Parallel Transforms

Table of Contents

• Include the Header
• Create a Unary Parallel-Transform Task
• Create a Binary Parallel-Transform Task

Taskflow provides template functions for constructing tasks to perform parallel transforms over ranges of items.

Include the Header

You need to include the header file, taskflow/algorithm/transform.hpp, for creating a parallel-transform task.

#include <taskflow/algorithm/transform.hpp>

Create a Unary Parallel-Transform Task

Parallel-transform transforms a range of items, possibly with a different type for the transformed data, and stores the result in another range. The task created by tf::Taskflow::transform(B first1, E last1, O d_first, C c) is equivalent to a parallel execution of the following loop:

while (first1 != last1) {
  *d_first++ = c(*first1++);
}

By default, tf::Taskflow::transform(B first1, E last1, O d_first, C c) creates a task to spawn a subflow (see Dynamic Tasking) that simultaneously applies the callable c to the object obtained by dereferencing every iterator in the range [first1, last1) and stores the result in another range beginning at d_first. It is user's responsibility for ensuring the range is valid within the execution of the parallel-transform task. Taskflow's parallel-transform tasks work on all iterable STL containers.

std::vector<int> src = {1, 2, 3, 4, 5};
std::vector<int> tgt(src.size());
taskflow.transform(src.begin(), src.end(), [](int i){ 
  std::cout << "transforming item " << i << " to " << i + 1 << '\n';
  return i + 1;
});

You can enable stateful iterators by creating a reference wrapper and pass the wrapped iterator to the argument of tf::Taskflow::transform. This is especially useful when the range is not known at the time of creating a parallel-transform task, but through another task.

std::vector<int> src, tgt;
std::vector<int>::iterator first, last, d_first;
 
tf::Task init = taskflow.emplace([&](){
  src.resize(1000);
  tgt.resize(1000);
  first   = src.begin();
  last    = src.end();
  d_first = tgt.begin();
});
 
tf::Task transform = taskflow.for_each(
  std::ref(first), std::ref(last), std::ref(d_first), 
  [&](int i) {
    std::cout << "transforming item " << i << " to " << i + 1 << '\n';
    return i+1;
  }
);
 
init.precede(transform);

When init finishes, the parallel-transform task transform will see first pointing to the beginning of src and last pointing to the end of src. Then, it simultaneously transforms these 1000 items by adding one to each element and stores the result in another range starting at d_first.

Create a Binary Parallel-Transform Task

You can use the overload, tf::Taskflow::transform(B1 first1, E1 last1, B2 first2, O d_first, C c), to perform parallel transforms on two source ranges pointed by first1 and first2 using the binary operator c and store the result in another range pointed by d_first. This method is equivalent to the parallel execution of the following loop:

while (first1 != last1) {
  *d_first++ = c(*first1++, *first2++);
}

The following example creates a parallel-transform task that adds two ranges of elements one by one and stores the result in a target range:

std::vector<int> src1 = {1, 2, 3, 4, 5};
std::vector<int> src2 = {5, 4, 3, 2, 1};
std::vector<int> tgt(src.size());
taskflow.transform(
  src1.begin(), src1.end(), src2.begin(), tgt.begin(), 
  [](int i, int j){ 
    return i + j;
  }
);