DOCSLIB.ORG

Object-Oriented Programming

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Object-Oriented Programming CSE 143 Introduction •Classes have given us tremendous benefit •Encapsulation: make details of implementation and representation private Object-Oriented •Classes help break down and organize the task of Programming writing a program •Classes are designed to enable other benefits [Chapter 8] •Inheritance (subclassing) •Dynamic dispatch (polymorphism) •Two very powerful programming tools! •They help us write less code 2/20/99 336 2/20/99 337 Classes Using Other Classes Hierarchies of Organization •We've seen one way that a class can make use of •Often, we classify things in a hierarchy from another class general to specific •Use an instance of that class as a member variable Animal class PolyLine Mammal Fish Reptile { Canine Feline private: Tuna Shark PointArray pa; }; Wolf Dog Cat Iguana Croc •Objects have more of a "is-a-kind-of" relationship •We might call this a "has-a" relationship • • A Dog "is-a-kind-of" Canine, a Shark "is-a-kind-of" A PolyLine "has-a" PointArray Animal •This is often just called a “is-a” relation 2/20/99 338 2/20/99 339 Inheritance Example: A Point Class •Inheritance is a way to encode the "is-a-kind-of" •Let's say we had the relation in OO languages following class • class Point •Shark declares that it "is-a-kind-of" Fish by inheriting We can use inheritance to { create a class of colored public: from it Point( double x, double y ); • points based on this class A derived class inherits from a base class by putting double getX(); ": public BaseClassName" in the class declaration double getY(); GHULYHGÃFODVV EDVHÃFODVV void print( ostream& os ); RUÃVXEFODVV RUÃVXSHUFODVV private: class Shark : public Fish { double xpos; // Shark-specific stuff here double ypos; }; }; 2/20/99 340 2/20/99 341 CSE 143 1 ColorPoint with inheritance Rules of Inheritance class ColorPoint : public Point • { All data and methods in base class (superclass) • public: ColorPoint "is-a" Point ColorPoint( double x, double y, are automatically inherited by derived (sub) class • Therefore ColorPoint has to be Color c ); •Changes in base class are automatically able to do anything Point can // getX() is inherited from Point • All fields and methods of Point // getY() is inherited from Point propagated into derived classes are "inherited" by ColorPoint - // New accessor method for the •Public members of base class visible to derived they are transparently included! // Color field Color getColor(); class and clients that use it • Derived class can add new // We still need to redefine methods, fields // the print method! •Private members of base class still not visible to • Derived class can override base void print( ostream& os ); derived class or clients class behaviour private: // xpos is inherited from Point •Can also declare “protected” members in base // ypos is inherited from Point Color color; class which are visible in derived class, but not }; visible to clients. 2/20/99 342 2/20/99 343 ColorPoint Implementation ColorPoint Client ColorPoint::ColorPoint( double x, double y, Color c ) : Point( x, y ) { Point p( 1.0, 0.0 ); color = c; ColorPoint cp1( 3.14, -45.5, RED ); } // No problem: ColorPoint::print is defined Color ColorPoint::getColor() { cp1.print( cout ); return color; } // No problem: calls Point::getX() and Point::getY() // on Point subset of ColorPoint to access private void ColorPoint::print( ostream& os ) { // xpos and ypos fields os << "(" << getX() << ", " << getY() cout << cp1.getX() << " " << cp1.getY() << endl; << ")/" << color; } // p can refer to any Point, including one that is // actually a ColorPoint. But p can only be used to • // access fields that belong to all Points (i.e., New notation: “: member(values, …)” specifies // can’t be used to access getColor base class constructor to initialize base class p = cp1; fields in derived class object 2/20/99 344 2/20/99 345 Invoking Overriden Methods Scope Resolution •Observation: ColorPoint::print does the same •It turns out that the :: operator allows us to thing as Point::print, and then prints out a Color explicitly call an overriden method from the •So perhaps we can call Point::print from within derived class ColorPoint::print void ColorPoint::print( ostream& os ) { •What happens if we try it this way? Point::print( os ); os << ", " << Color; } void ColorPoint::print( ostream& os ) { •BaseClass::method( arguments ) can be used print( os ); // trying to call print method in superclass BaseClass os << ", " << Color; as long as really is a parent class } (either direct base class or more distant ancestor) 2/20/99 346 2/20/99 347 CSE 143 2 Inheritance and Constructors Substituting Derived Classes •Constructors are not inherited! •Recall that an instance of a •Can’t be, because their name specifies which class derived class can always be substituted for an Point p( 1.0, 9.0 ); they’re part of! ColorPoint cp( 6.0, 7.0, red ); instance of a base class • p = cp; Instead, constructor of base class is called • Derived class guaranteed to automatically before constructor of derived class have (at least) the same data void printPoint( Point pt ) • and interface as base class { Can indicate base class constructor explicitly using the • pt.print( cout ); “:class(arguments)” notation to pass parameters (like in But you may not get the } behaviour you want! ColoredPoint example) printPoint( p ); •If omitted, default constructor of base class is called printPoint( cp ); •Constructors are called in "inside-out" order 2/20/99 348 2/20/99 349 Pointers And Inheritance Static And Dynamic Types •You can also substitute a •In C++, every variable has a static and a dynamic pointer to a derived class type for a pointer to a base Point *pptr = new Point( 1.0, 9.0 ); ColorPoint *cpptr = •Static type is declared type of variable class new ColorPoint( 6.0, 7.0, red ); Every variable has a single static type that never changes • There's still that guarantee Point *fooptr = cpptr; •Dynamic type is type of object the variable actually about data and interface void printPoint( Point *ptr ) • Also holds for reference types { contains or refers to ofstream ofs( "point.out" ); Dynamic type can change during the program! • No information disappears!! ptr->print( ofs ); • ofs.close(); •Up to now, these have always been identical Unfortunately, we still have } the same problems… •But not any more! printPoint( pptr ); printPoint( cpptr ); Point *myPointPointer = new ColorPoint( 3.14, 2.78, green ); 2/20/99 350 2/20/99 351 Static Dispatch Dynamic Dispatch •"Dispatching" is the act of deciding which piece of •C++ has a mechanism for declaring individual code to execute when a method is called methods as dynamically dispatched •Static dispatch means that the decision is made •If an overriding function exists, call it statically, i.e. at compile time •In base class, label the function with virtual •Decision made based on static type of receiver keyword Point *myPointPointer = new ColorPoint( 3.14, 2.78, green ); •Overriding versions in subclasses may or may not have // myPointPointer is a Point*, so call Point::print the virtual keyword; but use consistently for better style myPointPointer->print( cout ); •Rule of thumb: If you may ever need to override a •Idea: make the decision at runtime, based on the method, make it virtual! dynamic type of the object •Even safer rule: Unless you have a good reason not to, member functions in a class hierarchy should be virtual. 2/20/99 352 2/20/99 353 CSE 143 3 Example Of Dynamic Dispatch Dynamically-Dispatched Calls Point *p = new ColorPoint( 3.13, 5.66, ochre ); class Point { p->print( cout ); public: virtual void print( ostream& os ); … •The compiler notices that Point::print is defined }; as virtual class ColorPoint : public Point { •Instead of just calling Point::print, it inserts public: virtual void print( ostream& os ); extra code to look at information attached to the … new }; object by to decide what function to call •This is slightly slower than static dispatch Point *p = new ColorPoint( 3.13, 5.66, ochre ); •Almost always too minor a speed penalty to worry about p->print( cout ); // calls ColorPoint::print( ) 2/20/99 354 2/20/99 355 A User-Interface Example Abstract Classes •Many user interface class Widget { • public: Some classes are so abstract that instances of toolkits are void getPosition( int& x, int& y ); them shouldn't even exist implemented as class virtual void draw( Screen& theScreen ); … •What does it mean to have an instance of Widget? of hierarchies private: Animal? •Helps manage int xpos; int ypos; • complexity }; An abstract class is one that should not or can not be instantiated - it only defines an interface class Button : public Widget { • public: • Question: how should virtual void draw( Screen& theScreen ); A concrete class can have instances we implement … • }; It may not make sense to attempt to fully Widget::draw? implement all functions in an abstract class class Toolbar : public Widget { public: •What should Widget::draw do? virtual void draw( Screen& theScreen ); }; 2/20/99 356 2/20/99 357 Pure Virtual Functions Summary •A “pure virtual” function is not implemented in the base class - must implement in derived class •Object-Oriented Programming •Syntax: append "= 0" to base method declaration •Inheritance class Widget { •Use for "is-a-kind-of", not "has-a" relations public: virtual void draw( Screen& theScreen ) = 0; •Classification hierarchies }; •Base class (superclass), derived class (subclass) •Compiler guarantees that class with pure virtual •Overriding functions functions cannot be instantiated • • Static and dynamic types If you call a pure virtual function, you’ll use the • version from some derived class Dynamic Dispatch • Widget *w = new Button(); Virtual functions w->draw( myScreen ); •Abstract classes, pure virtual functions 2/20/99 358 2/20/99 359 CSE 143 4.
Load more

Recommended publications
  • Behavioral Subtyping, Specification Inheritance, and Modular Reasoning Gary T
    Computer Science Technical Reports Computer Science 9-3-2006 Behavioral Subtyping, Specification Inheritance, and Modular Reasoning Gary T. Leavens Iowa State University David A. Naumann Iowa State University Follow this and additional works at: http://lib.dr.iastate.edu/cs_techreports Part of the Software Engineering Commons Recommended Citation Leavens, Gary T. and Naumann, David A., "Behavioral Subtyping, Specification Inheritance, and Modular Reasoning" (2006). Computer Science Technical Reports. 269. http://lib.dr.iastate.edu/cs_techreports/269 This Article is brought to you for free and open access by the Computer Science at Iowa State University Digital Repository. It has been accepted for inclusion in Computer Science Technical Reports by an authorized administrator of Iowa State University Digital Repository. For more information, please contact [email protected]. Behavioral Subtyping, Specification Inheritance, and Modular Reasoning Abstract Behavioral subtyping is an established idea that enables modular reasoning about behavioral properties of object-oriented programs. It requires that syntactic subtypes are behavioral refinements. It validates reasoning about a dynamically-dispatched method call, say E.m(), using the specification associated with the static type of the receiver expression E. For languages with references and mutable objects the idea of behavioral subtyping has not been rigorously formalized as such, the standard informal notion has inadequacies, and exact definitions are not obvious. This paper formalizes behavioral subtyping and supertype abstraction for a Java-like sequential language with classes, interfaces, exceptions, mutable heap objects, references, and recursive types. Behavioral subtyping is proved sound and semantically complete for reasoning with supertype abstraction. Specification inheritance, as used in the specification language JML, is formalized and proved to entail behavioral subtyping.
    [Show full text]
  • Advances in Programming Languages Lecture 18: Concurrency and More in Rust
    Advances in Programming Languages Lecture 18: Concurrency and More in Rust Ian Stark School of Informatics The University of Edinburgh Friday 24 November 2016 Semester 1 Week 10 https://blog.inf.ed.ac.uk/apl16 Topic: Programming for Memory Safety The final block of lectures cover features used in the Rust programming language. Introduction: Zero-Cost Abstractions (and their cost) Control of Memory: Ownership Concurrency and more This section of the course is entirely new — Rust itself is not that old — and I apologise for any consequent lack of polish. Ian Stark Advances in Programming Languages / Lecture 18: Concurrency and More in Rust 2016-11-24 The Rust Programming Language The Rust language is intended as a tool for safe systems programming. Three key objectives contribute to this. Zero-cost abstractions Basic References Memory safety https://www.rust-lang.org https://blog.rust-lang.org Safe concurrency The “systems programming” motivation resonates with that for imperative C/C++. The “safe” part draws extensively on techniques developed for functional Haskell and OCaml. Sometimes these align more closely than you might expect, often by overlap between two aims: Precise control for the programmer; Precise information for the compiler. Ian Stark Advances in Programming Languages / Lecture 18: Concurrency and More in Rust 2016-11-24 Previous Homework Watch This Peter O’Hearn: Reasoning with Big Code https://is.gd/reasoning_big_code Talk at the Alan Turing Institute about how Facebook is using automatic verification at scale to check code and give feedback to programmers. Ian Stark Advances in Programming Languages / Lecture 18: Concurrency and More in Rust 2016-11-24 Guest Speaker ! Probabilistic Programming: What is it and how do we make it better? Maria Gorinova University of Edinburgh 3.10pm Tuesday 29 November 2016 Ian Stark Advances in Programming Languages / Lecture 18: Concurrency and More in Rust 2016-11-24 Optional: Review and Exam Preparation ! The final lecture will be an exam review, principally for single-semester visiting students.
    [Show full text]
  • Developing Rust on Bare-Metal
    RustyGecko - Developing Rust on Bare-Metal An experimental embedded software platform Håvard Wormdal Høiby Sondre Lefsaker Master of Science in Computer Science Submission date: June 2015 Supervisor: Magnus Lie Hetland, IDI Co-supervisor: Antonio Garcia Guirado, IDI Marius Grannæs, Silicon Labs Norwegian University of Science and Technology Department of Computer and Information Science Preface This report is submitted to the Norwegian University of Science and Technology in fulfillment of the requirements for master thesis. This work has been performed at the Department of Computer and Information Science, NTNU, with Prof. Magnus Lie Hetland as the supervisor, Antonio Garcia Guirado (ARM), and Marius Grannæs (Silicon Labs) as co-supervisors. The initial problem description was our own proposal and further developed in co-operation with our supervisors. Acknowledgements Thanks to Magnus Lie Hetland, Antonio Garcia Guirado, and Marius Grannæs for directions and guidance on technical issues and this report. Thanks to Silicon Labs for providing us with development platforms for us to develop and test our implementation on. Thanks to Antonio Garcia Guirado for the implementation of the CircleGame application for us to port to Rust and use in our benchmarks. Thanks to Itera w/Tommy Ryen for office space. A special thanks to Leslie Ho and Siri Aagedal for all the support and for proofreading the thesis. Sondre Lefsaker and H˚avard Wormdal Høiby 2015-06-14 i Project Description The Rust programming language is a new system language developed by Mozilla. With the language being statically compiled and built on the LLVM compiler infras- tructure, and because of features like low-level control and zero-cost abstractions, the language is a candidate for use in bare-metal systems.
    [Show full text]
  • Xtend User Guide
    Xtend User Guide September 10, 2013 Contents I. Getting Started 5 1. Introduction 6 2. Hello World 7 3. The Movies Example 9 3.1. The Data . .9 3.2. Parsing The Data . 10 3.3. Answering Some Questions . 11 3.3.1. Question 1 : What Is The Number Of Action Movies? . 11 3.3.2. Question 2 : What Is The Year The Best Movie From The 80's Was Released? . 12 3.3.3. Question 3 : What Is The The Sum Of All Votes Of The Top Two Movies?................................. 13 II. Reference Documentation 14 4. Java Interoperability 15 4.1. Type Inference . 15 4.2. Conversion Rules . 15 4.3. Interoperability with Java . 16 5. Classes and Members 17 5.1. Package Declaration . 17 5.2. Imports . 17 5.3. Class Declaration . 18 5.4. Constructors . 19 5.5. Fields . 19 5.6. Methods . 20 5.6.1. Abstract Methods . 20 5.6.2. Overriding Methods . 21 5.6.3. Declared Exceptions . 21 5.6.4. Inferred Return Types . 21 5.6.5. Generic Methods . 22 2 5.6.6. Dispatch Methods . 22 5.7. Annotations . 26 5.8. Extension Methods . 26 5.8.1. Extensions from the Library . 27 5.8.2. Local Extension Methods . 28 5.8.3. Extension Imports . 28 5.8.4. Extension Provider . 29 5.9. Interface Declaration . 29 5.10. Annotation Type Declaration . 30 5.11. Enum Type Declaration . 30 6. Expressions 31 6.1. Literals . 31 6.1.1. String Literals . 31 6.1.2. Character Literals . 32 6.1.3. Number Literals .
    [Show full text]
  • The Rust Programming Language
    The Rust Programming Language The Rust Team 2016-10-01 2 Contents 1 Introduction 11 Contributing ...................................... 11 2 Getting Started 13 Installing Rust ..................................... 13 Hello, world! ...................................... 16 Hello, Cargo! ...................................... 19 Closing Thoughts .................................... 23 3 Tutorial: Guessing Game 25 Set up .......................................... 25 Processing a Guess ................................... 26 Generating a secret number .............................. 30 Comparing guesses ................................... 33 Looping ......................................... 37 Complete! ........................................ 41 4 Syntax and Semantics 43 Variable Bindings ....................................... 43 Patterns ......................................... 43 Type annotations .................................... 44 Mutability ........................................ 44 Initializing bindings .................................. 45 Scope and shadowing .................................. 46 Functions ........................................... 48 Primitive Types ........................................ 53 Booleans ......................................... 53 4 CONTENTS char ........................................... 53 Numeric types ..................................... 54 Arrays .......................................... 55 Slices ........................................... 56 str ...........................................
    [Show full text]
  • COMP322 - Introduction to C++
    COMP322 - Introduction to C++ Lecture 09 - Inheritance continued Robert D. Vincent School of Computer Science 10 March 2010 Recall from last time I Inheritance describes the creation of derived classes from base classes. I Derived classes inherit all data and functions from their base classes, while certain functions such as constructors are handled specially. I Polymorphism is a product of the combination of virtual functions and C++ assignment compatibility rules. I We can create abstract types which contain “pure” functions that may have no implementation. We can’t actually create objects from these types, but we can create references or pointers to abstract types. Static dispatch If a member function is not virtual, the choice of function to call is made at compile time: class A{ int f(); }; class B: public A{ int f(); }; int main() { B b; A *pa = &b; pa->f(); // Calls A::f() because pa is of type 'A *' } This is called either “static dispatch” or “static binding”, and it is the default behavior in C++. Dynamic dispatch If a member function is virtual, the choice of function to call is made at run time: class A{ virtual int f(); }; class B: public A{ int f(); }; int main() { B b; A *pa = &b; pa->f(); // Calls B::f() because pa points to a 'B *' } Called either “dynamic dispatch” or “run-time binding”, this is both more useful and less efficient than static dispatch. Dynamic dispatch internals I Dynamic dispatch is implemented by adding a layer of indirection to a function call. I Any object with dynamic dispatch contains an automatically-generated pointer to a vtable.
    [Show full text]
  • Virtual Functions
    Virtual Functions OK, today’s topic in Object Oriented C++ is function dispatch. First, let’s review overloading. What is function overloading? Two methods or functions with the same signature. Why do you overload a function? For polymorphism. Now, mechanically, how do you overload something? Not only the method name, but the number and type of arguments and the return type must all match. Why? The name and arguments so that you know it’s an overload; the return type so that the compiler always knows the data types at run time… OK, so let’s create an example of overloading. Let’s say I have an Animal class with a “warm up” method. It might look like: Class Animal { Public: bool Warm-up() const {return false;} void Behave() { if (cold()) warm-up(); } }; Notice that is returns false, because there is no way for a generic Animal to warm up. But if we get more specific… Class Mammal : public Animal { Public: Bool Warm-up() const {shiver(); return true;} }; Notice that this is an example of function overloading, because the name, arguments and return type all match. Now, what happens if I write the following code? Mammal m; m.Behave(); In particular, which version of Behave() gets called? Notice that m is a mammal, but Behave() is a method of Animal, so inside Animal() the declared type of the object is Animal, not Mammal. If I did this example in Java, the answer is that the mammal would shiver and return true. But as written, in C++ the animal does nothing and Warm_up returns false.
    [Show full text]
  • Object Oriented Programming Introduction
    Object Oriented Programming Introduction Object Oriented Programming Introduction to Computer Science II Definition Christopher M. Bourke Objected-Oriented Programming is a programming paradigm that uses [email protected] objects (data structures which have data fields and methods) which interact together within a computer program. Object Oriented Programming Object Oriented Programming Key Ideas Keep in mind When talking about OOP, keep in mind: I Modularity of design I Definitions are often blurred, authors/contexts may differ slightly I Encapsulation of data into Objects I Languages can have some, all, or none OOP features I Abstraction of an object’s behavior and representation I Programs can ignore or utilize some subset of features I Hierarchical organization of objects with subtypes inheriting behavior and state I Non-OOP languages can support some (or all!) OOP features without being “OOP” I Polymorphism ability of objects to have multiple forms in different contexts at different times I Debate still going as to “best” practices, how to do OOP, etc. I Understanding some concepts requires understanding other concepts Object Oriented Programming MotivationI Advantages Consider the evolution of computer languages I Facilitates modeling of real-world agents & problems I Machine Code Assembly Imperative Language I Supports modularity & decomposition → → I Clear demarcation of responsibilities and functionality I Each iteration still forces you to think in terms of a computer’s structure I Separates interface (the what) from implementation
    [Show full text]
  • Multimethods
    11-A1568 01/23/2001 12:39 PM Page 263 11 Multimethods This chapter defines, discusses, and implements multimethods in the context of Cϩϩ . The Cϩϩ virtual function mechanism allows dispatching a call depending on the dynamic type of one object. The multimethods feature allows dispatching a function call depending on the types of multiple objects. A universally good implementation requires language support, which is the route that languages such as CLOS, ML, Haskell, and Dylan have taken. Cϩϩ lacks such support, so its emulation is left to library writers. This chapter discusses some typical solutions and some generic implementations of each. The solutions feature various trade-offs in terms of speed, flexibility, and dependency management. To describe the technique of dispatching a function call depending on mul- tiple objects, this book uses the terms multimethod (borrowed from CLOS) and multiple dis- patch. A particularization of multiple dispatch for two objects is known as double dispatch. Implementing multimethods is a problem as fascinating as dreaded, one that has stolen lots of hours of good, healthy sleep from designers and programmers.1 The topics of this chapter include • Defining multimethods • Identifying situations in which the need for multiobject polymorphism appears • Discussing and implementing three double dispatchers that foster different trade-offs • Enhancing double-dispatch engines After reading this chapter, you will have a firm grasp of the typical situations for which multimethods are the way to go. In addition, you will be able to use and extend several ro- bust generic components implementing multimethods, provided by Loki. This chapter limits discussion to multimethods for two objects (double dispatch).
    [Show full text]
  • Multiple Dispatch and Roles in OO Languages: Ficklemr
    Multiple Dispatch and Roles in OO Languages: FickleMR Submitted By: Asim Anand Sinha [email protected] Supervised By: Sophia Drossopoulou [email protected] Department of Computing IMPERIAL COLLEGE LONDON MSc Project Report 2005 September 13, 2005 Abstract Object-Oriented Programming methodology has been widely accepted because it allows easy modelling of real world concepts. The aim of research in object-oriented languages has been to extend its features to bring it as close as possible to the real world. In this project, we aim to add the concepts of multiple dispatch and first-class relationships to a statically typed, class based langauges to make them more expressive. We use Fickle||as our base language and extend its features in FickleMR . Fickle||is a statically typed language with support for object reclassification, it allows objects to change their class dynamically. We study the impact of introducing multiple dispatch and roles in FickleMR . Novel idea of relationship reclassification and more flexible multiple dispatch algorithm are most interesting. We take a formal approach and give a static type system and operational semantics for language constructs. Acknowledgements I would like to take this opportunity to express my heartful gratitude to DR. SOPHIA DROSSOPOULOU for her guidance and motivation throughout this project. I would also like to thank ALEX BUCKLEY for his time, patience and brilliant guidance in the abscence of my supervisor. 1 Contents 1 Introduction 5 2 Background 8 2.1 Theory of Objects . 8 2.2 Single and Double Dispatch . 10 2.3 Multiple Dispatch . 12 2.3.1 Challenges in Implementation .
    [Show full text]
  • Interface-Based Programming in C++
    Term Paper Daniel Sarnow Interface-based Programming in C++ Fakultät Technik und Informatik Faculty of Engineering and Computer Science Studiendepartment Informatik Department of Computer Science Daniel Sarnow Title of the paper Interface-based Programming in C++ Keywords C++, Compiler Optimizations, Callgrind, Interface, polymorphism Abstract In many object oriented literature runtime polymorphism is often the only way mentioned to realize interface-based programming. In comparison to other languages, C++ oers more than just that. In C++ interface-based programming can also be achieved through link-time or compile-time polymorphism. This paper will show how interface-based programming can be done in C++ when using these three types of polymorphism. The dierent approaches are compared with each other at dierent compiler optimization levels. The paper will also show that compiler optimization mechanisms like inlining and devirtualization have been improved to minimize the overhead caused by dynamic dispatch even more. Contents 1 Introduction1 2 Approaches2 2.1 Runtime Polymorphism . .2 2.2 Link-Time Polymorphism . .4 2.3 Compile-Time Polymorphism . .5 2.3.1 Implicit Interface . .5 2.3.2 Curiously Recurring Template Pattern (CRTP) . .5 3 Comparing the Approaches7 3.1 Runtime Performance . .7 3.1.1 GCC Optimization Flags . 11 3.1.2 Measurement of execution time of the Approaches . 12 3.2 Memory Usage . 13 4 Conclusion 15 iii List of Tables 2.1 Necessity to use a certain interface implementation . .2 3.1 Execution time in milliseconds . 13 4.1 Necessity to use a certain interface implementation (II) . 15 List of Figures 2.1 Class diagram: Dynamic Approach .
    [Show full text]
  • DEPARTMENT of INFORMATICS Integration and Test of RUST Tool
    DEPARTMENT OF INFORMATICS TECHNISCHE UNIVERSITÄT MÜNCHEN Bachelor’s Thesis in Informatics: Games Engineering Integration and Test of RUST Tool Support for MPI Matthias Kübrich DEPARTMENT OF INFORMATICS TECHNISCHE UNIVERSITÄT MÜNCHEN Bachelor’s Thesis in Informatics: Games Engineering Integration and Test of RUST Tool Support for MPI Integration und Test von RUST Tool Unterstützung für MPI Author: Matthias Kübrich Supervisor: Prof. Dr. Martin Schulz Advisor: Bengisu Elis Submission Date: October 15, 2020 I confirm that this bachelor’s thesis in informatics: games engineering is my own work and I have documented all sources and material used. Munich, October 15, 2020 Matthias Kübrich Acknowledgments I would like to thank my advisor Bengisu Elis for all the invaluable help and advice she offered me in writing this thesis, including all the work that led to its completion. I also would like to thank my supervisor Prof. Dr. Martin Schulz, who was quick to suggest this work’s very subject, which turned out to be highly interesting and enjoyable. Abstract With increasingly high performance expectations in modern computer platforms, par- allel computing becomes more and more critical. However, implementing bug-free programs to compute efficiently in parallel is, in general, significantly more demanding than implementing their sequential counterpart. Parallel applications have always been complicated to implement, maintain, or inspect. The Rust programming language aims to provide static programming with zero-cost guarantees for memory-safety. This technology, used to create MPI-Tools, can significantly improve any developers’ workflow creating or maintaining HPC applications. This thesis presents the Rust-Library Rmpi, which provides MPI access with memory safety, a framework implementation for the creation of PMPI -and QMPI-Tools for the programming language Rust, as well as an extensive performance analysis.
    [Show full text]