Key RAJA features shown in this example:

  • RAJA::forall loop execution template
  • RAJA::RangeSegment iteration space construct
  • RAJA reduction types
  • RAJA reduction policies

In the Vector Dot Product (Sum Reduction) example, we showed how to use the RAJA sum reduction type. The following example uses all supported RAJA reduction types: min, max, sum, min-loc, max-loc.


Multiple RAJA reductions can be combined in any RAJA loop kernel execution method, and reduction operations can be combined with any other kernel operations.

We start by allocating an array (the memory manager in the example uses CUDA Unified Memory if CUDA is enabled) and initializing its values in a manner that makes the example mildly interesting and able to show what the different reduction types do. Specifically, the array is initialized to a sequence of alternating values (‘1’ and ‘-1’). Then, two values near the middle of the array are set to ‘-100’ and ‘100’:

// Define array length
  const int N = 1000000;

// Allocate array data and initialize data to alternating sequence of 1, -1.
  int* a = memoryManager::allocate<int>(N);

  for (int i = 0; i < N; ++i) {
    if ( i % 2 == 0 ) {
      a[i] = 1;
    } else {
      a[i] = -1; 

// Set min and max loc values
  const int minloc_ref = N / 2;
  a[minloc_ref] = -100;

  const int maxloc_ref = N / 2 + 1;
  a[maxloc_ref] = 100;

We also define a range segment to iterate over the array:

  RAJA::RangeSegment arange(0, N);

With these parameters and data initialization, all the code examples presented below will generate the following results:

  • the sum will be zero
  • the min will be -100
  • the max will be 100
  • the min loc will be N/2
  • the max loc will be N/2 + 1

A sequential kernel that exercises all RAJA sequential reduction types is:

  using EXEC_POL1   = RAJA::seq_exec;
  using REDUCE_POL1 = RAJA::seq_reduce;
  RAJA::ReduceSum<REDUCE_POL1, int> seq_sum(0);
  RAJA::ReduceMin<REDUCE_POL1, int> seq_min(std::numeric_limits<int>::max());
  RAJA::ReduceMax<REDUCE_POL1, int> seq_max(std::numeric_limits<int>::min());
  RAJA::ReduceMinLoc<REDUCE_POL1, int> seq_minloc(std::numeric_limits<int>::max(), -1);
  RAJA::ReduceMaxLoc<REDUCE_POL1, int> seq_maxloc(std::numeric_limits<int>::min(), -1);

  RAJA::forall<EXEC_POL1>(arange, [=](int i) {
    seq_sum += a[i];


    seq_minloc.minloc(a[i], i);
    seq_maxloc.maxloc(a[i], i);


  std::cout << "\tsum = " << seq_sum.get() << std::endl;
  std::cout << "\tmin = " << seq_min.get() << std::endl;
  std::cout << "\tmax = " << seq_max.get() << std::endl;
  std::cout << "\tmin, loc = " << seq_minloc.get() << " , " 
                               << seq_minloc.getLoc() << std::endl;
  std::cout << "\tmax, loc = " << seq_maxloc.get() << " , " 
                               << seq_maxloc.getLoc() << std::endl;

Note that each reduction object takes an initial value at construction. Also, within the kernel, updating each reduction is done via an operator or method that is basically what you would expect (i.e., ‘+=’ for sum, ‘min()’ for min, etc.). After the kernel executes, the reduced value computed by each reduction object is retrieved after the kernel by calling a ‘get()’ method on the reduction object. The min-loc/max-loc index values are obtained using ‘getLoc()’ methods.

For parallel multi-threading execution via OpenMP, the example can be run by replacing the execution and reduction policies with:

  using EXEC_POL2   = RAJA::omp_parallel_for_exec;
  using REDUCE_POL2 = RAJA::omp_reduce;

Similarly, the kernel containing the reductions can be run in parallel on a CUDA GPU using these policies:

  using EXEC_POL3   = RAJA::cuda_exec<CUDA_BLOCK_SIZE>;
  using REDUCE_POL3 = RAJA::cuda_reduce;


Each RAJA reduction type requires a reduction policy that must be compatible with the execution policy for the kernel in which it is used.

The file RAJA/examples/tut_reductions.cpp contains the complete working example code.