Addt: Automatic Data Distribution Tool for Porting Programs to Distributed Memory Machines

Data parallel languages like HPF require the user to specify the distribution of the data in the application. This distribution is speciied over the processors on which the application is to execute. A good data distribution depends on the number of processors, the communication latency and the underlying compilation system. Hence it is diicult to do this manually. In this paper we present a method for automatically generating a data distribution for a distributed memory environment. This method has been implemented in an automatic data distribution tool called ADDT. ADDT reads in a shared memory parallel program in Fortran and generates as output a HPF program. Our method works by breaking up the input program into sub parts called distribution blocks (DB). (A somewhat similar idea has been suggested by other researchers). It then obtains a data distribution for the arrays in each DB using the data layouts from preceding DB's. The 1 2 algorithms used to obtain a data distribution for a DB are also discussed in this paper. We present experimental results obtained by running ADDT on some example programs.