Concurrent Programming

01

Sachintha Gunasena MBCShttp://lk.linkedin.com/in/sachinthadtg

Learning Outcomes

Sachintha Gunasena MBCShttp://lk.linkedin.com/in/sachinthadtg

Learning Outcomes• Demonstrate foundational computing knowledge

of concurrent systems, their performance opportunities and how to implement them using:• Java concurrent features and their semantics• Java packages and APIs for concurrent programs• Conventional synchronisation algorithms, data-

structures and APIs• Wait-free and lock-free synchronisation controls.

Learning Outcomes Cont.d• Apply knowledge of computing principles and

technical skills to parallelise tasks to improve their performance and response characteristics by:• Using abstraction and computational thinking• Developing, implementing and testing the

effectiveness of alternate Java programs with different levels of concurrency

• Critiquing the approach used to solve a problem by evaluating its strengths and weaknesses

starting with the basics…

Programming• the process of writing computer programs.• example

• Output

class First { public static void main(String[] arguments) { System.out.println("Let's do something using Java technology."); }}

OOP• Object

• An object is a software bundle of related state and behavior. Software objects are often used to model the real-world objects that you find in everyday life. This lesson explains how state and behavior are represented within an object, introduces the concept of data encapsulation, and explains the benefits of designing your software in this manner.

• Class• A class is a blueprint or prototype from which objects are created. This section

defines a class that models the state and behavior of a real-world object. It intentionally focuses on the basics, showing how even a simple class can cleanly model state and behavior.

• Inheritance• Inheritance provides a powerful and natural mechanism for organizing and

structuring your software. This section explains how classes inherit state and behavior from their superclasses, and explains how to derive one class from another using the simple syntax provided by the Java programming language.

OOP Cont.d• Encapsulation

• If a class disallows calling code from accessing internal object data and forces access through methods only, this is a strong form of abstraction or information hiding known as encapsulation

• Interface• An interface is a contract between a class and the outside world. When a

class implements an interface, it promises to provide the behavior published by that interface. This section defines a simple interface and explains the necessary changes for any class that implements it.

• Package• A package is a namespace for organizing classes and interfaces in a

logical manner. Placing your code into packages makes large software projects easier to manage. This section explains why this is useful, and introduces you to the Application Programming Interface (API) provided by the Java platform.

Concurrency• a property of systems in which several computations

are executing simultaneously, and potentially interacting with each other.

• The computations may be executing on multiple cores in the same chip, preemptively time-shared threads on the same processor, or executed on physically separated processors.

• A number of mathematical models have been developed for general concurrent computation including Petri nets, process calculi, the Parallel Random Access Machine model, the Actor model and the Reo Coordination Language.

Concurrent Programming

• Concurrent object-oriented programming is a programming paradigm which combines object-oriented programming (OOP) together with concurrency.

• While numerous programming languages, such as Java, combine OOP with concurrency mechanisms like threads, the phrase "concurrent object-oriented programming" primarily refers to systems where objects themselves are a concurrency primitive, such as when objects are combined with the actor model.

Parallel Computing

• a form of computation in which many calculations are carried out simultaneously, operating on the principle that large problems can often be divided into smaller ones, which are then solved at the same time.

Concurrent vs Parallel

• you can have two threads (or processes) executing concurrently on the same core through context switching.

• When the two threads (or processes) are executed on two different cores (or processors), you have parallelism.

• in concurrency, parallelism is only "virtual", while the other true parallelism.

• Therefore, every parallel program is concurrent, but the converse is not necessarily true

Why Concurrency?• Efficiency

• Time (load sharing)• Cost (resource sharing)• Availability• Multiple access• Convenience

• • Perform several tasks at once• Modeling power• Describing systems that are inherently parallel

Real World Example• Computer systems are used for modeling

objects in the real world• Object-oriented programming

• The world often includes parallel operation• example:

• Limited number of seats on the same plane• Several booking agents active at the same

time

Terms• Multiprocessing

• the use of more than one processing unit in a system

• Parallel execution• processes running at the same time

Terms Cont.d• Interleaving

• several tasks active, only one running at a time

• Multitasking• the OS runs interleaved executions

• Concurrency• multiprocessing, multitasking, or any

combination

Moore’s Law

Why it matters?• The “end of Moore’s law as we knew it” has important

implications on the software construction process• Computing is taking an irreversible step toward parallel

architectures• Hardware construction of ever faster sequential CPUs has hit

physical limits• Clock speed no longer increases for every new processor

generation• Moore’s Law expresses itself as exponentially increasing

number of processing cores per chip• If we want programs to run faster on the next processor

generation, the software must exploit more concurrency

Amdahl’s Law

Amdahl’s Law Cont.d• We go from 1 processor to n. What gain may we

expect?• Amdahl’s law severely limits our hopes!• Define gain as:• speed up = ( 1 / (( 1-p)+( p/n)) )• where p = parallelizable %; n= number of

processors• Not everything can be parallelized!

Types of Parallel Computation

• Flynn’s taxonomy: classification of computer architectures

• Considers relationship of instruction streams to data streams:• SISD: No parallelism (uniprocessor)• SIMD: Vector processor, GPU• MIMD: Multiprocessing (predominant

today)

MIMD Variants• SPMD (Single Program Multiple Data):

• All processors run same program, but at independent speeds; no lockstep as in SIMD

• MPMD (Multiple Program Multiple Data):• Often manager/worker strategy: manager

distributes tasks, workers return result to manager

Shared Memory Model

• All processors share a common memory • Shared-memory communication

Memory

Processor 1

Processor 2

Processor 4

Processor 3

Distributed Memory Model

• Each processor has own local memory, inaccessible to others

• Message passing communication• Common for SPMD architecture

Processor

Processor

Processor

Memory

Memory

Memory

Message Passing

Client Server Model• Specific case of the distributed model• Examples: Database-centered systems, World-

Wide Web

Processor

Processor

Processor

Memory

Memory

Memory

SCOOP Mechanism• Simple Concurrent Object-Oriented Programming• Evolved through previous two decades; CACM (1993) and

chap. 32 of Object-Oriented Software Construction, 2nd edition, 1997

• Prototype-implementation at ETH in 2007• Implementation integrated within EiffelStudio in 2011 (by

Eiffel Software)• Current reference: ETH PhD Thesis by Piotr Nienaltowski,

2008; articles by Benjamin Morandi, Sebastian Nanz and Bertrand Meyer, 2010-2011

SCOOP Preview: a sequential programtransfer (source, target: ACCOUNT;

amount: INTEGER)-- If possible, transfer amount from source to target.do

if source.balance >= amount thensource.withdraw (amount)target.deposit (amount)

endendTypical calls:

transfer (acc1, acc2, 100)transfer (acc1, acc3, 100)

In a concurrent setting, using SCOOP

transfer (source, target: separate ACCOUNT; amount: INTEGER)

-- If possible, transfer amount from source to target.do

if source.balance >= amount thensource.withdraw (amount)target.deposit (amount)

endendTypical calls:transfer (acc1, acc2, 100)transfer (acc1, acc3, 100)

A better SCOOP versiontransfer (source, target: separate ACCOUNT;

amount: INTEGER)-- Transfer amount from source to target.require

source.balance >= amountdosource.withdraw (amount)target.deposit (amount)ensuresource.balance = old source.balance –

amounttarget.balance = old target.balance +

amountend

Dining Philosophers

• find out about it…

Programming: Then & Now

Sequential Programming

• Used to be messy• Still hard but key improvements:

• Structured programming• Data abstraction & object technology• Design by Contract• Genericity, multiple inheritance• Architectural techniques

Concurrent Programming

• Used to be messy• Example: threading models in most popular

approaches• Development level: sixties/seventies• Only understandable through operational

reasoning

References

• http://www.javaworld.com/article/2078679/java-concurrency/java-concurrency-modern-threading-for-not-quite-beginners.html

• https://books.google.mv/books?id=-x1S4neCSOYC&pg=PA5&lpg=PA5&dq=what+is+using+concurrency+constructs&source=bl&ots=Dxdk47Ddy-&sig=ZxzRD163royyROmEOF2lAtSZBz8&hl=en&sa=X&ved=0CEkQ6AEwBmoVChMIx-rc1ZbYxwIVQQeOCh3LkguH#v=onepage&q=what%20is%20using%20concurrency%20constructs&f=false

• http://book.realworldhaskell.org/read/concurrent-and-multicore-programming.html• https://books.google.mv/books?id=mHozgJ7ngq0C&pg=PA1&lpg=PA1&dq=what+is

+using+concurrency+constructs&source=bl&ots=TOT_g-IESw&sig=xSG7JzGquRD9NZmnbkt6tv6Vr2A&hl=en&sa=X&ved=0CFcQ6AEwCWoVChMIx-rc1ZbYxwIVQQeOCh3LkguH#v=onepage&q=what%20is%20using%20concurrency%20constructs&f=false

• http://gee.cs.oswego.edu/dl/cpj/mechanics.html• Read stack overflow / java documentation / etc

Next Up…

• Concurrent Programming…

Sachintha Gunasena MBCShttp://lk.linkedin.com/in/sachinthadtg

Thank you.

Sachintha Gunasena MBCShttp://lk.linkedin.com/in/sachinthadtg