Add occa::forLoop

[dmed256]

API

Basic Example

Here we generate a for-loop that goes through [0, N) and tiled by tileSize

occa::forLoop()
  .outer(occa::range(0, N, tileSize))
  .inner(tileSize)
  .run(scope, OCCA_FUNCTION([=](const int outerIndex, const int innerIndex) -> void {
    // ...
  }));

Indices + Multiple Dimensions

We give an example where an index array is passed rather than a simple occa::range
Additionally, this @inner loop has 2 dimensions so the expected OCCA_FUNCTION should be taking in an int2 for the inner indices

occa::array<int> indices;
// ...
occa::forLoop()
  .outer(indices)
  .inner(X, Y)
  .run(scope, OCCA_FUNCTION([=](const int outerIndex, const int2 innerIndex) -> void {
    // ...
  }));

Implementation

It would be nice to template the forLoop<outerSize, innerSize> to avoid code repetition, but there are 2 not-so-great UX issues:

• Template errors would be hard to decipher
• No easy way to set intTuple<outerSize> -> int / int2 / int3
• No way to enforce the only-outer or outer+inner loops when calling run that I'm aware of

We can also add tile methods to make it simple to generate the outer/inner loop combinations

class outerForLoop<O> : public outerForLoop {
  occa::iteration outerIterations[<O>];
  
  class innerForLoop<I> : public innerForLoop {
    occa::iteration innerIterations[<I>];

    void run(occa::function<void(int<O>, int<I>)> fn);

    void run(occa::scope scope,
             occa::function<void(int<O>, int<I>)> fn);
  };
    
  innerForLoop1 inner(occa::iteration iteration1);
  
  innerForLoop2 inner(occa::iteration iteration1,
                      occa::iteration iteration2);
  
  innerForLoop3 inner(occa::iteration iteration1,
                      occa::iteration iteration2,
                      occa::iteration iteration3);

  void run(occa::function<void(int<O>)> fn);

  void run(occa::scope scope,
           occa::function<void(int<O>)> fn);
};  

class forLoop {
  forLoop();
  
  outerForLoop1 outer(occa::iteration iteration1);
  
  outerForLoop2 outer(occa::iteration iteration1,
                      occa::iteration iteration2);
  
  outerForLoop3 outer(occa::iteration iteration1,
                      occa::iteration iteration2,
                      occa::iteration iteration3);
};