Distributed input/output processing in data-driven multiprocessors
- Southern Methodist Univ., Dallas, TX (United States). Dept. of Computer Science and Engineering
- University of Southern California, Los Angeles, CA (United States). Dept. of Electrical Engineering Systems
Data-flow principles of execution provide an elegant way to ensure at runtime that instructions can be executed asynchronously in a parallel environment. However, while the conventional von Neumann model of interpretation has a very rigid ordering of instructions, it is the very asynchronous character of the dataflow model of execution that introduces conflicts when ``state`` tasks (such as I/O operations) must share common data objects. In order to execute I/O operations safely and in parallel, an algorithm to detect and classify cases of potential conflicts (hazards) has been developed; it is described in this paper. It is based upon localizing the effect of I/O operations by splitting the data-flow graph into two subgraphs: (a) the computation subgraph, and (b) the I/O subgraph. The scheme presented in this paper thus enables the creation and interaction of both subgraphs, which in turn yields a deterministic execution. Furthermore, the proposed scheme enables the distributed execution of I/O operations as permitted by data dependencies.
- Research Organization:
- University of Southern California, Los Angeles, CA (United States). Dept. of Electrical Engineering Systems
- Sponsoring Organization:
- USDOE, Washington, DC (United States); Department of Defense, Washington, DC (United States); National Aeronautics and Space Administration, Washington, DC (United States)
- DOE Contract Number:
- FG03-87ER25043
- OSTI ID:
- 10191954
- Report Number(s):
- CONF-901221-2; ON: DE94000323; BR: KC0701030; CNN: Agreement NCC 2-539
- Resource Relation:
- Conference: 2. IEEE symposium on parallel and distributed processing,Dallas, TX (United States),9-13 Dec 1990; Other Information: PBD: [1992]
- Country of Publication:
- United States
- Language:
- English
Similar Records
Solving Partial Differential Equations in a data-driven multiprocessor environment
Eps'88: Combining the best features of von Neumann and dataflow computing