Introduction to Concurrent Programming

Size: px

Start display at page:

Download "Introduction to Concurrent Programming"

Emory Chambers
6 years ago
Views:

Σχολή Ηλεκτρολόγων Μηχανικών και Μηχανικών Υπολογιστών Τομέας Τεχνολογίας Πληροφορικής και Υπολογιστών Εθνικό Μετσόβιο Πολυτεχνείο Γλώσσες Προγραμματισμού ΙΙ Διδάσκοντες: Νικόλαος

1 Σχολή Ηλεκτρολόγων Μηχανικών και Μηχανικών Υπολογιστών Τομέας Τεχνολογίας Πληροφορικής και Υπολογιστών Εθνικό Μετσόβιο Πολυτεχνείο Γλώσσες Προγραμματισμού ΙΙ Διδάσκοντες: Νικόλαος Παπασπύρου, Κωστής Σαγώνας Introduction to Concurrent Programming Based on the companion slides for the book The Art of Multiprocessor Programming by Maurice Herlihy & Nir Shavit, 2008

2 Άδεια Χρήσης Το παρόν εκπαιδευτικό υλικό υπόκειται σε άδειες χρήσης Creative Commons. Για εκπαιδευτικό υλικό, όπως εικόνες, που υπόκειται σε άδεια χρήσης άλλου τύπου, αυτή πρέπει να αναφέρεται ρητώς.

3 From the New York Times SAN FRANCISCO, 7 May Intel said on Friday that it was scrapping its development of two microprocessors, a move that is a shift in the company's business strategy. Introduction to Concurrent Programming 2

4 Moore s Law Transistor count still rising Clock speed flattening sharply Introduction to Concurrent Programming 3

5 Your Old Desktop: A Uniprocesor cpu memory Introduction to Concurrent Programming 4

6 Till recently in the Enterprise: The Shared Memory Multiprocessor (SMP) cache cache Bus cache Bus shared memory Introduction to Concurrent Programming 5

7 On Your New Desktop: The Multicore processor (CMP) All on the same chip cache cache cache Bus Bus shared memory Sun T2000 Niagara Introduction to Concurrent Programming 6

8 Multicores Are Here Intel's Intel ups ante with 4-core chip. New microprocessor, due this year, will be faster, use less electricity... [San Fran Chronicle] AMD will launch a dual-core version of its Opteron server processor at an event in New York on April 21. [PC World] Sun s Niagara will have eight cores, each core capable of running 4 threads in parallel, for 32 concurrently running threads.. [The Inquierer] Introduction to Concurrent Programming 7

9 Why do we care? Time no longer cures software bloat The free ride is over When you double your program s path length You can t just wait 6 months Your software must somehow exploit twice as much concurrency Introduction to Concurrent Programming 8

10 Traditional Scaling Process Speedup 1.8x 3.6x 7x User code Traditional Uniprocessor Time: Moore s law Introduction to Concurrent Programming 9

11 Multicore Scaling Process Speedup 1.8x 3.6x 7x User code Multicore Unfortunately, not so simple Introduction to Concurrent Programming 10

12 Real-World Scaling Process Speedup 1.8x 2x 2.9x User code Multicore Parallelization and Synchronization require great care Introduction to Concurrent Programming 11

13 Sequential Computation thread memory object object Introduction to Concurrent Programming 12

14 Concurrent Computation threads memory object object Introduction to Concurrent Programming 13

15 Asynchrony Sudden unpredictable delays Cache misses (short) Page faults (long) Scheduling quantum used up (really long) Introduction to Concurrent Programming 14

16 Model Summary Multiple threads Sometimes called processes Single shared memory Objects live in memory Unpredictable asynchronous delays Introduction to Concurrent Programming 15

17 Road Map We are going to focus on principles first, then practice Start with idealized models Look at simplistic problems Emphasize correctness over pragmatism Correctness may be theoretical, but incorrectness has practical impact Introduction to Concurrent Programming 16

18 Concurrency Jargon Hardware Processors, cores Software Threads, processes Sometimes OK to confuse them, sometimes not Introduction to Concurrent Programming 17

19 Parallel Primality Testing Challenge Print primes from 1 to Given Ten-processor multiprocessor One thread per processor Goal Get ten-fold speedup (or close) Introduction to Concurrent Programming 18

20 Load Balancing P 0 P 1 P 9 Split the work evenly Each thread tests range of 10 9 Introduction to Concurrent Programming 19

21 Procedure for Thread i void primeprint { int i = ThreadID.get(); // IDs in {0..9} for (j = i* , j < (i+1)*10 9 ; j++) { if (isprime(j)) print(j); } } Introduction to Concurrent Programming 20

22 Issues Higher ranges have fewer primes Yet larger numbers harder to test Thread workloads Uneven Hard to predict rejected Need dynamic load balancing Introduction to Concurrent Programming 21

23 Shared Counter each thread takes a number 17 Introduction to Concurrent Programming 22

24 Procedure for Thread i int counter = new Counter(1); void primeprint { long j = 0; while (j < ) { j = counter.getandincrement(); if (isprime(j)) print(j); } } Introduction to Concurrent Programming 23

25 Procedure for Thread i Counter counter = new Counter(1); void primeprint { long j = 0; while (j < ) { j = counter.getandincrement(); } } if (isprime(j)) print(j); Shared counter object Introduction to Concurrent Programming 24

26 Where Things Reside int counter = new Counter(1); void primeprint { long j = 0; while (j < ) { j = counter.getandincrement(); if (isprime(j)) print(j); } } code Local variables cache cache Bus cache Bus 1 shared memory shared counter Introduction to Concurrent Programming 25

27 Procedure for Thread i Counter counter = new Counter(1); void primeprint { Stop when every long j = 0; value taken while (j < ) { j = counter.getandincrement(); if (isprime(j)) print(j); } } Introduction to Concurrent Programming 26

28 Procedure for Thread i Counter counter = new Counter(1); void primeprint { long j = 0; while (j < ) { j = counter.getandincrement(); if (isprime(j)) print(j); } } Increment & return each new value Introduction to Concurrent Programming 27

29 Counter Implementation public class Counter { private long value; } public long getandincrement() { return value++; } OK for single thread, not for concurrent threads Introduction to Concurrent Programming 28

30 What It Means public class Counter { private long value; } public long getandincrement() { return value++; temp = value; } value = value + 1; return temp; Introduction to Concurrent Programming 29

31 Not so good Value read 1 write 2 read 2 write 3 read 1 write 2 time Introduction to Concurrent Programming 30

32 Is this problem inherent? read write write read If we could only glue reads and writes Introduction to Concurrent Programming 31

33 Challenge public class Counter { private long value; } public long getandincrement() { temp = value; value = temp + 1; return temp; } Make these steps atomic (indivisible) Introduction to Concurrent Programming 32

34 Hardware Solution public class Counter { private long value; public long getandincrement() { temp = value; value = temp + 1; return temp; } } ReadModifyWrite() instruction Introduction to Concurrent Programming 33

35 An Aside: Java public class Counter { private long value; } public long getandincrement() { synchronized { temp = value; value = temp + 1; } return temp; } Synchronized block Introduction to Concurrent Programming 34

36 An Aside: Java public class Counter { private long value; } public long getandincrement() { synchronized { temp = value; value = temp + 1; } return temp; } Mutual Exclusion Introduction to Concurrent Programming 35

37 Mutual Exclusion or Alice & Bob share a pond A B Introduction to Concurrent Programming 36

38 Alice has a pet A B Introduction to Concurrent Programming 37

39 Bob has a pet A B Introduction to Concurrent Programming 38

40 The Problem A B The pets don t get along Introduction to Concurrent Programming 39

41 Formalizing the Problem Two types of formal properties in asynchronous computation: Safety Properties Nothing bad happens ever Liveness Properties Something good happens eventually Introduction to Concurrent Programming 40

42 Formalizing our Problem Mutual Exclusion Both pets never in pond simultaneously This is a safety property No Deadlock if only one wants in, it gets in if both want in, one gets in. This is a liveness property Introduction to Concurrent Programming 41

43 Simple Protocol Idea Just look at the pond Gotcha Trees obscure the view Introduction to Concurrent Programming 42

44 Interpretation Threads can t see what other threads are doing Explicit communication required for coordination Introduction to Concurrent Programming 43

45 Mobile Phone Protocol Idea Bob sends SMS to Alice (or vice-versa) Gotcha Bob takes shower Alice recharges battery Bob out shopping for pet food Introduction to Concurrent Programming 44

46 Interpretation Message-passing doesn t work Recipient might not be Listening There at all Communication must be Persistent (like writing) Not transient (like speaking) Introduction to Concurrent Programming 45

47 Can Protocol cola cola Introduction to Concurrent Programming 46

48 Bob conveys a bit A B cola Introduction to Concurrent Programming 47

49 Bob conveys a bit A B cola Introduction to Concurrent Programming 48

50 Can Protocol Idea Cans on Alice s windowsill Strings lead to Bob s house Bob pulls strings, knocks over cans Gotcha Cans cannot be reused Bob runs out of cans Introduction to Concurrent Programming 49

51 Interpretation Cannot solve mutual exclusion with interrupts Sender sets fixed bit in receiver s space Receiver resets bit when ready Requires unbounded number of interrupt bits Introduction to Concurrent Programming 50

52 Flag Protocol A B Introduction to Concurrent Programming 51

53 Alice s Protocol (sort of) A B Introduction to Concurrent Programming 52

54 Bob s Protocol (sort of) A B Introduction to Concurrent Programming 53

55 Alice s Protocol Raise flag Wait until Bob s flag is down Unleash pet Lower flag when pet returns Introduction to Concurrent Programming 54

56 Bob s Protocol Raise flag Wait until Alice s flag is down Unleash pet Lower flag when pet returns danger! Introduction to Concurrent Programming 55

57 Bob s Protocol (2 nd try) Bob defers Raise flag to Alice While Alice s flag is up Lower flag Wait for Alice s flag to go down Raise flag Unleash pet Lower flag when pet returns Introduction to Concurrent Programming 56

58 The Flag Principle Raise the flag Look at other s flag Flag Principle: If each raises and looks, then Last to look must see both flags up Introduction to Concurrent Programming 57

59 Proof of Mutual Exclusion Assume both pets in pond Derive a contradiction By reasoning backwards Consider the last time Alice and Bob each looked before letting the pets in Without loss of generality assume Alice was the last to look Introduction to Concurrent Programming 58

60 Proof Bob last raised flag Alice last raised her flag Alice s last look Bob s last looked time QED Alice must have Introduction seen to Concurrent Bob s Programming flag. A contradiction 59

61 Proof of No Deadlock If only one pet wants in, it gets in. Deadlock requires both continually trying to get in. If Bob sees Alice s flag, he gives her priority (a gentleman ) QED Introduction to Concurrent Programming 60

62 Remarks Protocol is unfair Bob s pet might never get in Protocol uses waiting If Bob is eaten by his pet, Alice s pet might never get in Introduction to Concurrent Programming 61

63 Moral of Story Mutual Exclusion cannot be solved by transient communication (mobile phones) interrupts (cans) It can be solved by one-bit shared variables that can be read or written Introduction to Concurrent Programming 62

64 The Arbiter Problem (an aside) Pick a point Pick a point Introduction to Concurrent Programming 63

65 The Fable Continues Alice and Bob fall in love & marry Then they fall out of love & divorce She gets the pets (who now get along fine with each other) He has to feed them Leading to a new coordination problem: Producer-Consumer Introduction to Concurrent Programming 64

66 Bob Puts Food in the Pond A Introduction to Concurrent Programming 65

67 Alice releases her pets to Feed mmm mmm B Introduction to Concurrent Programming 66

68 Producer/Consumer Alice and Bob can t meet Each has restraining order on other So he puts food in the pond And later, she releases the pets Avoid Releasing pets when there s no food Putting out food if uneaten food remains Introduction to Concurrent Programming 67

69 Producer/Consumer Need a mechanism so that Bob lets Alice know when food has been put out Alice lets Bob know when to put out more food Introduction to Concurrent Programming 68

70 Surprise Solution A B cola Introduction to Concurrent Programming 69

71 Bob puts food in Pond A B cola Introduction to Concurrent Programming 70

72 Bob knocks over Can A B cola Introduction to Concurrent Programming 71

73 Alice Releases Pets A yum yum B cola Introduction to Concurrent Programming 72

74 Alice Resets Can when Pets are Fed A B cola Introduction to Concurrent Programming 73

75 Pseudocode while (true) { while (can.isup()){}; pet.release(); Bob s code pet.recapture(); can.reset(); } while (true) { while (can.isdown()){}; pond.stockwithfood(); can.knockover(); } Alice s code Introduction to Concurrent Programming 74

76 Correctness Mutual Exclusion Pets and Bob never together in pond No Starvation safety liveness if Bob always willing to feed, and pets always famished, then pets eat infinitely often. Producer/Consumer The pets never enter pond unless there is food, and Bob never provides food if there is unconsumed food. safety Introduction to Concurrent Programming 75

77 Could Also Solve Using Flags A B Introduction to Concurrent Programming 76

78 Waiting Both solutions use waiting while(mumble){} Waiting is problematic If one participant is delayed So is everyone else But delays are common & unpredictable Introduction to Concurrent Programming 77

79 The Fable drags on Bob and Alice still have issues But they need to communicate So they agree to use billboards Introduction to Concurrent Programming 78

80 Billboards are Large B A 3 C 1 3 D 2 E 1 Letter Tiles From Scrabble box Introduction to Concurrent Programming 79

81 Write One Letter at a Time W 4 A 1 S 1 H4 A 1 B 3 C 3 D 2 E 1 Introduction to Concurrent Programming 80

82 To post a message W A S H T H E C A R whe w Introduction to Concurrent Programming 81

83 Let s send another message L S E L L L A V A 1 A M P3 3 S 1 Introduction to Concurrent Programming 82

84 Uh-Oh S 1 E 1 L 1 L 1 T 1 H 4 E 1 C A R L 1 OK Introduction to Concurrent Programming 83

85 Readers/Writers Devise a protocol so that Writer writes one letter at a time Reader reads one letter at a time Reader sees Old message or new message No mixed messages Introduction to Concurrent Programming 84

86 Readers/Writers (cont d) Easy with mutual exclusion But mutual exclusion requires waiting One waits for the other Everyone executes sequentially Remarkably We can solve R/W without mutual exclusion Introduction to Concurrent Programming 85

87 Why do we care? We want as much of the code as possible to execute concurrently (in parallel) A larger sequential part implies reduced performance Amdahl s law: this relation is not linear Introduction to Concurrent Programming 86

88 Amdahl s Law Speedup = OldExecutionTime NewExecutionTime of computation given n CPUs instead of 1 Introduction to Concurrent Programming 87

89 Amdahl s Law Sequential fraction Speedup = Number of processors 1 1 p + p n Parallel fraction Introduction to Concurrent Programming 88

90 Example Ten processors 60% concurrent, 40% sequential How close to 10-fold speedup? Speedup=2.17= Introduction to Concurrent Programming 89

91 Example Ten processors 80% concurrent, 20% sequential How close to 10-fold speedup? Speedup=3.57= Introduction to Concurrent Programming 90

92 Example Ten processors 90% concurrent, 10% sequential How close to 10-fold speedup? Speedup=5.26= Introduction to Concurrent Programming 91

93 Example Ten processors 99% concurrent, 01% sequential How close to 10-fold speedup? Speedup=9.17= Introduction to Concurrent Programming 92

94 The Moral Making good use of our multiple processors (cores) means Finding ways to effectively parallelize our code Minimize sequential parts Reduce idle time: time in which threads wait without work Introduction to Concurrent Programming 93

95 Άδεια Χρήσης Το παρόν εκπαιδευτικό υλικό υπόκειται σε άδειες χρήσης Creative Commons. Για εκπαιδευτικό υλικό, όπως εικόνες, που υπόκειται σε άδεια χρήσης άλλου τύπου, αυτή πρέπει να αναφέρεται ρητώς.

Introduction to Concurrent Programming

Introduction to Concurrent Programming Based on the companion slides for the book The Art of Multiprocessor Programming by Maurice Herlihy & Nir Shavit, 2008 From the New York Times SAN FRANCISCO, 7 May