Intel Threading Building Blocks. Outfitting C++ for Multi-core Processor Parallelism James Reinders

Spis treści książki

• Intel Threading Building Blocks
  • SPECIAL OFFER: Upgrade this ebook with OReilly
  • Foreword
  • Note from the Lead Developer of Intel Threading Building Blocks
  • Preface
    • Assumptions This Book Makes
    • Contents of This Book
    • Conventions Used in This Book
      • Informal Class Declarations
    • Using Code Examples
    • How to Contact Us
    • Acknowledgments
  • 1. Why Threading Building Blocks?
    • Overview
    • Benefits
      • Comparison with Raw Threads and MPI
      • Comparison with OpenMP
      • Recursive Splitting, Task Stealing, and Algorithms
  • 2. Thinking Parallel
    • Elements of Thinking Parallel
    • Decomposition
      • Data Parallelism
      • Task Parallelism
      • Pipelining (Task and Data Parallelism Together)
      • Mixed Solutions
      • Achieving Parallelism
    • Scaling and Speedup
      • How Much Parallelism Is There in an Application?
        • Amdahls Law
        • Gustafsons observations regarding Amdahls Law
        • What did they really say?
        • Serial versus parallel algorithms
    • What Is a Thread?
      • Programming Threads
      • Safety in the Presence of Concurrency
    • Mutual Exclusion and Locks
    • Correctness
    • Abstraction
    • Patterns
    • Intuition
  • 3. Basic Algorithms
    • Initializing and Terminating the Library
    • Loop Parallelization
      • parallel_for
        • Grain size
        • Automatic grain size
        • Notes on automatic grain size
        • parallel_for with partitioner
      • parallel_reduce
        • Advanced example
        • Parallel_reduce with partitioner
      • Advanced Topic: Other Kinds of Iteration Spaces
        • Notes on blocked_range2d
      • parallel_scan
        • Parallel_scan with partitioner
    • Recursive Range Specifications
      • Splittable Concept
    • Recursive Range Specifications
      • Model Types: Splittable Ranges
        • split Class
        • Range Concept
      • Model Types
        • blocked_range<Value> Template Class
        • blocked_range2d Template Class
        • Partitioner Concept
      • Model Types: Partitioners
        • simple_partitioner Class
        • auto_partitioner Class
        • parallel_for<Range,Body> Template Function
        • parallel_reduce<Range,Body> Template Function
        • parallel_scan<Range,Body> Template Function
        • pre_scan_tag and final_scan_tag Classes
    • Summary of Loops
  • 4. Advanced Algorithms
    • Parallel Algorithms for Streams
      • Cook Until Done: parallel_while
        • Notes on parallel_while scaling
          • parallel_while Template Class
      • Working on the Assembly Line: Pipeline
        • Throughput of pipeline
        • Nonlinear pipelines
          • pipeline Class
          • filter Class
      • parallel_sort
        • parallel_sort<RandomAccessIterator, Compare> Template Function
  • 5. Containers
    • concurrent_queue
      • Iterating over a concurrent_queue for Debugging
      • When Not to Use Queues
        • Template Class
    • concurrent_vector
      • concurrent_vector Template Class
    • concurrent_vector
      • Whole Vector Operations
      • Concurrent Operations
      • Parallel Iteration
      • Capacity
      • Iterators
    • concurrent_hash_map
      • More on HashCompare
        • Template Class
      • Whole-Table Operations
      • Concurrent Access
        • const_accessor
        • accessor class
      • Concurrent Operations: find, insert, erase
      • Parallel Iteration
      • Capacity
      • Iterators
  • 6. Scalable Memory Allocation
    • Limitations
    • Problems in Memory Allocation
    • Memory Allocators
      • Which Library to Link into Your Application
      • Using the Allocator Argument to C++ STL Template Classes
    • Replacing malloc, new, and delete
      • Replace malloc, free, realloc, and calloc
      • Replace new and delete
      • Allocator Concept
      • Model Types
        • scalable_allocator<T> Template Class
        • cache_aligned_allocator<T> Template Class
        • aligned_space Template Class
  • 7. Mutual Exclusion
    • When to Use Mutual Exclusion
    • Mutexes
      • Mutex Flavors
      • Reader-Writer Mutexes
      • Upgrade/Downgrade
      • Lock Pathologies
        • Deadlock
        • Convoying and priority inversion
    • Mutexes
      • Mutex Concept
        • mutex Class
        • spin_mutex Class
        • queuing_mutex Class
      • ReaderWriterMutex Concept
        • Model Types
          • spin_rw_mutex Class
          • queuing_rw_mutex Class
    • Atomic Operations
      • Why atomic<T> Has No Constructors
      • Memory Consistency and Fences
        • atomic<T> Template Class
  • 8. Timing
    • tick_count Class
    • tick_count::interval_t Class
  • 9. Task Scheduler
    • When Task-Based Programming Is Inappropriate
    • Much Better Than Raw Native Threads
      • Oversubscription
      • Fair Scheduling
      • High Coding Overhead
      • Load Imbalance
      • Portability
    • Initializing the Library Is Your Job
    • Example Program for Fibonacci Numbers
    • Task Scheduling Overview
    • How Task Scheduling Works
    • Recommended Task Recurrence Patterns
      • Blocking Style with Children
      • Continuation-Passing Style with Children
        • Recycling the parent as the continuation
        • Recycling the parent as a child
    • Making Best Use of the Scheduler
      • Recursive Chain Reaction
      • Continuation Passing
        • Scheduler bypass
        • Recycling
        • Empty tasks
        • Lazy copying
    • Task Scheduler Interfaces
      • task_scheduler_init Class
      • task Class
      • empty_task Class
      • task_list Class
    • Task Scheduler Summary
  • 10. Keys to Success
    • Key Steps to Success
    • Relaxed Sequential Execution
    • Safe Concurrency for Methods and Libraries
    • Debug Versus Release
    • For Efficiencys Sake
    • Enabling Debugging Features
      • The TBB_DO_ASSERT Macro
      • Do Not Ship Your Program Built with TBB_DO_ASSERT
      • The TBB_DO_THREADING_TOOLS Macro
      • Debug Versus Release Libraries
    • Mixing with Other Threading Packages
    • Naming Conventions
      • The tbb Namespace
      • The tbb::internal Namespace
      • The _ _TBB Prefix
  • 11. Examples
    • The Aha! Factor
    • A Few Other Key Points
    • parallel_for Examples
      • ParallelAverage
      • Seismic
      • Matrix Multiply
      • ParallelMerge
      • SubstringFinder
    • The Game of Life
      • Implementation
      • Automaton
      • Automata: Implementation
      • Extending the Application
      • Futher Reading
    • Parallel_reduce Examples
      • ParallelSum
      • ParallelSum without Having to Specify a Grain Size
      • ParallelPrime
    • CountStrings: Using concurrent_hash_map
      • CountStrings: Using concurrent_hash_map
        • Switching from an STL map
    • Quicksort: Visualizing Task Stealing
    • A Better Matrix Multiply (Strassen)
    • Advanced Task Programming
      • Start a Large Task in Parallel with the Main Program
      • Two Mouths: Feeding Two from the Same Task in a Pipeline
    • Packet Processing Pipeline
      • Parallel Programming for an Internet Device
      • Local Network Router Example
      • Pipelined Components for the Local Network Router
        • Network address translation (NAT)
        • Application Layer Gateway (ALG)
        • Packet forwarding
        • Our example
      • Implementation
        • The Threading Building Blocks pipeline
        • Synchronization with the pipeline item and concurrent hash maps
      • Filter Classes
        • Class get_next_packet : public tbb::filter
        • Class output_packet : public tbb::filter
        • Class translator : public tbb::filter
        • Class gateway : public tbb::filter
        • Class forwarding : public tbb::filter
        • Additional reading
    • Memory Allocation
      • Replacing new and delete
      • Replacing malloc, calloc, realloc, and free
    • Game Threading Example
      • Threading Architecture: Physics + Rendering
      • Overview of Keys to Scalability
      • A Frame Loop
      • Domain Decomposition Data Structure Needs
      • Think Tasks, Not Threads
      • Load Balancing Versus Task Stealing
      • Synchronization Between Physics Threads
      • Integrating the Example into a Real Game
      • How to Measure Performance
    • Physics Interaction and Update Code
    • Open Dynamics Engine
      • Look for Hotspots
      • Improving on the First Solution
      • The Code
  • 12. History and Related Projects
    • Libraries
    • Languages
    • Pragmas
    • Generic Programming
      • Concepts in Generic Programming
      • Pseudosignatures in Generic Programming
      • Models in Generic Programming
    • Caches
    • Costs of Time Slicing
    • Quick Introduction to Lambda Functions
    • Further Reading
  • Index
  • About the Author
  • Colophon
  • SPECIAL OFFER: Upgrade this ebook with OReilly

pokaż cały spis treści

ukryj recenzje