194 ,001197 .001201 ,001205 .001209 .001213 .001216 ,000500 .000501 .000502 .000503 .000504 .000505 .000506 ,000507 .000510 .000511 .000512 .000513 ,000514 .000515 .000516 ,000517 ,000520 .000521 .000522 .000523 .000524 .000525 .000526 .000527 .000530 .000531 .000532 .000533 .000534 .000535 ,000536 .000537 ,000540 .000541 .000542 ,000543 .000544 ,000545 ,000546 ,000547 .000550 ,000551 .000552 ,000553 .000554 .000555 .000556 .000557 ,000560 .000561 ,000562 .000563 ,000564 .000565 .000566 ,000567 ,000570 .000571 .000572 .000573 .000574 .000575 ,000576 .000577 ,001220 .001224 .001228 .001232 .001235 .001239 ,001243 .001247 .001251 .001255 .001258 ,001262 .001266 .001270 .001274 .001277 .001281 .001285 .001289 .001293 .001296 .001300 ,001304 .001308 .001312 .001316 .001319 .001323 .001327 .001331 .001335 .001338 .001342 ,001346 .001350 .001354 .001358 ,001361 .001365 .001369 ,001373 .001377 .001380 .001384 .001388 .001392 .001396 .001399 ,001403 .001407 ,001411 .001415 .001419 ,001422 .001426 .001430 .001434 ,001438 ,001441 .001445 .001449 .001453 .001457 ,001461 ,000600 .000601 .000602 .000603 .000604 .000605 .000606 .000607 ,000610 ,000611 .000612 .000613 .000614 ,000615 .000616 .000617 ,000620 ,000621 .000622 .000623 .000624 .000625 .000626 .000627 .000630 .000631 .000632 .000633 .000634 ,000635 .000636 ,000637 .000640 ,000641 .000642 .000643 .000644 .000645 ,000646 ,000647 .000650 ,800651 ,000652 ,000653 ,000654 .000655 .001)656 ,000657 ,000660 ,000661 ,000662 .000663 ,000664 ,000665 .000666 ,000667 .000670 ,000671 ,000672 ,000673 .000674 ,000675 ,000676 .000677 .001464 .001468 .001472 .001476 .001480 .001483 .001487 .001491 ,001495 .001499 .001502 .001506 .001510 .001514 .001518 .001522 .001525 .001529 .001533 .001537 ,001541 .001544 .001548 .001552 .001556 ,001560 .001564 .001567 .001571 .001575 .001579 .001583 .001586 .001590 .001594 ,001598 .001602 ,001605 ,001609 ,001613 ,001617 .001621 .001625 .001628 .001632 ,001636 .001640 ,001644 ,001647 ,001651 .001655 ,001659 ,001663 .001667 ,001670 .001674 ,001678 ,001682 ,001686 .001689 .001693 ,001697 .000700 .000701 .000702 .000703 .000704 .000705 .000706 .000707 .000710 .000711 .000712 .000713 .000714 .000715 .000716 .000717 .000720 ,000721 .000722 .000723 .000724 .000725 .000726 .000727 .000730 .000731 .000732 .000733 ,000734 ,000735 ,000736 .000737 .000740 .000741 ,000742 ,000743 ,000744 .000745 .000746 .000747 ,000750 ,000751 ,000752 .000753 .000754 .000755 ,000756 ,000757 .000760 .000761 .000762 ,000763 ,000764: ,000765 ,000766 .000767 ,000770 ,00.0771 .000772 .000773 .000774 .000775 .000776 .000777 ,001708 .001712 .001716 .001720 ,001724 .001728 .001731 .001735 .001739 .001743 .001747 ,001750 ,001754 .001758 .001762 .001766 ,001770 .001773 ,001777 ,001781 .001785 .001789 .001792 .001796 .001800 .001804 .001808 .001811 ,001815 .001819 .001823 .001827 .001831 .001834 ,001838 ,001842 ,001846 .001850 ,001853 .001857 ,001861 .001865 .001869 .001873 .001876 .001880 ,001884 .001888 ,001892 ,001895 ,001899 .001903 .001907 ,001911 .001914,001918 ,001922 ,001926 .001930 :001934 .001937 ,001941 ,001945 .001949 104 - ,OOI7.Dl ,001705 Appendix C: Table of Powers of Two 2- 11 2" 1t 1 2 4 8 0 1 2 3 1.0 0,5 0,25 0,125 16 32 64 128 4 5 6 7 0,062 0,031 0,015 0,007 5 25 625 812 5 256 512 1 024 2 048 8 9 10 11 0,003 0,001 0,000 0,000 906 953 976 488 4 096 8 192 16 384 32 768 12 13 14 15 0.000 0.000 0,000 0,000 244 140 122' 070 061 035 030 517 625 312 5 156 25 578 125 25 125 562 5 281 25 65 131 262 524 536 072 144 288 16 17 18 19 0.000 0,000 0,000 0.000 015 007 003 001 258 629 814 907 789 394 69'7 348 062 531 265 632 5 25 625 812 5 1 2 4 8 048 097 194 388 576 152 304 608 20 21 22 23 0.000 0,000 0,000 0.000 000 000 000 000 953 476 238 119 674 8'37 418 209 316 158 579 289 406 203 101 550 25 125 562 5 781 25 16 33 67 134 777 554 108 217 216 432 864 728 24 25 26 27 0.000 0.000 0.000 0.000 000 000 000 000 059 029 014 00-7 604 802 901 450 644 322 161 580 775 387 193 596 390 695 847 923 625 312 5 656 25 828 125 268435 536 870 1 073 741 2 147 483 456 912 824 648 28 29 30 31 0.000 0.000 0.000 0.000 000 000 000 000 003 001 000 000 725 862 931 465 290 645 322 661 298 149 574 287 461 230 615 307 914 957 478 739 062 031 515 257 5 25 625 812 5 4 8 17 34 294 589 179 359 967 934 869 738 296 592 184 368 32 33 34 35 0.000 0.000 0.000 0.000 000 000 000 000 000 000 000 000 232 116 058 029 830 415 207 103 643 321 660 830 653 826 913 456 869 934 467 733 628 814 407 703 906 453 226 613 25 125 562 5 281 25 68 137 274 549 719 438 877 755 476 953 906 813 736 472 944 888 36 37 38 39 0.000 0.000 0.000 0.000 000 000 000 000 000 000 000 000 014 007 003 001 551 275 637 818 915 957 978 989 228 614 807 403 366 183 091 545 851 425 712 856 806 903 951 475 640 320 660 830 625 312 5 156 25 078 125 Powers of Two 105 COMMENT SHEET IBM 7094 II DATA PROCESSING SYSTEM - VOLUME I CUSTOMER ENGINEERING INSTRUCTION-MAINTENANCE. FORM 223-2721-0 FROM NAME OFFICE NO. CHECK ONE OF THE COMMENTS AND EXPLAIN IN THE SPACE PROVIDED D SUGGESTED ADDITION (PAGE o SUGGESTED DELETION (PAGE D ERROR (PAGE , TIMING CHART, DRAWING, PROCEDURE, FOLD ETC.) EXPLANA TION 1111 Zl -I .JI C)I ZI 0 1 .JI (I I .. I jl U FOLD FOLD NO POSTAGE NECESSARY IF MAILED IN U. S. A. FOLD ON TWO LINES, STAPLE, AND MAIL ~TAPLE ~OLD -- - FOLD - -- - - - ---- ---------------.-----------------FIRST CLASS PERM IT NO. 81 POUGHKEEPSIE. N.Y. BUSINESS REPLY MAIL NO POSTAGE STAMP NECESSARY IF MAILED IN U. S. A. 1&1 Z .J POSTAGE WILL BE PAID BY t!J IBM CORPORATION o Z .J ( P.O. BOX 390 ...:J o POUGHKEEPSIE, N. Y. 12602 ATTN: CE MANUALS, DEPARTMENT 8"96 )LD FOLD 10j64:1.SM VO-112 STAPLE STAPLE I I I I I 223·2721 ·0 ITmlllli (J) International Business Machines Corporation Data Processing Division 112 East Post Road, White Plains , N. Y. 10601
Source Exif Data:
File Type : PDF File Type Extension : pdf MIME Type : application/pdf PDF Version : 1.3 Linearized : No XMP Toolkit : Adobe XMP Core 4.2.1-c043 52.372728, 2009/01/18-15:56:37 Create Date : 2014:07:26 13:11:29-08:00 Modify Date : 2014:07:26 12:29:21-07:00 Metadata Date : 2014:07:26 12:29:21-07:00 Producer : Adobe Acrobat 9.55 Paper Capture Plug-in Format : application/pdf Document ID : uuid:43fb9fb0-2940-3e49-8cbc-0b62cb77d8fe Instance ID : uuid:a03fa793-365e-4f44-ba44-fe614c63db06 Page Layout : SinglePage Page Mode : UseOutlines Page Count : 109EXIF Metadata provided by EXIF.tools