The two most important problems that allow for the efficient use of parallel processing power can be identified as detection of parallelism in computer programs and scheduling of the resulting parallel tasks on a target machine. This work is concerned with both of these problems. It is only recently that some researchers are concentrating on the detection of parallelism in ordinary (non-scientific) application programs. The research to be developed subsequent to the completion of the planning period will concentrate on detection of parallelism and determination of optimal task granularity in ordinary computer programs and processor scheduling and synchronization in multiprocessor environments. The research planning period mainly consists of plans for spending certain periods of time in residence in leading research sites around the country in order to get involved in the ongoing research in these sites and broaden the proposer's basis for doing independent research.
