Compiling Dynamic Mappings with Array

Array remappings are useful to many applications on distributed memory parallel machines. They are integrated into High Performance For-tran, a Fortran 90-based data-parallel language. This paper describes techniques for handling dynamic mappings through simple array copies: array remappings are translated into copies between statically mapped distinct versions of the array. It discusses the language restrictions required to do so. The remapping graph which captures all remapping and liveness information is presented, as well as additional data-ow optimizations that can be performed on this graph, so as to avoid useless remappings at runtime. Such useless remappings appear for arrays that are not used after a remapping. Live array copies are also kept to avoid other ow-dependent useless remappings. Finally the code generation and runtime management issues needed by our scheme are discussed. These techniques are implemented in our prototype HPF compiler.