A basic architecture supporting LGDG computation
Pages 23 - 33
Abstract
In order to combine the benefits of dataflow and control-flow computation while avoiding the pitfalls of both, the authors propose a two-level model of large-grain dataflow computation, called LGDG computation. A formalism has been provided in a previous paper to prove the determinism of parallel program execution under this model. The current paper presents the basic LGDG computer architecture which is organized at two levels, called the graph level and the node level. The kernel of the node-level architecture is characterized by its non-branch RISC (N-RISC) feature, which ensures an optimal utilization of pipeline processing. This kernel is supported by various co-processors based on the principle of function migration. We will show in this paper how the graph-level architecture can be reduced to a matching unit, extended by a writable control store, through node migration and node aggregation. As a result, the matching overhead is drastically reduced, thus eliminating the most severe bottleneck of existing 'fine-grain' dataflow computers.
References
[1]
Arvind, and iannucci, R.A., Two Fundamental Issues in Multiprocessing, DFVLR - Conference 1987 on 'Parallel Processing in Science and Engineering', June 25-26, 1987.
[2]
P. C. Treleaven, D.R. Brownbridge, and R.P. Hopkins, Data-Driven and Demand-Driven Computer Architecture, Computing Surveys, Vol. 14, No. I, pp. 93-143. March 1982.
[3]
Arvind and V. Kathail, A Mul@le Processor Data Flow Machine that Supports Generalized Procedures, Transaction of IEEE, pp.291-302. 1981.
[4]
P.C. Yew, Architecture of the Cedar Parallel Supercomputer, CSRD Rpt. No. 609, Center for Supercomputing Research and Development, University of Illinois, August 1986.
[5]
K.C. Dai, Large-Grain Dataflow Computation and Its Architectural Support, PhD Thesis, Technical Universit>, of Berlin, West Germany, 1988.
[6]
J.R. Gurd, et al., Fine.Grain Parallel Computing: The Dataflow Approach, in 'Future Parallel Computers' pp.82-153, (Eds. P. Treleaven, M. Vanneschi), An Advanced Course, Pisa, Italy, June 1986, Lecture Notes in Computer Science 272, Springer-Verlag, 1987.
[7]
A.V. Aho, R. Sethi, and J.D. Ullman, Compilers, Principles, Techniques, and Tools, Addison-Wesley Publishing Company, 1986.
[8]
K.C. Dai and W.K, Giloi, A Two-Level Model of Large.Grain Dataflow Computation Proc. of CA-DSP '89 International Symposium on Computer Architecture & Digital Signal Processing, pp. 200-205. Hong-Kong, 11-14th Oct. 1989,
[9]
P. Brinch Hansen, Edison Programs, Software-Practice and Experience, pp.397-414. 11 (4) 1981.
[10]
K.C. Dai, EDISON-80, A Language for Modular Pro. gramming of Parallel Processes, Information processhag Letters, North-Holland, pp.61-72. Jan. 1986.
[11]
K.C. Dai and W.K. Giloi, Unfolding of Parallelism in the LGDG Confutation, to be published.
[12]
Arvind and R.E. ThomasJ.Structares: An Efficient Data Type for Functional Languages, Laboratory for Computer Science, MIT, 12 September 1980.
[13]
D.E. Culler and Arvind, Resource Requirements of Dataflow Programs, Proc. of The 15th Annual International Symposium on Computer Architecture, pp.141- 150. May 1988.
[14]
K. Kawakami, and J.R. Gurd, A Scalable Dataflow Structure Store, IEEE Transaction, pp.243-250. 1986.
[15]
W.K. Giloi, The DRAMA Principle and Data Type Architectures, in J. Niedereichholz (ed.), Datenbanktechnologic, Teubner, 81-100. Stuttgart 1979.
[16]
W.K. Giloi and P. Behr, Hierarchical Function Distribution - A Design Principle for Advanced Multicom. purer Architectures, Proc. l Oth International Symposium on Computer Architecture, pp. 318-325. IEEE Catalog No. 83CH1889-5, 1983.
[17]
R.S. Nikhil, Arvind, Can Dataflow subsume yon Neumann Computing, Proc. of the 16th annual Int. Syrup. on Computer Architecture, Jerusalem, Israel, pp. 262- 272. May 28th - June 1st, 1989.
[18]
A. Veen, Dataflow Machine Architecture, ACM Computing Surveys, Vol.18, No.4, pp. 365-396. December 1986.
[19]
Arvind, R.S. Nikhil, and K.K. Pingali, l.Structures: Data Structures for Parallel Computing, ACM Trans. of Programming Languages and Systems, Vol. 11, No.4, pp. 598-632, Oct. 1989.
[20]
W.K. Giloi and H.K. Berg, Data Structure Architectures . A Major Operational Principle, Proc. 5th Arm. Symposium on Computer Architecture, IEEE New York, 1978.
[21]
P.M. Kogge, The Architecture of Pipelined Computers, McGraw-Hill advanced computer science series, McGraw-Hill Book Company, 1981.
[22]
D.J. Lilja, Reducing the Branch Penalty in Pipelined Processors, Computer, pp.47-55. July 1988.
[23]
U. Briining, Realisierungsprinzipien von Mikrorechnern hoher numerischer Le/stung, PhD thesis, Technical University Berlin, 1987.
[24]
K.C. Dai and W.K. Giloi, Towards Incorporating the Principle of DSA into the LGDG Computation, to be published.
[25]
R.A. Iannucci, Towards a Dataflow Ivon Neumann Hybrid Architecture, Proc. of 15th Int'l Symp. on Comp. Arch. pp.131-140. 1988.
[26]
B.R. Preiss, and V.C.Hamacher, Data Flow on a Queue Machine, Proe. of the 12th Annual Symposium on Computer Architecture (Boston, Mass., June 17-19, 1985). IEEE, pp. 342-351. New York.
[27]
R. Buehrer and K. Ekanadham, Incorporating Data Flow Ideas into yon Neumann Processors for Parallel Execution, IEEE Trans. on Comp. vol. C-36, no. 12, pp. 1515-1521. Dee. 1987.
[28]
Am29000 32-bit Streamlined instruction Processor - Users Manual, Advanced Micro Devices, 1988.
Index Terms
- A basic architecture supporting LGDG computation
Recommendations
A basic architecture supporting LGDG computation
ICS '90: Proceedings of the 4th international conference on SupercomputingIn order to combine the benefits of dataflow and control-flow computation while avoiding the pitfalls of both, the authors propose a two-level model of large-grain dataflow computation, called LGDG computation. A formalism has been provided in a ...
Internal architecture of Alpha 21164 microprocessor
COMPCON '95: Proceedings of the 40th IEEE Computer Society International ConferenceThe internal architecture of a 1200 MIPS/600 MFLOPS (peak) high-performance CMOS ALPHA microprocessor chip is described. This second-generation implementation is the world's fastest microprocessor. It contains a quad-issue superscalar instruction unit, ...
Architectural differences of efficient sequential and parallel computers
In this paper we try to conclude what kind of a computer architecture is efficient for executing sequential problems, and what kind of an architecture is efficient for executing parallel problems from the processor architect's point of view. For that ...
Comments
Information & Contributors
Information
Published In
Sept. 1990
489 pages
Copyright © 1990 ACM.
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from [email protected]
Publisher
Association for Computing Machinery
New York, NY, United States
Publication History
Published: 01 June 1990
Published in SIGARCH Volume 18, Issue 3b
Check for updates
Author Tags
Qualifiers
- Article
Contributors
Other Metrics
Bibliometrics & Citations
Bibliometrics
Article Metrics
- View Citations5Total Citations
- 360Total Downloads
- Downloads (Last 12 months)48
- Downloads (Last 6 weeks)12
Reflects downloads up to 12 Aug 2024
Other Metrics
Citations
View Options
Get Access
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Sign in