Fundamentals of Computer Systems

Transcription

1 Fundamentals of Computer Systems Boolean Logic Stephen A. Edwards Columbia University Summer 2017

2 Boolean Logic George Boole

3 Boole s Intuition Behind Boolean Logic Variables,,... represent classes of things No imprecision: A thing either is or is not in a class If is sheep and is white things, are all white sheep, and = =. If is men and is women, + is both men and women, and + = + + =. If is men, is women, and Z is European, Z( +) is European men and women and Z( +) = Z +Z.

4 The Axioms of (Any) Boolean Algebra A Boolean Algebra consists of A set of values A An and operator An or operator + A not operator A false value 0 A A true value 1 A

5 The Axioms of (Any) Boolean Algebra A Boolean Algebra consists of A set of values A An and operator An or operator + Axioms A not operator A false value 0 A A true value 1 A + = + = +( +Z) = ( +) +Z ( Z) = ( ) Z +( ) = ( +) = ( +Z) = ( ) +( Z) +( Z) = ( +) ( +Z) + = 1 = 0

6 The Axioms of (Any) Boolean Algebra A Boolean Algebra consists of A set of values A An and operator An or operator + Axioms A not operator A false value 0 A A true value 1 A + = + = +( +Z) = ( +) +Z ( Z) = ( ) Z +( ) = ( +) = ( +Z) = ( ) +( Z) +( Z) = ( +) ( +Z) + = 1 = 0 We will use the first non-trivial Boolean Algebra: A = {0,1}. This adds the law of excluded middle: if 0 then = 1 and if 1 then = 0.

7 Simplifying a Boolean Expression ou are a New orker if you were born in New ork or were not born in New ork and lived here ten years. Axioms +( ) + = + = +( +Z) = ( +) +Z ( Z) = ( ) Z +( ) = ( +) = ( +Z) = ( ) +( Z) +( Z) = ( +) ( +Z) + = 1 = 0 Lemma: 1 = ( +) = ( +) if = =

8 Simplifying a Boolean Expression ou are a New orker if you were born in New ork or were not born in New ork and lived here ten years. Axioms +( ) = ( +) ( +) + = + = +( +Z) = ( +) +Z ( Z) = ( ) Z +( ) = ( +) = ( +Z) = ( ) +( Z) +( Z) = ( +) ( +Z) + = 1 = 0 Lemma: 1 = ( +) = ( +) if = =

9 Simplifying a Boolean Expression ou are a New orker if you were born in New ork or were not born in New ork and lived here ten years. Axioms +( ) = ( +) ( +) = 1 ( +) + = + = +( +Z) = ( +) +Z ( Z) = ( ) Z +( ) = ( +) = ( +Z) = ( ) +( Z) +( Z) = ( +) ( +Z) + = 1 = 0 Lemma: 1 = ( +) = ( +) if = =

10 Simplifying a Boolean Expression ou are a New orker if you were born in New ork or were not born in New ork and lived here ten years. Axioms +( ) = ( +) ( +) = 1 ( +) = + Lemma: + = + = +( +Z) = ( +) +Z ( Z) = ( ) Z +( ) = ( +) = ( +Z) = ( ) +( Z) +( Z) = ( +) ( +Z) + = 1 = 0 1 = ( +) = ( +) if = =

11 More properties 0 +0 = = = = = 1 +0 = +1 = 1 + = + = + = = = = = = 1 0 = 0 1 = = ( +) = ( +) =

12 More Examples +Z( +Z) = +Z +ZZ = +Z = ( +Z) +( +Z) +Z = + +Z +Z = +Z +Z = +Z

13 More Examples Z +( +Z) = Z + +Z Expand = (Z + +Z) Factor w.r.t. = (Z + +Z +Z) Z Z = (Z +Z + +Z) Reorder = ( (Z +Z) + +Z ) Factor w.r.t. = ( + +Z) + = 1 = (1 +Z) 1 +Z = 1 = 1 = ( + +Z)( +Z) = +Z + + Z +Z +Z Z = +Z + +Z +Z = +Z

14 Sum-of-products form Can always reduce a complex Boolean expression to a sum of product terms: + ( +(Z +) +Z ) = +( +Z + +Z) = + +Z + + Z = +Z + Z (can do better) = ( +Z) + Z = ( +Z) + Z = Z + Z = + Z

15 What Does This Have To Do With Logic Circuits? Claude Shannon

16 Shannon s MS Thesis We shall limit our treatment to circuits containing only relay contacts and switches, and therefore at any given time the circuit between any two terminals must be either open (infinite impedance) or closed (zero impedance).

17 Shannon s MS Thesis It is evident that with the above definitions the following postulates hold. 0 0 = 0 A closed circuit in parallel with a closed circuit is a closed circuit = 1 An open circuit in series with an open circuit is an open circuit = 0 +1 = 1 An open circuit in series with a closed circuit in either order is an open circuit. 0 1 = 1 0 = 0 A closed circuit in parallel with an open circuit in either order is an closed circuit = 0 A closed circuit in series with a closed circuit is a closed circuit. 1 1 = 1 An open circuit in parallel with an open circuit is an open circuit. At any give time either = 0 or = 1

18 Alternate Notations for Boolean Logic Operator Math Engineer Schematic Copy x or Complement x AND x y or OR x y + +

19 Definitions Literal: a Boolean variable or its complement E.g., Implicant: A product of literals E.g., Z Minterm: An implicant with each variable once E.g., Z Z Z Maxterm: A sum of literals with each variable once E.g., + +Z + +Z + +Z

20 Be Careful with Bars

21 Be Careful with Bars Let s check all the combinations of and :

22 Truth Tables A truth table is a canonical representation of a Boolean function Minterm Maxterm Each row has a unique minterm and maxterm The minterm is 1 maxterm is 0 for only its row

23 Sum-of-minterms and Product-of-maxterms Two mechanical ways to translate a function s truth table into an expression: Minterm Maxterm F The sum of the minterms where the function is 1: F = + The product of the maxterms where the function is 0: F = ( +)( +)

24 Expressions to Schematics F = +

25 Expressions to Schematics F = +

26 Expressions to Schematics F = +

27 Expressions to Schematics F = +

28 Expressions to Schematics F = + + = F

29 Expressions to Schematics F = + = ( +)( +) + = F ( +)( +) = F

30 Minterms and Maxterms: Another Example The minterm and maxterm representation of functions may look very different: Minterm Maxterm F The sum of the minterms where the function is 1: F = + + The product of the maxterms where the function is 0: F = +

31 Expressions to Schematics 2 F = + + = = F + = F

32 The Menagerie of Gates

33 The Menagerie of Gates Buffer Inverter AND NAND OR NOR OR NOR

34 De Morgan s Theorem + = = + Proof by Truth Table:

35 De Morgan s Theorem in Gates AB = A +B = A +B = A B =

36 Bubble Pushing A B C D F Apply De Morgan s Theorem: Transform NAND into OR with inverted inputs

37 Bubble Pushing A B C D F Apply De Morgan s Theorem: Transform NAND into OR with inverted inputs Two bubbles on a wire cancel

38 Bubble Pushing A B C D F Apply De Morgan s Theorem: Transform NAND into OR with inverted inputs Two bubbles on a wire cancel

39 PONG PONG, Atari 1973 Built from TTL logic gates; no computer, no software Launched the video arcade game revolution

40 Horizontal Ball Control in PONG M L R A B The ball moves either left or right. Part of the control circuit has three inputs: M ( move ), L ( left ), and R ( right ). It produces two outputs A and B. Here, means I don t care what the output is; I never expect this input combination to occur.

41 Horizontal Ball Control in PONG M L R A B E.g., assume all the s are 0 s and use Minterms: A = MLR +MLR B = MLR +MLR +MLR 3 inv + 4 AND3 + 1 OR2 + 1 OR3

42 Horizontal Ball Control in PONG M L R A B Assume all the s are 1 s and use Maxterms: A = (M +L+R)(M +L+R) B = M +L+R 3 inv + 3 OR3 + 1 AND2

43 Horizontal Ball Control in PONG M L R A B Choosing better values for the s and being much more clever: A = M B = MR 1 NAND2 (!)

44 Karnaugh Maps Basic trick: put similar variable values near each other so simple functions are obvious M L R A B The M s are already arranged nicely

45 Karnaugh Maps Basic trick: put similar variable values near each other so simple functions are obvious M L R A B Let s rearrange the L s by permuting two pairs of rows

46 Karnaugh Maps Basic trick: put similar variable values near each other so simple functions are obvious M L R A B Let s rearrange the L s by permuting two pairs of rows

47 Karnaugh Maps Basic trick: put similar variable values near each other so simple functions are obvious M L R A B Let s rearrange the L s by permuting two pairs of rows

48 Karnaugh Maps Basic trick: put similar variable values near each other so simple functions are obvious M L R A B Let s rearrange the L s by permuting two pairs of rows

49 Karnaugh Maps Basic trick: put similar variable values near each other so simple functions are obvious M L R A B Let s rearrange the L s by permuting two pairs of rows

50 Karnaugh Maps Basic trick: put similar variable values near each other so simple functions are obvious M L R A B Let s rearrange the L s by permuting two pairs of rows

51 Karnaugh Maps Basic trick: put similar variable values near each other so simple functions are obvious M L R A B Let s rearrange the L s by permuting two pairs of rows

52 Karnaugh Maps Basic trick: put similar variable values near each other so simple functions are obvious M L R A B Let s rearrange the L s by permuting two pairs of rows

53 Karnaugh Maps Basic trick: put similar variable values near each other so simple functions are obvious M L R A B The R s are really crazy; let s use the second dimension

54 Karnaugh Maps Basic trick: put similar variable values near each other so simple functions are obvious M L R A B The R s are really crazy; let s use the second dimension

55 Karnaugh Maps Basic trick: put similar variable values near each other so simple functions are obvious M L R A B The R s are really crazy; let s use the second dimension

56 Karnaugh Maps Basic trick: put similar variable values near each other so simple functions are obvious M L R A B MR M

57 Maurice Karnaugh s Maps Transactions of the AIEE, 1953

58 Karnaugh Maps All functions can be expressed with a map. There is one square for each minterm in a function s truth table minterm 0 0 m0 0 1 m1 1 0 m2 1 1 m m0 m1 1 m2 m3

59 Karnaugh Maps All functions can be expressed with a map. There is one square for each minterm in a function s truth table minterm 0 0 m0 0 1 m1 1 0 m2 1 1 m m0 m1 1 m2 m3

60 Karnaugh Maps All functions can be expressed with a map. There is one square for each minterm in a function s truth table minterm 0 0 m0 0 1 m1 1 0 m2 1 1 m m0 m1 1 m2 m3

61 Karnaugh Maps Fill out the table with the values of some function. F

62 Karnaugh Maps When two cells share an edge and both are 1, those two terms can be combined to form a single, simpler term F = + +

63 Karnaugh Maps When two cells share an edge and both are 1, those two terms can be combined to form a single, simpler term F = F = F = +

64 Karnaugh Maps When two cells share an edge and both are 1, those two terms can be combined to form a single, simpler term F = F = F = F = +

65 Karnaugh Maps Circle contiguous groups of 1s Circles may be 1 1, 1 2, 1 4, 2 1, 2 2, 2 4, etc. Each circle represents an implicant The bigger the circle, the simpler the implicant Circle all and only 1 s to implement the function A Prime Implicant is a circle that can t be made bigger An Essential Prime Implicant is a prime implicant that covers a 1 covered by no other prime F = +

66 3-Variable Karnaugh Maps Use gray ordering on edges with multiple variables Gray encoding: order of values such that only one bit changes at a time Two minterms are considered adjacent if they differ in only one variable (this means maps wrap ) Z m0 m1 m3 m2 1 m4 m5 m7 m6

67 3-Variable Karnaugh Maps Use gray ordering on edges with multiple variables Gray encoding: order of values such that only one bit changes at a time Two minterms are considered adjacent if they differ in only one variable (this means maps wrap ) Z m0 m1 m3 m2 m4 m5 m7 m6 Z Z m0 m1 m3 m2 m4 m5 m7 m6

68 4-Variable Karnaugh Maps An extension of 3-variable maps. C D A B B A D C B D C A

69 The Seven-Segment Decoder Example a W Z a b c d e f g e f d g c b

70 Karnaugh Map for Seg. a W Z a Z W The Karnaugh Map Sum-of-Products Challenge Cover all the 1 s and none of the 0 s using as few literals (gate inputs) as possible. Few, large rectangles are good. Covering s is optional.

71 Karnaugh Map for Seg. a W Z a Z W The minterm solution: cover each 1 with a single implicant. a = W Z +W Z +W Z = 32 literals W Z +W Z +W Z + W Z +W Z 4 inv + 8 AND4 + 1 OR8

72 Karnaugh Map for Seg. a W Z a Z Merging implicants helps Recall the distributive law: AB +AC = A(B +C) W a = W Z +W + W Z +W = 12 literals 4 inv + 1 AND4 + 2 AND3 + 1 AND2 + 1 OR4

73 Karnaugh Map for Seg. a W Z a Z W Missed one: Remember this is actually a torus. a = Z +W = 11 literals W Z +W 4 inv + 3 AND3 + 1 AND2 + 1 OR4

74 Karnaugh Map for Seg. a W Z a Z W Taking don t-cares into account, we can enlarge two implicants: a = Z +W = 9 literals W Z +W 3 inv + 1 AND3 + 3 AND2 + 1 OR4

75 Karnaugh Map for Seg. a W Z a Z W Can also compute the complement of the function and invert the result. Covering the 0 s instead of the 1 s: a = W Z + Z +W = 9 literals 5 inv + 1 AND4 + 1 AND3 + 1 AND2 + 1 OR3

76 Karnaugh Map for Seg. a W Z a Z W To display the score, PONG used a TTL chip with this solution in it: (13) (12) OUTPUT a OUTPUT b

77 Another Karnaugh Map Example Z W Consider building a minimal two-level circuit for this function. Start by choose a large number of adjacent 1 s and s in a cube shape.

78 Another Karnaugh Map Example Z Z W W Here s a big group and the Karnaugh map of the corresponding implicant.

79 Another Karnaugh Map Example Z Z W W The implicant covers 4 1 s, so it only consists of two terms.

80 Another Karnaugh Map Example Z Z Z W W W Not all the 1 s are covered, so we need to choose another group of adjacent 1 s and s. Here is the Karnaugh map of the corresponding implicant.

81 Another Karnaugh Map Example Z Z Z W W W W Z WZ This implicant only covers 2 1 s, so it has three terms.

82 Another Karnaugh Map Example Z Z Z W W W W Z W Z WZ WZ + Together, these two implicants cover all the 1 s. ORing the two implicants together gives the answer.

83 Boolean Laws and Karnaugh Maps W Z W Z +W Z+ W Z +W Z+ W Z +W Z+ W Z +W Z Factor out the W s

84 Boolean Laws and Karnaugh Maps W Z (W +W) Z+ (W +W) Z+ (W +W) Z+ (W +W) Z Use the identities W +W = 1 and 1 =.

85 Boolean Laws and Karnaugh Maps W Z Z+ Z+ Z+ Z Factor out the s

86 Boolean Laws and Karnaugh Maps W Z ( +) Z+ ( +) Z Apply the identities again

87 Boolean Laws and Karnaugh Maps W Z Z+ Z Factor out Z

88 Boolean Laws and Karnaugh Maps W Z (Z +Z) Simplify

89 Boolean Laws and Karnaugh Maps W Z Done