PWR035: avoid non-consecutive array access to improve performance

Issue

Non-consecutive array access may impact performance.

Actions

Consider using techniques like loop fusion, loop interchange, loop tiling or changing the data layout to avoid non-consecutive access in hot loops.

Relevance

Accessing an array in a non-consecutive order is less efficient than accessing consecutive positions because the latter maximises locality of reference. 

Code examples

Consider the example code below to illustrate the presence of non-consecutive access patterns. The elements of array a are accessed in a non-consecutive manner. In the scope of the outer loop for_i, all the iterations access the first row of the array. Thus, the code exhibits repeated accesses to all the elements of the first row, a total number of times equal to rows.

void example(float **a, unsigned rows, unsigned cols) {
  for (unsigned i = 0; i < rows; ++i) {
    for (unsigned j = 0; j < cols; ++j) {
      a[0][j] = 0.f;
    }
  }
}

Related resources

• PWR035 examples at GitHub

References

• Memory access pattern (non-consecutive array access)
• Locality of reference
• Loop interchange
• Loop tiling