Parallel Programming with OmniThreadLibrary
Parallel Programming with OmniThreadLibrary
About the Book
Parallel programming is hard. Your best bet to a stable application is a good multithreading framework and for Delphi that means OmniThreadLibrary (www.omnithreadlibrary.com).
This book - which is, incidentally, written by the very author of the OmniThreadLibrary - walks you through different parts of the library. It gives an overview of low-level multithreading approach (programming with tasks), high-level approach (programming with abstractions) and describes various helper functions and data structures implemented in the library.
The book also walks you through multiple examples which demonstrate different aspects of multithreaded programming - from parallel execution of http requests to creation of database connection pools and multithreaded user interfaces.
A printed version is available on Lulu. If you'd like to buy both electronic and paper edition, buy the printed version first as that will give you a coupon for the e-book.
Packages
The Book
PDF
EPUB
WEB
English
The Book + Webinars
This package contains the book and three hour-long webinars on using high-level multithreading abstractions.
Includes:
High-Level Multithreading with OmniThreadLibrary
This is a collection of three one-hour webinars on using high-level multithreading abstractions in the OmniThreadLibrary.
PDF
EPUB
WEB
English
Organisation License (unlimited use)
Unlimited copies for 1) staff members employed by the purchasing organisation, and for educational organisations, 2) for every registered student of the purchasing organisation. Also contains three webinars.
Includes:
High-Level Multithreading with OmniThreadLibrary
This is a collection of three one-hour webinars on using high-level multithreading abstractions in the OmniThreadLibrary.
PDF
EPUB
WEB
English
Table of Contents
- About me
- Credits
-
Introduction
- Formatting conventions
- Learn more
- Release notes
-
1. Introduction to multi-threading
-
1.1 Multi-threading as a source of problems
- 1.1.1 Reading and writing shared data
- 1.1.2 Modifying shared data
- 1.1.3 Writes masquerading as reads
-
1.1 Multi-threading as a source of problems
-
2. Introduction to OmniThreadLibrary
- 2.1 Requirements
- 2.2 License
-
2.3 Installation
- 2.3.1 Installing with GetIt
- 2.3.2 Installing with Delphinus
- 2.3.3 Installing design package
- 2.4 Why use OmniThreadLibrary?
- 2.5 Tasks vs. threads
- 2.6 Locking vs. messaging
-
2.7 Message loop required
- 2.7.1 OmniThreadLibrary and console
- 2.7.2 OmniThreadLibrary task started from another task
- 2.7.3 OmniThreadLibrary task started from a TThread
-
2.8 TOmniValue
- 2.8.1 Data access
- 2.8.2 Type testing
- 2.8.3 Clearing the content
- 2.8.4 Operators
- 2.8.5 Using with generic types
- 2.8.6 Array access
- 2.8.7 Handling records
- 2.8.8 Object ownership
- 2.8.9 Working with TValue
- 2.8.10 Low-level methods
- 2.9 TOmniValueObj
- 2.10 Fluent interfaces
-
3. High-level multi-threading
-
3.1 Introduction
- 3.1.1 A life cycle of an abstraction
- 3.1.2 Anonymous methods, procedures, and methods
- 3.1.3 Pooling
-
3.2 Blocking collection
- 3.2.1 IOmniBlockingCollection
- 3.2.2 Bulk import and export
- 3.2.3 Throttling
- 3.3 Task configuration
-
3.4 Async
- 3.4.1 Handling exceptions
- 3.5 Async/Await
-
3.6 Future
- 3.6.1 IOmniFuture<T> interface
- 3.6.2 Completion detection
- 3.6.3 Cancellation
- 3.6.4 Handling exceptions
- 3.6.5 Examples
-
3.7 Join
- 3.7.1 IOmniParallelJoin interface
- 3.7.2 IOmniJoinState interface
- 3.7.3 Cancellation
- 3.7.4 Handling exceptions
-
3.8 Parallel task
- 3.8.1 IOmniParallelTask interface
- 3.8.2 Example
- 3.8.3 Handling exceptions
- 3.8.4 Examples
-
3.9 Background worker
- 3.9.1 Basics
- 3.9.2 IOmniBackgroundWorker interface
- 3.9.3 Task initialization
- 3.9.4 Work item configuration
- 3.9.5 Work item interface
- 3.9.6 Examples
-
3.10 Pipeline
- 3.10.1 Background
- 3.10.2 Basics
-
3.10.3 IOmniPipeline interface
- 3.10.3.1 Example
- 3.10.4 Generators, mutators, and aggregators
- 3.10.5 Throttling
- 3.10.6 Parallel stages
- 3.10.7 Exceptions
- 3.10.8 Examples
-
3.11 Parallel for
- 3.11.1 IOmniParallelSimpleLoop interface
- 3.11.2 Iterating over an array
- 3.11.3 Examples
-
3.12 ForEach
- 3.12.1 Cooperation
-
3.12.2 Iterating over …
- 3.12.2.1 … Number ranges
- 3.12.2.2 … Enumerable collections
- 3.12.2.3 … Thread-safe enumerable collections
- 3.12.2.4 … Blocking collections
- 3.12.2.5 … Anything
- 3.12.3 Providing external input
- 3.12.4 IOmniParallelLoop interface
- 3.12.5 Preserving output order
- 3.12.6 Aggregation
- 3.12.7 Cancellation
- 3.12.8 Task initialization and finalization
- 3.12.9 Handling exceptions
- 3.12.10 Examples
-
3.13 Fork/Join
- 3.13.1 IOmniForkJoin interface
- 3.13.2 IOmniCompute interface
- 3.13.3 IOmniCompute<T> interface
- 3.13.4 Exceptions
- 3.13.5 Examples
-
3.14 Map
- 3.14.1 IOmniParallelMapper<T1,T2> interface
-
3.15 Timed task
- 3.15.1 IOmniTimedTask interface
-
3.1 Introduction
-
4. Low-level multi-threading
- 4.1 Low-level for the impatient
- 4.2 Four ways to create a task
- 4.3 IOmniTaskControl and IOmniTask interfaces
- 4.4 Task controller needs an owner
- 4.5 Communication subsystem
- 4.6 Processor groups and NUMA nodes
-
4.7 Thread pooling
- 4.7.1 Execution flow
- 4.7.2 IOmniThreadPool interface
- 4.7.3 Task exit code
- 4.7.4 Monitoring thread pool operations
- 4.7.5 Processor groups and NUMA nodes
-
4.8 Lock-free collections
- 4.8.1 Bounded Stack
- 4.8.2 Bounded queue
- 4.8.3 Message queue
- 4.8.4 Dynamic queue
-
4.8.5 Observing lock-free collections
- 4.8.5.1 Examples
- 4.8.6 Benchmarks
- 4.9 Event monitor
-
4.10 Simple tasks
- 4.10.1 Name
- 4.10.2 UniqueID
- 4.10.3 Parameters
- 4.10.4 Termination
- 4.10.5 ExitCode
- 4.10.6 Exceptions
- 4.10.7 Sending messages to a task
- 4.10.8 Receiving messages from a task
- 4.10.9 ChainTo
- 4.10.10 Join / Leave
- 4.10.11 MonitorWith / RemoveMonitor
- 4.10.12 Enforced
- 4.10.13 Unobserved
- 4.10.14 Cancellation token / CancelWith
- 4.10.15 Lock / WithLock
- 4.10.16 WithCounter
- 4.10.17 SetPriority
- 4.10.18 SetQueueSize
-
4.11 TOmniWorker tasks
- 4.11.1 WaitForInit
- 4.11.2 Task
- 4.11.3 Receiving messages
- 4.11.4 RegisterComm
- 4.11.5 Invoke
- 4.11.6 Windows message & APC processing
- 4.11.7 Timers
- 4.11.8 TerminateWhen
- 4.11.9 UserData
- 4.12 Task groups
-
4.13 IOmniTask interface
- 4.13.1 Name and ID
- 4.13.2 Parameters
- 4.13.3 Termination
- 4.13.4 Exit status
- 4.13.5 Exceptions
- 4.13.6 Communication
- 4.13.7 Timers
- 4.13.8 RegisterWaitObject
- 4.13.9 CancellationToken
- 4.13.10 Lock
- 4.13.11 Counter
- 4.13.12 Processor groups and NUMA nodes
- 4.13.13 Internal and obsolete functions
-
5. Synchronization
-
5.1 Critical sections
- 5.1.1 IOmniCriticalSection
- 5.1.2 TOmniCS
-
5.1.3 Locked<T>
- 5.1.3.1 Why not use TMonitor?
- 5.2 TOmniMREW
- 5.3 Cancellation token
- 5.4 Waitable value
- 5.5 Inverse semaphore
-
5.6 Initialization
- 5.6.1 Pessimistic initialization
- 5.6.2 Optimistic initialization
- 5.7 TWaitFor
- 5.8 TOmniLockManager<K>
- 5.9 TOmniSingleThreadUseChecker
-
5.1 Critical sections
-
6. Miscellaneous
- 6.1 TOmniTwoWayChannel
- 6.2 TOmniValueContainer
- 6.3 TOmniCounter
- 6.4 TOmniAlignedInt32 and TOmniAlignedInt64
- 6.5 TOmniRecordWrapper
- 6.6 TOmniRecord
- 6.7 IOmniAutoDestroyObject
- 6.8 IOmniIntegerSet
-
6.9 Environment
- 6.9.1 IOmniAffinity
-
7. How-to
- 7.1 Background file scanning
- 7.2 Web download and database storage
- 7.3 Parallel for with synchronized output
- 7.4 Using taskIndex and task initializer in parallel for
- 7.5 Background worker and list partitioning
- 7.6 Parallel data production
-
7.7 Building a connection pool
- 7.7.1 From theory to practice
- 7.7.2 Running the demo
-
7.8 QuickSort and parallel max
- 7.8.1 QuickSort
- 7.8.2 Parallel max
- 7.9 Parallel search in a tree
-
7.10 Multiple workers with multiple frames
- 7.10.1 The worker
- 7.10.2 The frame
- 7.10.3 The form
-
7.11 OmniThreadLibrary and databases
- 7.11.1 Database model
-
7.11.2 Frame and worker
- 7.11.2.1 Connecting to the database
- 7.11.2.2 Retrieving the data
- 7.11.3 Main program
- 7.12 OmniThreadLibrary and COM/OLE
-
7.13 Using a message queue with a TThread worker
-
7.13.1 Sending data from multiple producers to a single worker
- 7.13.1.1 Initialization and cleanup
- 7.13.1.2 Sending data to the worker
- 7.13.1.3 Receiving the data
-
7.13.2 Sending data from a worker to a form
- 7.13.2.1 Initialization and cleanup
- 7.13.2.2 Sending data to the form
- 7.13.2.3 Receiving the data
-
7.13.1 Sending data from multiple producers to a single worker
- A. Units
- B. Demo applications
- C. Examples
-
D. Hooking into OmniThreadLibrary
- Exception notifications
- Thread notifications
- Pool notifications
-
E. ForEach internals
- Source provider
- Data manager
- Local queue
- Output ordering
- F. Hyperlinks
- Notes
The Leanpub 60 Day 100% Happiness Guarantee
Within 60 days of purchase you can get a 100% refund on any Leanpub purchase, in two clicks.
Now, this is technically risky for us, since you'll have the book or course files either way. But we're so confident in our products and services, and in our authors and readers, that we're happy to offer a full money back guarantee for everything we sell.
You can only find out how good something is by trying it, and because of our 100% money back guarantee there's literally no risk to do so!
So, there's no reason not to click the Add to Cart button, is there?
See full terms...
Earn $8 on a $10 Purchase, and $16 on a $20 Purchase
We pay 80% royalties on purchases of $7.99 or more, and 80% royalties minus a 50 cent flat fee on purchases between $0.99 and $7.98. You earn $8 on a $10 sale, and $16 on a $20 sale. So, if we sell 5000 non-refunded copies of your book for $20, you'll earn $80,000.
(Yes, some authors have already earned much more than that on Leanpub.)
In fact, authors have earnedover $14 millionwriting, publishing and selling on Leanpub.
Learn more about writing on Leanpub
Free Updates. DRM Free.
If you buy a Leanpub book, you get free updates for as long as the author updates the book! Many authors use Leanpub to publish their books in-progress, while they are writing them. All readers get free updates, regardless of when they bought the book or how much they paid (including free).
Most Leanpub books are available in PDF (for computers) and EPUB (for phones, tablets and Kindle). The formats that a book includes are shown at the top right corner of this page.
Finally, Leanpub books don't have any DRM copy-protection nonsense, so you can easily read them on any supported device.
Learn more about Leanpub's ebook formats and where to read them