| description | title | ms.date | helpviewer_keywords | ||
|---|---|---|---|---|---|
Learn more about how to write a parallel_for_each Loop |
How to: write a parallel_for_each Loop |
09/12/2022 |
|
This example shows how to use the concurrency::parallel_for_each algorithm to compute the count of prime numbers in a std::array object in parallel.
The following example computes the count of prime numbers in an array two times. The example first uses the std::for_each algorithm to compute the count serially. The example then uses the parallel_for_each algorithm to perform the same task in parallel. The example also prints to the console the time that is required to perform both computations.
[!code-cppconcrt-parallel-count-primes#1]
The following sample output is for a computer that has four cores.
serial version:
found 17984 prime numbers
took 125 ms
parallel version:
found 17984 prime numbers
took 63 ms
To compile the code, copy it and then paste it in a Visual Studio project, or paste it in a file that is named parallel-count-primes.cpp and then run the following command in a Visual Studio Command Prompt window.
cl.exe /EHsc parallel-count-primes.cpp
The lambda expression that the example passes to the parallel_for_each algorithm uses the InterlockedIncrement function to enable parallel iterations of the loop to increment the counter simultaneously. If you use functions such as InterlockedIncrement to synchronize access to shared resources, you can present performance bottlenecks in your code. You can use a lock-free synchronization mechanism, for example, the concurrency::combinable class, to eliminate simultaneous access to shared resources. For an example that uses the combinable class in this manner, see How to: Use combinable to improve performance.