Taskflow Algorithms » Parallel Find

Include the Header

You need to include the header file, taskflow/algorithm/find.hpp, for using parallel-find algorithms.

#include <taskflow/algorithm/find.hpp>

What is a Find Algorithm?

A find algorithm allows you to find an element in a range [first, last) that satisfies a specific criteria. The algorithm returns an iterator to the first found element in the range or returns last if there is no such iterator. Taskflow provides the following parallel-find algorithms:

• tf::Taskflow::find_if(B first, E last, T& result, UOP predicate, P&& part)
• tf::Taskflow::find_if_not(B first, E last, T& result, UOP predicate, P&& part)
• tf::Taskflow::min_element(B first, E last, T& result, C comp, P&& part)
• tf::Taskflow::max_element(B first, E last, T& result, C comp, P&& part)

Create a Parallel Find-If Task

tf::Taskflow::find_if performs parallel iterations to find the first element in the range [first, last) that makes the given predicate return true. It resembles a parallel implementation of the following loop:

template<typename InputIt, typename UnaryPredicate>
InputIt find_if(InputIt first, InputIt last, UnaryPredicate predicate) {
  for(; first != last; ++first) {
    if(predicate(*first)) {
      return first;
    }
  }
  return last;
}

The example below creates a task to find the element that is equal to 22 from an input range of 10 elements. The result will be stored in the forth argument passed by reference:

std::vector<int> input = {1, 9, 22, 3, -6, 13, 12, 0, 9, 11};
std::vector<int>::iterator result;
taskflow.find_if(
  input.begin(), input.end(), [](int i){ return i == 22; }, result
);
executor.run(taskflow);
assert(*result == 22);

Capture Iterators by Reference

You can pass iterators by reference using std::ref to marshal parameters update between dependent tasks. This is especially useful when the range iterators are not known at the time of creating a find-if task, but need initialization from another task.

std::vector<int> input;
std::vector<int>::iterator result, first, last;

// task to set up the range iterators
tf::Task init = taskflow.emplace([&](){
  input = {1, 9, 22, 3, -6, 13, 12, 0, 9, 11};
  first = input.begin(),
  last  = input.end();
});

// task to perform parallel find
tf::Task task = taskflow.find_if(
  std::ref(first), std::ref(last), result, [](int i){ return i == 22; }
);

init.precede(task);

executor.run(taskflow);
assert(*result == 22);

In the above example, when init finishes, input has been initialized to 10 elements with first and last pointing to the data range of input. The find-if task will then work on this initialized range as a result of passing iterators by reference.

Create a Parallel Find-If-Not Task

tf::Taskflow::find_if_not performs parallel iterations to find the first element in the range [first, last) that makes the given predicate return false. It resembles a parallel implementation of the following loop:

template<typename InputIt, typename UnaryPredicate>
InputIt find_if(InputIt first, InputIt last, UnaryPredicate predicate) {
  for(; first != last; ++first) {
    if(!predicate(*first)) {
      return first;
    }
  }
  return last;
}

The example below creates a task to find the element that is NOT equal to 22 from an input range of 10 elements. The result will be stored in the forth argument passed by reference:

std::vector<int> input = {1, 1, 22, 1, 1, 1, 1, 1, 1, 1};
std::vector<int>::iterator result;
taskflow.find_if_not(
  input.begin(), input.end(), result, [](int i){ return i == 1; }
);
executor.run(taskflow);
assert(*result == 22);

Similar to Capture Iterators by Reference, iterators of tf::Taskflow::find_if_not are templated to allow passing iterators by reference using std::ref. This is especially useful when the range iterators are not known at the time of creating a find-if-not task, but need initialization from another task.

Find the Smallest and the Largest Elements

tf::Taskflow::min_element finds the smallest element in a range [first, last) using the given comparison function object. The example below finds the smallest element, i.e., -1, from an input range of 10 elements and stores the iterator to that smallest element in result:

std::vector<int> input = {1, 1, 1, 1, 1, -1, 1, 1, 1, 1};
std::vector<int>::iterator result;
taskflow.min_element(
  input.begin(), input.end(), std::less<int>(), result
);
executor.run(taskflow).wait();
assert(*result == -1);

Similarly, tf::Taskflow::max_element finds the largest element in a range [first, last) using the given comparison function object. The example below finds the largest element, i.e., 2, from an input range of 10 elements and stores the iterator to that largest element in result:

std::vector<int> input = {1, 1, 1, 1, 1, 2, 1, 1, 1, 1};
std::vector<int>::iterator result;
taskflow.max_element(
  input.begin(), input.end(), std::less<int>(), result
);
executor.run(taskflow).wait();
assert(*result == 2);

Configure a Partitioner

You can configure a partitioner for parallel-find tasks (tf::Taskflow::find_if, tf::Taskflow::find_if_not, tf::Taskflow::min_element, tf::Taskflow::max_element) to run with different scheduling methods, such as guided partitioning, dynamic partitioning, and static partitioning. The following example creates two parallel-find tasks using two different partitioners, one with the static partitioning algorithm and another one with the guided partitioning algorithm:

std::vector<int> vec(1024, -1);
std::vector<int>::iterator result;

tf::ExecutionPolicy<tf::StaticPartitioner> static_partitioner;
tf::ExecutionPolicy<tf::GuidedPartitioner> guided_partitioner;

// create a parallel-find task with a static partitioner
taskflow.find_if(
  vec.begin(), vec.end(), result, [&](int i) { return i == -1; }, static_partitioner
);

// create a parallel-find task with a guided partitioner
taskflow.find_if(
  vec.begin(), vec.end(), result, [&](int i) { return i == -1; }, guided_partitioner
);