JPH054258U - Parallel distributed processor - Google Patents

Parallel distributed processor

Info

Publication number
JPH054258U
JPH054258U JP4715391U JP4715391U JPH054258U JP H054258 U JPH054258 U JP H054258U JP 4715391 U JP4715391 U JP 4715391U JP 4715391 U JP4715391 U JP 4715391U JP H054258 U JPH054258 U JP H054258U
Authority
JP
Japan
Prior art keywords
processor
processors
shared memory
parallel distributed
distributed processing
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
JP4715391U
Other languages
Japanese (ja)
Inventor
直彦 阿部
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Mitsubishi Heavy Industries Ltd
Original Assignee
Mitsubishi Heavy Industries Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Mitsubishi Heavy Industries Ltd filed Critical Mitsubishi Heavy Industries Ltd
Priority to JP4715391U priority Critical patent/JPH054258U/en
Publication of JPH054258U publication Critical patent/JPH054258U/en
Withdrawn legal-status Critical Current

Links

Landscapes

  • Multi Processors (AREA)
  • Advance Control (AREA)

Abstract

(57)【要約】 【目的】 本考案は、処理速度を向上させるために多数
のプロセッサを並列に使用する分散処理装置において、
最小のハードウエアで、処理速度が速く、システムの拡
張や形態の変更が容易な分散処理装置を提供することを
目的とする。 【構成】 並列処理で演算を行う各プロセッサ11をそ
れぞれアドレス/データバス12を介して共有メモリ1
3に結合し、この共有メモリ13を介してプロセッサ1
1間でデータの授受を行うことによりプロセッサ間の分
離にインターフェイスを不要にし、このような結合によ
るプロセッサを超格子構造に連結したことを特徴として
いる。
(57) [Summary] [Object] The present invention provides a distributed processing device that uses a large number of processors in parallel in order to improve the processing speed.
It is an object of the present invention to provide a distributed processing device which requires a minimum of hardware, has a high processing speed, and is capable of easily expanding a system and changing its form. [Structure] Each processor 11 that performs an arithmetic operation in parallel processing is connected to a shared memory 1 via an address / data bus 12.
3 and the processor 1 via this shared memory 13.
By exchanging data between one and the other, an interface is not required for separation between processors, and the processor is connected by a superlattice structure by such coupling.

Description

【考案の詳細な説明】[Detailed description of the device]

【0001】[0001]

【産業上の利用分野】[Industrial applications]

本考案は、処理速度の向上のために多数プロセッサを用いる計算機(並列分散 処理装置)に関する。 The present invention relates to a computer (parallel distributed processing device) that uses multiple processors to improve processing speed.

【0002】[0002]

【従来の技術】[Prior Art]

従来の並列分散処理装置は図2のようなものであり、各プロセッサ21間はシ リアルバス22でデータ授受を行っており、接続形態は格子状である。 A conventional parallel distributed processing device is as shown in FIG. 2, and data is transmitted and received between each processor 21 by a serial bus 22, and the connection form is a grid.

【0003】[0003]

【考案が解決しようとする課題】[Problems to be solved by the device]

しかしながら、図2に示すような従来の装置では次のような問題点がある。 However, the conventional device as shown in FIG. 2 has the following problems.

【0004】 (a)シリアルバス用の特殊なインターフェイス回路23を、各プロセッサ2 1毎に接続するため、その数はプロセッサ分必要であり、ハードウェア規模が増 大する。 (b)プロセッサ数の増加に伴い、最も遠いプロセッサ間でのデータ授受に要 する時間が増大する。 (c)プロセッサの追加によるハードウェアおよびソフトウェアの改修規模が 大きい。(A) The special interface circuit 23 for the serial bus is connected to each processor 21. Therefore, the number of processors is required for each processor 21 and the hardware scale increases. (B) As the number of processors increases, the time required for data transfer between the furthest processors increases. (C) The scale of hardware and software repairs due to the addition of processors is large.

【0005】 (d)各プロセッサにおける演算時間とシリアルバスによるデータ入力のタイ ミングのオーバラップにより発生する無駄時間(ソフトウェアのオーバヘッド) が大である。(D) A large amount of dead time (software overhead) is generated due to the overlap of the calculation time in each processor and the timing of data input by the serial bus.

【0006】 本考案は上記のような点に鑑みなされたもので最小のハードウェアで、より処 理速度の速い、システムの拡張および形態変更が容易な並列分散処理装置を提供 することを目的とする。The present invention has been made in view of the above points, and an object of the present invention is to provide a parallel distributed processing device with minimum hardware, faster processing speed, and easy system expansion and configuration change. To do.

【0007】[0007]

【課題を解決するための手段】[Means for Solving the Problems]

本考案の共有メモリによる超格子接続型並列分散処理装置は、プロセッサ間の データ授受を共有メモリを介して行い、また、プロセッサを超格子構造に接続す ることを特徴としている。 The superlattice connection type parallel distributed processing apparatus using the shared memory of the present invention is characterized in that data is exchanged between the processors via the shared memory and the processors are connected to the superlattice structure.

【0008】[0008]

【作用】[Action]

前記共有メモリによる超格子接続を特徴とする並列分散処理装置によれば、各 プロセッサはメモリへのアクセスによって他のプロセッサとのデータ授受が可能 であり、特別な付加回路を必要としない。また、データ入力を各プロセッサで必 要なタイミングで自由に実行できるため、ソフトウェアのオーバヘッドがない。 メモリによる各プロセッサがアイソレートされているため、同種、異種を問わず 容易ににプロセッサの追加が可能であり、プロセッサの追加によるソフトウェア の改修が容易である。 According to the parallel distributed processing device characterized by the superlattice connection by the shared memory, each processor can exchange data with another processor by accessing the memory, and no special additional circuit is required. Further, since data input can be freely executed by each processor at a necessary timing, there is no software overhead. Since each processor is isolated by memory, it is possible to easily add processors of the same type or different types, and it is easy to modify software by adding processors.

【0009】[0009]

【実施例】【Example】

以下図1を参照して本考案の一実施例としての共有メモリによる超格子接続並 列分散処理装置について説明する。 A superlattice-connected parallel distributed processing apparatus using a shared memory as an embodiment of the present invention will be described below with reference to FIG.

【0010】 12個のプロセッサ(実施例に於いては16ビット汎用プロセッサ使用)11 を図1の如く三角柱の超格子構造に接続する。各プロセッサ間に双方向同時アク セス可能なデュアルポートメモリ13を共有メモリとして配置し、プロセッサと メモリ間をプロセッサ直結のアドレス/データ・バス12で接続する。共有メモ リ上に読出・書込フラグを設け、各プロセッサはデータの書込および読出し毎に フラグをオン・オフすることにより、各プロセッサはデータの新旧および連続性 を管理する。Twelve processors (using a 16-bit general-purpose processor in the embodiment) 11 are connected to a triangular prism superlattice structure as shown in FIG. A dual port memory 13 capable of bidirectional simultaneous access is arranged as a shared memory between the processors, and the processors and the memories are connected by an address / data bus 12 directly connected to the processors. A read / write flag is provided on the shared memory, and each processor turns on / off the flag each time data is written or read, so that each processor manages old and new data and continuity.

【0011】[0011]

【考案の効果】[Effect of the device]

以上のように本考案によれば下記に示す特有の効果が得られる。 (a)共有メモリでプロセッサ間データ授受を行うことにより、特殊なインタ ーフェイス回路が不要となり、ハードウェアの小型化が可能である。 As described above, according to the present invention, the following unique effects can be obtained. (A) By exchanging data between the processors in the shared memory, no special interface circuit is required, and the hardware can be downsized.

【0012】 (b)共有メモリにより各プロセッサがアイソレートされるため、ハードウェ アおよびソフトウェアの設計労力が軽減される。また、プロセッサ追加による性 能向上が容易である。(B) Since each processor is isolated by the shared memory, the effort of designing hardware and software is reduced. In addition, it is easy to improve performance by adding a processor.

【0013】 (c)共有メモリによるプロセッサ間データ授受によるソフトウェアのオーバ ヘッド低減、および超格子接続によるプロセッサ間距離の短縮(最も離れている プロセッサ間のデータ授受を仲介するプロセッサ数の低減)による処理速度が向 上する。(C) Processing by reducing software overhead by exchanging data between processors using a shared memory and shortening the distance between processors by superlattice connection (reducing the number of processors that mediate data exchange between the furthest processors) Speed improves.

【図面の簡単な説明】[Brief description of drawings]

【図1】本考案の1実施例に係わる並列分散処理装置の
構成図。
FIG. 1 is a block diagram of a parallel distributed processing device according to an embodiment of the present invention.

【図2】従来の並列分散処理装置を示す構成図。FIG. 2 is a configuration diagram showing a conventional parallel distributed processing device.

【符号の説明】[Explanation of symbols]

11…プロセッサ、12…アドレス/データ・バス、1
3…デュアルポートメモリ。
11 ... Processor, 12 ... Address / data bus, 1
3-Dual port memory.

Claims (1)

【実用新案登録請求の範囲】 【請求項1】 演算を行うプロセッサと、各プロセッサ
間のデータ授受を仲介する共有メモリと、プロセッサと
共有メモリを結ぶアドレス/データ・バスとにより構成
され、これらが超格子構造に連結されていることを特徴
とする並列分散処理装置。
[Claims for utility model registration] Claims: 1. A processor for performing an operation, a shared memory for mediating data exchange between the processors, and an address / data bus connecting the processor and the shared memory. A parallel distributed processing device characterized by being connected to a superlattice structure.
JP4715391U 1991-06-21 1991-06-21 Parallel distributed processor Withdrawn JPH054258U (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP4715391U JPH054258U (en) 1991-06-21 1991-06-21 Parallel distributed processor

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP4715391U JPH054258U (en) 1991-06-21 1991-06-21 Parallel distributed processor

Publications (1)

Publication Number Publication Date
JPH054258U true JPH054258U (en) 1993-01-22

Family

ID=12767151

Family Applications (1)

Application Number Title Priority Date Filing Date
JP4715391U Withdrawn JPH054258U (en) 1991-06-21 1991-06-21 Parallel distributed processor

Country Status (1)

Country Link
JP (1) JPH054258U (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS5162343U (en) * 1974-11-09 1976-05-17
JPS5425356U (en) * 1977-07-20 1979-02-19
JP2022526929A (en) * 2019-03-27 2022-05-27 グラフコアー リミテッド Network computer with multiple built-in rings

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS5162343U (en) * 1974-11-09 1976-05-17
JPS5425356U (en) * 1977-07-20 1979-02-19
JP2022526929A (en) * 2019-03-27 2022-05-27 グラフコアー リミテッド Network computer with multiple built-in rings
JP2022527066A (en) * 2019-03-27 2022-05-30 グラフコアー リミテッド Incorporation of rings in circular computer networks

Similar Documents

Publication Publication Date Title
KR920704231A (en) Cluster configuration for highly parallel scalar / vector multiprocessor systems
EP3729261B1 (en) A centralized-distributed mixed organization of shared memory for neural network processing
JPH0619760B2 (en) Information processing equipment
JP2003271574A (en) Data communication method for shared memory type multiprocessor system
JPH054258U (en) Parallel distributed processor
CN116745754A (en) System and method for accessing remote resource
WO2001025941A1 (en) Multiprocessor computer systems with command fifo buffer at each target device
JPH0358163A (en) Loosely coupled multiprocessor system
JP2643116B2 (en) Main memory controller
JPH03131951A (en) Data transfer system
JPH03204753A (en) Dma controller
JPS6029139B2 (en) Connection method between processing units
JPH0646413B2 (en) Data processing processor
JPH02108150A (en) Parallel decentralized processor of computer
JPH0926945A (en) Information processor
JPH01263858A (en) Multi-processor system
JPS6116115B2 (en)
JPS61281344A (en) Access controlling method of two port memory
JPH0214741B2 (en)
JP2001290790A (en) Disk controller
JPH04167191A (en) Fast data transfer system in decentralized memory type parallel computer
JPH02211571A (en) Information processor
JPS61153770A (en) Image processor
JPH0570867B2 (en)
JPS62274346A (en) Memory access circuit

Legal Events

Date Code Title Description
A300 Withdrawal of application because of no request for examination

Free format text: JAPANESE INTERMEDIATE CODE: A300

Effective date: 19950907