Cda Loop Transformations

In this paper we present a new loop transformation technique called Computation Decomposition and Alignment (CDA). Computation Decomposition rst decomposes the iteration space into ner computation spaces. Computation Alignment subsequently, linearly transforms each computation space independently. CDA is a general framework in that linear transformations and its recent extensions are just special cases of CDA. CDA's ne grained loop restructuring can incur considerable computational e ort, but can exploit optimization opportunities that earlier frameworks cannot. We present four optimization contexts in which CDA can be useful. Our initial experiments demonstrate that CDA adds a new dimension to performance optimization.