Ongoing Optimization: Relaxed Semantic with CppMem

21 August 2016

Contents[Show]

With the relaxed semantic, we have no synchronisations and ordering constraints on atomic operations.

Relaxed semantic

With the relaxed semantic, there is only the atomicity of the operations on atomics left.

// ongoingOptimizationRelaxedSemantic.cpp

#include <atomic>
#include <iostream>
#include <thread>

std::atomic<int> x{0};
std::atomic<int> y{0};

void writing(){  
  x.store(2000,std::memory_order_relaxed);  
  y.store(11,std::memory_order_relaxed);
}

void reading(){  
  std::cout << y.load(std::memory_order_relaxed) << " ";  
  std::cout << x.load(std::memory_order_relaxed) << std::endl;
}

int main(){
  std::thread thread1(writing);
  std::thread thread2(reading);
  thread1.join();
  thread2.join();
};

Now, the questions are very easy to answer. Does the program have well-defined behaviour? Which values for x and y are possible? On the one hand, all operations on x and y are atomic. So the program is well defined. On the other hand, there are no restrictions on the interleavings of the threads. In the end, thread 2 can see the operations on thread 1 in a different order. So this is the first time in our process of ongoing optimizations, that thread 2 can display x == 0 and y == 1. All combinations of x and y are possible.

sukzessiveOptimierungRelaxedSemantikEng

I'm curious, how the graph of CppMem will look like for x == 0 and y == 1?

CppMem

int main(){
  atomic_int x= 0;
  atomic_int y= 0;
  {{{ { 
      x.store(2000, memory_order_relaxed);
      y.store(11,memory_order_relaxed);
      }
  ||| {
      y.load(memory_order_relaxed);
      x.load(memory_order_relaxed);
      }
  }}}
}

That was the CppMem program. Now to the graph.

Execution for (y=0,x=2000)

The graph shows crystal clear the unintuitive behaviour.

relaxed

x reads the value 2000 from the writing thread, but y reads the value 0 from the main thread. What happens, although the reading of y is sequenced before the reading of x. Sequenced before exact means, that the operation e:Rrix sb is sequenced-before the operation f:Rrix.

What's next?

This was the last post in my mini-series about ongoing optimization. So, what's next? There are a lot of issues with the singleton pattern. I'm totally aware of that. But the singleton pattern is an ideal use case for a variable, which has to be initialized in a thread-safe way. From that point on, you can use it without synchronization.
So in the next post, I discuss different ways to initialize a singleton in a multithreading environment. You get the performance numbers and can reason about your uses cases for the thread-safe initialization of a variable.

Thanks a lot to my Patreon Supporters: Matt Braun, Roman Postanciuc, Tobias Zindl, Marko, G Prvulovic, Reinhold Dröge, Abernitzke, Frank Grimm, Sakib, Broeserl, António Pina, Sergey Agafyin, Андрей Бурмистров, Jake, GS, Lawton Shoemake, Animus24, Jozo Leko, John Breland, Louis St-Amour, Venkat Nandam, Jose Francisco, Douglas Tinkham, Kuchlong Kuchlong, Robert Blanch, Truels Wissneth, Kris Kafka, Mario Luoni, Neil Wang, Friedrich Huber, lennonli, Pramod Tikare Muralidhara, Peter Ware, Daniel Hufschläger, Alessandro Pezzato, Evangelos Denaxas, Bob Perry, Satish Vangipuram, Andi Ireland, Richard Ohnemus, Michael Dunsky, Leo Goodstadt, John Wiederhirn, Yacob Cohen-Arazi, Florian Tischler, Robin Furness, Michael Young, Holger Detering, Bernd Mühlhaus, Matthieu Bolt, Stephen Kelley, Kyle Dean, Tusar Palauri, Dmitry Farberov, Ralf Holly, Juan Dent, George Liao, Daniel Ceperley, and Jon T Hess.

Thanks in particular to Jon Hess, Lakshman, Christian Wittenhorst, Sherhy Pyton, Dendi Suhubdy, Sudhakar Belagurusamy, Richard Sargeant, Rusty Fleming, Ralf Abramowitsch, John Nebel, and Mipko.

My special thanks to Embarcadero

My special thanks to PVS-Studio

Mentoring Program in English

"Fundamentals for C++ Professionals": Reopens October 2022
"Design Patterns and Architectural Patterns in C++": Start January 2023

Do you want to stay informed about my mentoring programs? Write to This email address is being protected from spambots. You need JavaScript enabled to view it..

Seminars

I'm happy to give online seminars or face-to-face seminars worldwide. Please call me if you have any questions.

Bookable (Online)

German

Embedded Programmierung mit modernem C++: 28.06.2022 - 30.06.2022 (Termingarantie)
C++20: 23.08.2022 - 25.08.2022 (Termingarantie)
Clean Code: Best Practices für modernes C++: 11.10.2022 - 13.10.2022
Design Pattern und Architekturpattern mit C++: 08.11.2022 - 10.11.2022 (Termingarantie)

Standard Seminars (English/German)

Here is a compilation of my standard seminars. These seminars are only meant to give you a first orientation.

New

Contact Me

Phone: +49 7472 917441
Mobil:: +49 176 5506 5086
Mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
German Seminar Page: www.ModernesCpp.de
English Seminar Page: www.ModernesCpp.net

Modernes C++,

RainerGrimmSmall

Tags: Ongoing Optimization, CppMem, atomics, relaxed semantic

Comments

0 #1 Crystal x asli 2016-11-16 18:59

I simply could not go away your website prior to
suggesting that I actually enjoyed the usual info a person provide on your guests?
Is gonna be back continuously in order to inspect new posts

Quote

0 #2 人妻無料動画 2016-12-01 00:35

Excellent read, I just passed this onto a colleague who was doing some research on that.
And he just bought me lunch since I found it for him smile So
let me rephrase that: Thank you for lunch!

Quote

Refresh comments list
RSS feed for comments to this post

Add comment

JComments