Dataflow architectures offer the ability to trade program-level parallelism for machine level latency. Dataflow further offers a uniform synchronization paradigm, representing one end of a spectrum wherein the unit of scheduling is a single instruction. At the opposite extreme are the von Neumann architectures which schedule on a task, or process, basis. The spectrum is examined and an architecture which is a hybrid of dataflow and von Neumann organizations is proposed. The analysis attempts to discover those features of the dataflow architecture, lacking in a von Neumann machine, which are essential for tolerating latency and synchronization costs. These features are captured in the concept of a parallel machine language which can be grafted on top of an otherwise traditional von Neumann base. In such an architecture, the units of scheduling, called scheduling quanta, are bound at compile time rather than at instruction-set design time. The parallel machine language supports this notion using a large synchronization name space. A prototypical architecture is described, and results of simulation studies are presented. A comparison is made between the MIT tagged-token dataflow machine and the subject machine which presents a model for understanding the cost of synchronization in a parallel environment.<<ETX>>
[1]
Jarek Deminet.
Experience with Multiprocessor Algorithms
,
1982,
IEEE Transactions on Computers.
[2]
Harry F. Jordan.
Performance measurements on HEP - a pipelined MIMD computer
,
1983,
ISCA '83.
[3]
Very Long Instruction Word architectures and the ELI-512
,
1983,
ISCA '83.
[4]
David A. Patterson,et al.
Reduced instruction set computers
,
1985,
CACM.
[5]
Kattamuri Ekanadham,et al.
Incorporating Data Flow Ideas into von Neumann Processors for Parallel Execution
,
1987,
IEEE Transactions on Computers.
[6]
K. R. Traub,et al.
Sequential implementation of lenient programming languages
,
1988
.