Automatic Data Layout for High Performance Fortran

High Performance Fortran (HPF) is rapidly gaining acceptance as a language for parallel programming. The goal of HPF is to provide a simple yet efficient machine independent parallel programming model. Besides the algorithm selection, the data layout choice is the key intellectual step in writing an efficient HPF program. The developers of HPF did not believe that data layouts can be determined automatically in all cases, Therefore HPF requires the user to specify the data layout. It is the task of the HPF compiler to generate efficient code for the user supplied data layout. The choice of a good data layout depends on the HPF compiler used, the target architecture, the problem size, and the number of available processors. Allowing remapping of arrays at specific points in the program makes the selection of an efficient data layout even harder. Although finding an efficient data layout fully automatically may not be possible in all cases. HPF users will need support during the data layout selection process. In particular, this support is necessary if the user is not familiar with the characteristics of the target HPF compiler and target architecture, or even with HPF itself. Therefore, tools for automatic data layout and performance estimation will be crucial if the HPF is to find general acceptance in the scientific community. This paper discusses a framework for automatic data layout for use in a data layout assistant tool for a data-parallel language such as HPF. The envisioned tool can be used to generate a first data layout for a sequential Fortran program without data layout statements, or to extend a partially specified data layout in a HPF program to a totally specified data layout. Since the data layout assistant is not embedded in a compiler and will run only a few times during the tuning process of an application program, the framework can use techniques that may be too computationally expensive to be included in a compiler. A prototype data layout assistant tool based on our framework has been implemented as part of the D system currently under development at Rice University. The paper reports preliminary experimental results. The results indicate that the framework is efficient and generates data layouts of high quality.