Table of Contents for
Mastering C++ Multithreading

Version ebook / Retour

Cover image for bash Cookbook, 2nd Edition Mastering C++ Multithreading by Maya Posch Published by Packt Publishing, 2017
  1. Mastering C++ Multithreading
  2. Title Page
  3. Copyright
  4. Mastering C++ Multithreading
  5. Credits
  6. About the Author
  7. About the Reviewer
  8. www.PacktPub.com
  9. Why subscribe?
  10. Customer Feedback
  11. Table of Contents
  12. Preface
  13. What this book covers
  14. What you need for this book
  15. Who this book is for
  16. Conventions
  17. Reader feedback
  18. Downloading the example code
  19. Errata
  20. Piracy
  21. Questions
  22. Revisiting Multithreading
  23. Getting started
  24. The multithreaded application
  25. Makefile
  26. Other applications
  27. Summary
  28. Multithreading Implementation on the Processor and OS
  29. Defining processes and threads
  30. Tasks in x86 (32-bit and 64-bit)
  31. Process state in ARM
  32. The stack
  33. Defining multithreading
  34. Flynn's taxonomy
  35. Symmetric versus asymmetric multiprocessing
  36. Loosely and tightly coupled multiprocessing
  37. Combining multiprocessing with multithreading
  38. Multithreading types
  39. Temporal multithreading
  40. Simultaneous multithreading (SMT)
  41. Schedulers
  42. Tracing the demo application
  43. Mutual exclusion implementations
  44. Hardware
  45. Software
  46. Summary
  47. C++ Multithreading APIs
  48. API overview
  49. POSIX threads
  50. Windows support
  51. PThreads thread management
  52. Mutexes
  53. Condition variables
  54. Synchronization
  55. Semaphores
  56. Thread local storage (TLC)
  57. Windows threads
  58. Thread management
  59. Advanced management
  60. Synchronization
  61. Condition variables
  62. Thread local storage
  63. Boost
  64. Qt
  65. QThread
  66. Thread pools
  67. Synchronization
  68. QtConcurrent
  69. Thread local storage
  70. POCO
  71. Thread class
  72. Thread pool
  73. Thread local storage (TLS)
  74. Synchronization
  75. C++ threads
  76. Putting it together
  77. Summary
  78. Thread Synchronization and Communication
  79. Safety first
  80. The scheduler
  81. High-level view
  82. Implementation
  83. Request class
  84. Worker class
  85. Dispatcher
  86. Makefile
  87. Output
  88. Sharing data
  89. Using r/w-locks
  90. Using shared pointers
  91. Summary
  92. Native C++ Threads and Primitives
  93. The STL threading API
  94. Boost.Thread API
  95. The 2011 standard
  96. C++14
  97. C++17
  98. STL organization
  99. Thread class
  100. Basic use
  101. Passing parameters
  102. Return value
  103. Moving threads
  104. Thread ID
  105. Sleeping
  106. Yield
  107. Detach
  108. Swap
  109. Mutex
  110. Basic use
  111. Non-blocking locking
  112. Timed mutex
  113. Lock guard
  114. Unique lock
  115. Scoped lock
  116. Recursive mutex
  117. Recursive timed mutex
  118. Shared mutex
  119. Shared timed mutex
  120. Condition variable
  121. Condition_variable_any
  122. Notify all at thread exit
  123. Future
  124. Promise
  125. Shared future
  126. Packaged_task
  127. Async
  128. Launch policy
  129. Atomics
  130. Summary
  131. Debugging Multithreaded Code
  132. When to start debugging
  133. The humble debugger
  134. GDB
  135. Debugging multithreaded code
  136. Breakpoints
  137. Back traces
  138. Dynamic analysis tools
  139. Limitations
  140. Alternatives
  141. Memcheck
  142. Basic use
  143. Error types
  144. Illegal read / illegal write errors
  145. Use of uninitialized values
  146. Uninitialized or unaddressable system call values
  147. Illegal frees
  148. Mismatched deallocation
  149. Overlapping source and destination
  150. Fishy argument values
  151. Memory leak detection
  152. Helgrind
  153. Basic use
  154. Misuse of the pthreads API
  155. Lock order problems
  156. Data races
  157. DRD
  158. Basic use
  159. Features
  160. C++11 threads support
  161. Summary
  162. Best Practices
  163. Proper multithreading
  164. Wrongful expectations - deadlocks
  165. Being careless - data races
  166. Mutexes aren't magic
  167. Locks are fancy mutexes
  168. Threads versus the future
  169. Static order of initialization
  170. Summary
  171. Atomic Operations - Working with the Hardware
  172. Atomic operations
  173. Visual C++
  174. GCC
  175. Memory order
  176. Other compilers
  177. C++11 atomics
  178. Example
  179. Non-class functions
  180. Example
  181. Atomic flag
  182. Memory order
  183. Relaxed ordering
  184. Release-acquire ordering
  185. Release-consume ordering
  186. Sequentially-consistent ordering
  187. Volatile keyword
  188. Summary
  189. Multithreading with Distributed Computing
  190. Distributed computing, in a nutshell
  191. MPI
  192. Implementations
  193. Using MPI
  194. Compiling MPI applications
  195. The cluster hardware
  196. Installing Open MPI
  197. Linux and BSDs
  198. Windows
  199. Distributing jobs across nodes
  200. Setting up an MPI node
  201. Creating the MPI host file
  202. Running the job
  203. Using a cluster scheduler
  204. MPI communication
  205. MPI data types
  206. Custom types
  207. Basic communication
  208. Advanced communication
  209. Broadcasting
  210. Scattering and gathering
  211. MPI versus threads
  212. Potential issues
  213. Summary
  214. Multithreading with GPGPU
  215. The GPGPU processing model
  216. Implementations
  217. OpenCL
  218. Common OpenCL applications
  219. OpenCL versions
  220. OpenCL 1.0
  221. OpenCL 1.1
  222. OpenCL 1.2
  223. OpenCL 2.0
  224. OpenCL 2.1
  225. OpenCL 2.2
  226. Setting up a development environment
  227. Linux
  228. Windows
  229. OS X/MacOS
  230. A basic OpenCL application
  231. GPU memory management
  232. GPGPU and multithreading
  233. Latency
  234. Potential issues
  235. Debugging GPGPU applications
  236. Summary

Output

After compiling the application, running it produces the following output for the 50 total requests:

    $ ./dispatcher_demo.exe
Initialised.
Starting processing request 1...
Starting processing request 2...
Finished request 1
Starting processing request 3...
Finished request 3
Starting processing request 6...
Finished request 6
Starting processing request 8...
Finished request 8
Starting processing request 9...
Finished request 9
Finished request 2
Starting processing request 11...
Finished request 11
Starting processing request 12...
Finished request 12
Starting processing request 13...
Finished request 13
Starting processing request 14...
Finished request 14
Starting processing request 7...
Starting processing request 10...
Starting processing request 15...
Finished request 7
Finished request 15
Finished request 10
Starting processing request 16...
Finished request 16
Starting processing request 17...
Starting processing request 18...
Starting processing request 0...

At this point, we can already clearly see that even with each request taking almost no time to process, the requests are clearly being executed in parallel. The first request (request 0) only starts being processed after the sixteenth request, while the second request already finishes after the ninth request, long before this.

The factors which determine which thread, and thus, which request is processed first depends on the OS scheduler and hardware-based scheduling as described in chapter 2, Multithreading Implementation on the Processor and OS. This clearly shows just how few assumptions can be made about how a multithreaded application will be executed even on a single platform.

    Starting processing request 5...
Finished request 5
Starting processing request 20...
Finished request 18
Finished request 20
Starting processing request 21...
Starting processing request 4...
Finished request 21
Finished request 4

In the preceding code, the fourth and fifth requests also finish in a rather delayed fashion.

    
Starting processing request 23...
Starting processing request 24...
Starting processing request 22...
Finished request 24
Finished request 23
Finished request 22
Starting processing request 26...
Starting processing request 25...
Starting processing request 28...
Finished request 26
Starting processing request 27...
Finished request 28
Finished request 27
Starting processing request 29...
Starting processing request 30...
Finished request 30
Finished request 29
Finished request 17
Finished request 25
Starting processing request 19...
Finished request 0

At this point, the first request finally finishes. This may indicate that the initialization time for the first request will always be delayed as compared to the successive requests. Running the application multiple times can confirm this. It's important that if the order of processing is relevant, this randomness does not negatively affect one's application.

    Starting processing request 33...
Starting processing request 35...
Finished request 33
Finished request 35
Starting processing request 37...
Starting processing request 38...
Finished request 37
Finished request 38
Starting processing request 39...
Starting processing request 40...
Starting processing request 36...
Starting processing request 31...
Finished request 40
Finished request 39
Starting processing request 32...
Starting processing request 41...
Finished request 32
Finished request 41
Starting processing request 42...
Finished request 31
Starting processing request 44...
Finished request 36
Finished request 42
Starting processing request 45...
Finished request 44
Starting processing request 47...
Starting processing request 48...
Finished request 48
Starting processing request 43...
Finished request 47
Finished request 43
Finished request 19
Starting processing request 34...
Finished request 34
Starting processing request 46...
Starting processing request 49...
Finished request 46
Finished request 49
Finished request 45

Request 19 also became fairly delayed, showing once again just how unpredictable a multithreaded application can be. If we were processing a large dataset in parallel here, with chunks of data in each request, we might have to pause at some points to account for these delays, as otherwise, our output cache might grow too large.

As doing so would negatively affect an application's performance, one might have to look at low-level optimizations, as well as the scheduling of threads on specific processor cores in order to prevent this from happening.

    Stopped workers.
Joined threads.
Joined threads.
Joined threads.
Joined threads.
Joined threads.
Joined threads.
Joined threads.
Joined threads.
Joined threads.
Joined threads.
Clean-up done.

All 10 worker threads which were launched in the beginning terminate here as we call the stop() function of the Dispatcher.