Manchester data-flow: a progress report

The Manchester Data-Flow Machine, MDFM, has evolved continuously during the past decade. By the time the prototype uniprocessor hardware system was decommissioned, in 1989, the putative multi-processor architecture comprised separate Processing Elements and Structure Store Units, together with a “global” Allocator and Throttle Unit, all linked by a packet-based Interconnection Switch. The decisions leading to this design are well documented in the literature, but some of the reasoning behind these decisions is now out-of-date. This paper looks afresh at the MDFM architecture, in the light of new hardware and software technology and the new state-of-the-art in data-flow computing, and derives an updated architecture, NMDFM, for a Manchester-style multi-processor system. The variable parameters of the NMDFM are investigated in an attempt to find an “optimum” set for future simulation studies.

[1]  Allan Porterfield,et al.  The Tera computer system , 1990, ICS '90.

[2]  Yong Meng Teo,et al.  Performance issues in dataflow machines , 1987, Future Gener. Comput. Syst..

[3]  Allan Porterfield,et al.  The Tera computer system , 1990 .

[4]  Arvind,et al.  Some Relationships Between Asynchronous Interpreters of a Dataflow Language , 1977, Formal Description of Programming Concepts.

[5]  J. R. Gurd,et al.  A scalable dataflow structure store , 1986, ISCA 1986.

[6]  J. V. Woods,et al.  Design of a processing subsystem for the Manchester data-flow computer , 1981 .

[7]  Yong Meng Teo,et al.  Resource management in a multi-ring dataflow machine , 1989 .

[8]  David A. Patterson,et al.  Computer Architecture: A Quantitative Approach , 1969 .

[9]  David E. Culler,et al.  Managing resources in a parallel machine , 1986 .

[10]  Joseph A. Fisher.  The VLIW Machine: A Multiprocessor for Compiling Scientific Code , 1984, Computer.

[11]  Stephen K. Skedzielewski,et al.  Data Flow Graph Optimization in IF1 , 1985, FPCA.

[12]  Yasuhiro Inagami,et al.  The specification of a new Manchester Dataflow machine , 1989, ICS '89.

[13]  A. P. Wim Böhm,et al.  Code Optimization for Tagged-Token Dataflow Machines , 1989, IEEE Trans. Computers.

[14]  I. Watson,et al.  Pseudo-associative store with hardware hashing , 1983 .

[15]  Kenji Nishida,et al.  Evaluation of a prototype data flow processor of the SIGMA-1 for scientific computations , 1986, ISCA 1986.

[16]  John R. Gurd,et al.  Iterative Instructions in the Manchester Dataflow Computer , 1990, IEEE Trans. Parallel Distributed Syst..

[17]  John Glauert,et al.  SISAL: streams and iteration in a single assignment language. Language reference manual, Version 1. 2. Revision 1 , 1985 .

[18]  古谷 立美.  A Highly Parallel Computer Architecture , 1985 .

[19]  Michael O'Boyle,et al.  Program and data transformations for efficient execution on distributed memory architectures , 1993, Technical report series.

[20]  John Sargeant,et al.  Control of parallelism in the Manchester Dataflow Machine , 1987, FPCA.

[21]  Gregory M. Papadopoulos,et al.  Implementation of a general purpose dataflow multiprocessor , 1991 .

[22]  John R. Gurd,et al.  Tools for Performance Evaluation of Parallel Machines , 1987, ICS.

[23]  John R. Gurd,et al.  Processor Allocation in a Multi-ring Dataflow Machine , 1986, J. Parallel Distributed Comput..

[24]  Kenji Nishida,et al.  Evaluation of a Prototype Data Flow Processor of the SIGMA-1 for Scientific Computations , 1986, ISCA.

[25]  Ian Watson,et al.  The Manchester prototype dataflow computer , 1985, CACM.

[26]  D. Dent.  The ECMWF Model: Past, Present and Future , 1988 .

[27]  John Glauert,et al.  SISAL: streams and iteration in a single-assignment language. Language reference manual, Version 1. 1 , 1983 .