JPS59195757A - Load distribution control system - Google Patents

Load distribution control system

Info

Publication number
JPS59195757A
JPS59195757A JP7119083A JP7119083A JPS59195757A JP S59195757 A JPS59195757 A JP S59195757A JP 7119083 A JP7119083 A JP 7119083A JP 7119083 A JP7119083 A JP 7119083A JP S59195757 A JPS59195757 A JP S59195757A
Authority
JP
Japan
Prior art keywords
load
slave
processor
counter
slave processor
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.)
Granted
Application number
JP7119083A
Other languages
Japanese (ja)
Other versions
JPH0519742B2 (en
Inventor
Mitsuyoshi Sasakura
三好 笹倉
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.)
Casio Computer Co Ltd
Original Assignee
Casio Computer Co 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 Casio Computer Co Ltd filed Critical Casio Computer Co Ltd
Priority to JP7119083A priority Critical patent/JPS59195757A/en
Publication of JPS59195757A publication Critical patent/JPS59195757A/en
Publication of JPH0519742B2 publication Critical patent/JPH0519742B2/ja
Granted legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F15/00Digital computers in general; Data processing equipment in general
    • G06F15/16Combinations of two or more digital computers each having at least an arithmetic unit, a program unit and a register, e.g. for a simultaneous processing of several programs
    • G06F15/177Initialisation or configuration control

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Hardware Design (AREA)
  • Theoretical Computer Science (AREA)
  • Software Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Multi Processors (AREA)

Abstract

PURPOSE:To hold a proper load distribution state to execute an efficient distributed processing by recognizing the load state of each slave processor on a basis of the state of a holding job and the state of the idle number of each slave processor. CONSTITUTION:Each of slave processors 2001-200n is provided with the first counter 203 indicating the number of holding jobs and the second counter 201 indicating the idle number. By a master processor 100, the number of holding jobs indicated by the first counter 203 in each of slave processor 2001-200n is inputted to a comparison control part 102 and the idle number indicated by the second counter 201 is inputted to a comparison control part 103 at intervals of a certain time. In the comparison control part 102, the slave processor 200i having the smallest number of holding jobs is a load turn-on object. If plural slave processors have the smallest number of holding jobs, the slave processor 200j having the maximum idle number obtained by the comparison control part 103 is a load turn-on object.

Description

【発明の詳細な説明】 〔発明の技術分野〕 本発明はマルチプロセッサシステムにおける負荷分散制
御方式に関する。
DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a load distribution control method in a multiprocessor system.

〔従来技術とその問題点〕[Prior art and its problems]

従来、負荷分散方式を採るマルチプロセッサシステムに
おいては、負荷(、JOB)の分散制御を行なうプロセ
ッサ(以下マスクプロセッサと称す)が、その制御対像
となる複数のプロセッサ(以下スレーブプロセッサと称
す)の負荷配分を決定する際、その判断時機は一定して
おらず、成るe′?の処理状態となった時点で、JOB
管理テーブルを参照し、その際の各スレーブプロセッサ
におけるJOB待機状態から、負荷の配分(JOBの割
当て)を決定してhた。
Conventionally, in a multiprocessor system that employs a load distribution method, a processor (hereinafter referred to as a mask processor) that performs load (JOB) distributed control controls multiple processors to be controlled (hereinafter referred to as slave processors). When deciding on load distribution, the decision timing is not constant and becomes e'? When the processing state is reached, the JOB
With reference to the management table, load distribution (JOB assignment) was determined based on the JOB standby state in each slave processor at that time.

しかしながら、このような従来の負荷分散制御手段にお
いては、負荷配分の判定期間が一定で々く、成る特定の
処理状態となった際において、各スレーブプロセッサの
JOB待機状態から負荷の配分が決定されるため、常に
平均化した状態を保って適正な負荷分散状態を維持する
ことができず、成る時機、負荷配分に不当な偏りが生じ
たり、アイドリング状轢となるスレーブプロセッサが多
くなったりする等の不具合が生じて、効率の良い分散処
理が期待できないという問題があった。
However, in such conventional load distribution control means, the load distribution determination period is constant and when a specific processing state is reached, the load distribution is determined from the JOB waiting state of each slave processor. As a result, it is not possible to always maintain an averaged state and an appropriate load distribution state, which may lead to unreasonable bias in load distribution or an increase in the number of slave processors idling. There was a problem in that efficient distributed processing could not be expected due to this problem.

〔発明の目的〕[Purpose of the invention]

本発明は上記実情Vcaみなされたもので、常に平均化
した適正な負荷分散状態を維持でき、効率の良い分散処
理を実行できるようにした負荷分散制御方式を提供する
ことを目的とする。
The present invention is based on the above-mentioned actual situation Vca, and an object of the present invention is to provide a load distribution control method that can always maintain an averaged and appropriate load distribution state and execute efficient distributed processing.

〔発明の要点〕[Key points of the invention]

本発明は、マスタプロセッサが、インターバルタイマに
よる一定時間毎に、各スレーブ7’ 。
In the present invention, the master processor checks each slave 7' at regular intervals using an interval timer.

セッサの待機JOB状態と、アイPル回数状態とカラ、
各スレーブプロセッサの負荷状態を認識する構成として
、常に平均化された状態で適正な負荷分散状態を維持す
ることができるようにしたもので、これによシ、分散処
理を効率良く実行することのできる経済性に優れた処理
性能ノ高いマルチプロセッサシステムが構築でキル。
Sessa's standby JOB status, eye roll count status and color,
The configuration recognizes the load status of each slave processor, making it possible to maintain an appropriate load distribution status in an averaged state at all times.This makes it possible to efficiently execute distributed processing. A multiprocessor system with high processing performance and excellent economic efficiency can be constructed.

〔発明の実施例〕[Embodiments of the invention]

以下図面を参照して本発明の一実施例を説明する。第1
図は本発明の一実施例を示すブロック図である。図中、
100はマスタプロセッサ(M−CPU)、2n01,
2θ0.−200nはスレーブプロセッサ(S −CP
U )である。101乃至1n、9はマスタプロセッサ
1θθに設けられた本発明に係わる部分の構成要素をな
すもので、101は一定時間毎に読込み制御のためのタ
イミング信号(CTP)を発生するインターバルタイマ
である。
An embodiment of the present invention will be described below with reference to the drawings. 1st
The figure is a block diagram showing one embodiment of the present invention. In the figure,
100 is a master processor (M-CPU), 2n01,
2θ0. -200n is a slave processor (S-CP)
U ). Reference numerals 101 to 1n and 9 constitute components of a portion related to the present invention provided in the master processor 1θθ, and 101 is an interval timer that generates a timing signal (CTP) for read control at regular intervals.

102はスレーブプロセッサ20 /)、 、 20 
/)。
102 is a slave processor 20 /), , 20
/).

・・・200n各々に設けられた後述する待機JOB数
カ数カメンタ容を取込み、最小待機JOB数のスレーブ
プロセッサを認識するJOB数比較制御部であり、図示
しないテーブルを参照して最少待機JOB 11状態に
あるスレーブプロセッサのアドレス(s−cpu−AD
DRA)と、最少待機JOTl数状態にあるスレーブプ
ロセッサが複数台存在するか否かを示す判定信号(MM
S :ここではMMS = ”1”の際に″「、MMS
 = ” O”の際に1無”)を出力する。
. . . A JOB number comparison control unit that takes in the number of standby jobs (described later) provided in each of the 200n and recognizes the slave processor with the minimum number of standby jobs, and refers to a table (not shown) to determine the minimum number of standby jobs 11. Address of slave processor in state (s-cpu-AD
DRA) and a determination signal (MM
S: Here, when MMS = “1”, “”, MMS
= Outputs 1 (nothing) when “O”.

103はスレーブプロセッサ2001,20n鵞・・・
200n各々に設けられた後述する負荷カウンタのラッ
チ出力を取込み、その内容にもとづいて、滑少待機JO
B数状態にあるスレーブプロセッサが複数含有る際、即
ちJOB数比較制御部102より″1”レベルの判定信
号(MMS =″1”)が出力された際に、その複数台
のスレーブプロセッサのうちから、負荷を配分すべき1
台のスレーブプロセッサを決定する負荷カウント数比較
制御部であシ、図示しないテーブルを参照して負荷カウ
ント数(アイドリング回数)の最も大キいスレーブプロ
セッサのアドレス(S−CPU−ADDRB)を出力す
る。201乃至203はスレーブプロセッサz n /
)1 、 z n o2・・・200nに設けられた本
発明に係わる部分の構成要素をなすもので、201はイ
ンターバルタイマ101のタイミング信号(CTP )
発生周期をもって、その期間内におけるアイ−リング(
IDLING)の回数を計数する負荷カウンタ、202
は上記期間毎の負荷カウンタ20ノの計数値(すなわち
アイドリング回数)をラッチするラッチ回路(LATC
H)である。このラッチ回路202に貯えられた負荷カ
ウンタ201の計数値すなわちアイドリング5− 回数データは負荷カウント数比較制御部103に送られ
る。203はJOB投入の都度カウントアツプ(+1)
され、JOB終了の都度カウントダウン(−1)される
待機JOB数カ数カメンタシ、この計数値すなわち待機
JOB数データはJOB数比較制御部102に送られる
103 is the slave processor 2001, 20n...
The latch output of the load counter (described later) provided in each of the 200n is taken in, and based on the contents, the slippage standby JO
When there are multiple slave processors in the B number state, that is, when the JOB number comparison control unit 102 outputs a "1" level determination signal (MMS = "1"), one of the multiple slave processors The load should be distributed from 1
The load count comparison control unit determines the number of slave processors, and outputs the address (S-CPU-ADDRB) of the slave processor with the largest load count (idling number) by referring to a table (not shown). . 201 to 203 are slave processors z n /
)1, zno2...200n, which constitutes a component related to the present invention, and 201 is the timing signal (CTP) of the interval timer 101.
Eye ring (with a period of occurrence) within that period
A load counter that counts the number of times (IDLING) 202
is a latch circuit (LATC) that latches the count value of the load counter 20 (i.e., the number of idling times) for each period.
H). The count value of the load counter 201 stored in the latch circuit 202, that is, the 5-times of idling data is sent to the load count comparison control section 103. 203 counts up (+1) every time a job is submitted.
The number of waiting jobs is counted down (-1) each time a job ends, and this counted value, that is, the number of waiting jobs data is sent to the job number comparison control section 102.

第2図は本発明の一実施例における動作を説明するため
のフローチャートである。
FIG. 2 is a flowchart for explaining the operation in one embodiment of the present invention.

ここで第1図、及び第2図を参照して一実施例の動作を
説明する。マスタプロセッサ100に設けられたインタ
ーバルタイマ101は、一定時間毎に読込み制御のため
のタイミング信号(CTP)を発生し、このタイミング
信号(CTP)が各スレーブプロセッサ200..20
0g・・・200nに供給される。各スレーブプロセッ
サzoom。
The operation of one embodiment will now be described with reference to FIGS. 1 and 2. An interval timer 101 provided in the master processor 100 generates a timing signal (CTP) for read control at regular intervals, and this timing signal (CTP) is transmitted to each slave processor 200 . .. 20
0g...200n is supplied. Each slave processor zoom.

200霊・・・200nに設けられた負荷カウンタ20
1・・・はそれぞれ上記インターバルタイマ101から
のタイミング信号(CTP)を受けて初期化された後、
自己プロセッサのアイドリング回数を計数する。又、待
機JOB数カ数カメンタ3・・・は自6− 己プロセッサへのJOB投入投入力ウントアツプ(+1
)され、JOB終了毎にカウントダウン(−1)されて
、常に現在の待機JOB数を示している。上記負荷カウ
ンタ201・・・で計数された負荷カウント数データ(
アイげリング回数データ)は、初期化に先立ち上記タイ
ミング信号(CTP)の立上りに同期してラッチ回路2
02・・・にラッチされる。更にこの各ラッチ回路20
2・・・【ラッチされた負荷カウント数データは、上記
各待機JOB数カ数カメンタ3・・・で計数された待機
JOB 数データト共にマスタプロセッサ100に送ら
れる。そして各スレーブプロセッサ2001゜2002
・・・200nの待機JOB数データはそれぞれJOB
数比較制御部102に入力され、負荷カウント数データ
はそれぞれ負荷カウント数比較制御部103に入力され
る。 JOB数比較制御部102は入力された各スレー
ブプロセッサ2001.200@ =・200nの待機
JOB数データと図示しないテーブルの参照と釦より、
命も待機JOB数の少ないスレーブプロセッサを示すプ
ロセッサアドレス(S−CPU−ADDRA)を出力す
るとともに、その最少待機JOB数状態にあるスレーブ
プロセッサが複数台存在するか否かを示す判定信号(V
MS)を出力する。ここで最少待機JOII数のスレー
ブプロセッサが1台である際、すなわち上記判定信号(
m、as)が単数のプロセッサであることを示す@ (
MMS−′O″)となっている際は1、ToB数叱較制
御部102より出力されたデロセy t 7 Y L/
 、X (S−CPU−ADDRA)が帰路的な負荷投
入対象トなるスレーブプロセッサを示すプロセツサア1
弓として用いられ、このデロセツサアPレスヲモつスレ
ーブプロセッサ2001(1−111,・・・n)が負
荷投入対象となる。この際、上記判定後において、負荷
投入対象となるスレーブプロセッサ2001に負荷が投
入されると、そのスレーブプロセッサ2001に対応し
て設けられた負荷カウンタがカウントアツプ(+1)さ
れる。又、最少待機JOB数のスレーブプロセッサが複
数台存在する際、すなわち上記判定信号(WS)が複数
のプロセッサであることを示す値(MMS =”1”)
となっている際は、次に負荷カウント数比較制御部10
.9より得られる、最大負荷カウント数を示す、すなわ
ちアイドリング回数の盃も多重スレーブプロセッサを示
すプロー1= ツt アl& L/ ス(S−CPU−
ADDRB)が最終的な負荷投入対象となるスレーブプ
ロセッサアドレスとして用いられ、このプロセッサアド
レスをもつスレーブプロセッサ200j(j=12−−
 n ) カ負荷投入対隼となる。この一連の動作フロ
ーを第2図に示す。
200 souls...Load counter 20 installed at 200n
1... are each initialized by receiving the timing signal (CTP) from the interval timer 101, and then
Count the number of times the self-processor is idling. In addition, the number of standby JOBs is 3..., which is the number of JOBs input to the self-processor (+1).
), and is counted down (-1) every time a job ends, and always indicates the current number of waiting jobs. Load count data counted by the load counter 201 (
Prior to initialization, the eigelring number data) is sent to the latch circuit 2 in synchronization with the rising edge of the timing signal (CTP).
02...is latched. Furthermore, each latch circuit 20
2... [The latched load count data is sent to the master processor 100 together with the standby job count data counted by each standby job count mentor 3. And each slave processor 2001゜2002
...200n standby JOB number data is JOB
The load count data is input to the load count comparison control section 102, and the load count data is input to the load count comparison control section 103, respectively. The job number comparison control unit 102 uses the input waiting job number data of each slave processor 2001.200@=200n, references to a table (not shown), and a button.
It outputs a processor address (S-CPU-ADDRA) indicating a slave processor with a small number of waiting jobs, and also outputs a determination signal (V
MS) is output. Here, when the number of slave processors with the minimum number of standby JOIIs is one, that is, the above judgment signal (
m, as) is a single processor @ (
MMS-'O''), it is 1, and the output from the ToB number comparison control unit 102 is 1.
, X (S-CPU-ADDRA) indicates the slave processor to which the return load is applied
The slave processors 2001 (1-111, . . . n) that are used as processors are loaded. At this time, after the above determination, when a load is applied to the slave processor 2001 to which the load is applied, the load counter provided corresponding to that slave processor 2001 is incremented (+1). Also, when there are multiple slave processors with the minimum number of waiting JOBs, that is, the determination signal (WS) is a value indicating that there are multiple processors (MMS = "1").
If so, then load count comparison control section 10
.. 9, which indicates the maximum load count number, that is, the number of idling times, also indicates multiple slave processors.
ADDRB) is used as the slave processor address to which the final load is applied, and the slave processor 200j (j=12--
n) Power input vs. Hayabusa. The flow of this series of operations is shown in FIG.

このようにして、インターバルタイマ10ノで定まる一
定期間毎に、各スレーブプロセッサ200!、200!
−20Onの待機JOB PIと負荷カウント数(74
1972回数)とにより、次の期間内における負荷投入
対象となるスレーブプロセッサが決定される。これKよ
り、成る時機に待機JOB数が偏ったり、アイp IJ
ソング態となるプロセッサが増えたシする不具合が解消
され、システム稼働期間に亘って、常に平均化された適
正負荷分散状態を維持でき、効率の良い9− 分散処理が実行できる。
In this way, each slave processor 20! , 200!
-20On standby JOB PI and load count number (74
1972 times), the slave processor to which the load will be applied within the next period is determined. From this, the number of waiting jobs will be biased at the right time, and Ip IJ
The problem caused by an increase in the number of processors in a song state is resolved, a properly averaged load distribution state can always be maintained over the system operating period, and efficient 9-way distributed processing can be executed.

尚、上記した実施例忙おいては、最少待機JOB M状
態にあるスレーブプロセッサが単数である際は、そのス
レーブプロセッサを最終的な負荷投入対象とし、最少待
機JOB数状態にあるスレーブプロセッサが複数存在す
る際は、負荷カウント数(アイドリング回数)の最も大
きいスレーブプロセッサを帰路的な負荷投入対象として
いるが、これに限らず、例えば帰少待機JOB数状態に
あるスレーブプロセッサが複数存在する際は、次にその
各スレーブプロセッサの負荷カウント数(741972
回数)を認識し、負荷カウント数の大きいスレーブプロ
セッサを最終的な負荷投入対象とする構成としてもよい
In the above-mentioned embodiment, when there is only one slave processor in the minimum waiting JOB M state, that slave processor is the final load target, and when there are multiple slave processors in the minimum waiting JOB number state. When it exists, the slave processor with the largest load count (number of idling times) is targeted for repatriation load. , then the load count number of each slave processor (741972
It is also possible to adopt a configuration in which the slave processor with a large load count is recognized as the final load injection target.

〔発明の効果〕〔Effect of the invention〕

以上詳記したように本発明によれば、マスタプロセッサ
が複数のスレーブプロセッサの負荷分散制御を行なうマ
ルチプロセッサシステムにおいて、マスタプロセッサが
、一定期間毎に、各スレーブプロセッサの待機JOB数
と負荷カウ10− ント数(アイドリング回数)とによシ、次の期間内にお
ける負荷投入対象となるスレーブプロセッサを決定する
構成としたことによ)、システム稼働期間に亘って、常
に平均化された適正負荷分散状態を維持でき、効率の良
い分散処理が実行できる負荷分散制御方式が提供できる
As described in detail above, according to the present invention, in a multiprocessor system in which a master processor performs load distribution control of a plurality of slave processors, the master processor checks the number of waiting JOBs of each slave processor and the load counter 10 at regular intervals. - Appropriate load distribution that is always averaged over the system operation period (depending on the number of idling operations) and determining the slave processor to which the load will be applied within the next period) A load distribution control method that can maintain the state and execute efficient distributed processing can be provided.

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

第1図は本発明の一実施例を示すブロック図、第2図は
上記実施例の動作を説明するためのフローチャートであ
る。 100・・・マスタプロセッサ(M−CPU)、101
・・・インターバルタイマ、102・・・JOB数比較
制御部、103・・・負荷カウント数比較制御部、20
01.2(10,・・・200n・・・スレーブプロセ
ッサ(S−CPU)、20!・・・負荷カウンタ、20
2・・・ラッチ回路、203・・・待機JOB数カ数カ
メンタ願人代理人  弁理土鈴 江 武 彦11−
FIG. 1 is a block diagram showing one embodiment of the present invention, and FIG. 2 is a flowchart for explaining the operation of the above embodiment. 100... Master processor (M-CPU), 101
...Interval timer, 102...JOB number comparison control section, 103...Load count number comparison control section, 20
01.2 (10,...200n...Slave processor (S-CPU), 20!...Load counter, 20
2...Latch circuit, 203...Waiting JOB several mentor applicant agent Patent attorney Tosuzu E Takehiko 11-

Claims (1)

【特許請求の範囲】[Claims] マスタプロセッサが複数のスレーブプロセッサの負荷分
散制御を行なうマルチプロセッサシステムにお騒で、前
記スレーブプロセッサ各々に、待機ジョブ数を示す第1
のカウンタ、及びアイドル回数を示す第2のカウンタを
設け、前記マスタプロセッサが、一定期間毎に、前記ス
レーブプロセッサ各々の第1のカウンタで示される待@
ゾヨプ数と、第2のカウンタで示されるアイドル回数と
をもとに1次の一定期間での負荷分散状態を行なうこと
を特徴とした負荷分散制御方式。
In a multiprocessor system in which a master processor performs load distribution control of a plurality of slave processors, a first message indicating the number of waiting jobs is sent to each of the slave processors.
and a second counter indicating the number of idle times, and the master processor is configured to count the wait times indicated by the first counter of each of the slave processors at regular intervals.
A load distribution control method characterized in that a load distribution state is performed in a primary fixed period based on the number of zoyoops and the number of idle times indicated by a second counter.
JP7119083A 1983-04-22 1983-04-22 Load distribution control system Granted JPS59195757A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP7119083A JPS59195757A (en) 1983-04-22 1983-04-22 Load distribution control system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP7119083A JPS59195757A (en) 1983-04-22 1983-04-22 Load distribution control system

Publications (2)

Publication Number Publication Date
JPS59195757A true JPS59195757A (en) 1984-11-06
JPH0519742B2 JPH0519742B2 (en) 1993-03-17

Family

ID=13453492

Family Applications (1)

Application Number Title Priority Date Filing Date
JP7119083A Granted JPS59195757A (en) 1983-04-22 1983-04-22 Load distribution control system

Country Status (1)

Country Link
JP (1) JPS59195757A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1111738A1 (en) * 1998-08-31 2001-06-27 Rohm Co., Ltd. Semiconductor device and substrate for semiconductor device

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS51144543A (en) * 1975-06-06 1976-12-11 Toshiba Corp Processing system of composite computer system
JPS57757A (en) * 1980-06-04 1982-01-05 Hitachi Ltd Job execution schedule system
JPS57120167A (en) * 1981-01-19 1982-07-27 Mitsubishi Electric Corp Load scattering method for composite computer system

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS51144543A (en) * 1975-06-06 1976-12-11 Toshiba Corp Processing system of composite computer system
JPS57757A (en) * 1980-06-04 1982-01-05 Hitachi Ltd Job execution schedule system
JPS57120167A (en) * 1981-01-19 1982-07-27 Mitsubishi Electric Corp Load scattering method for composite computer system

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1111738A1 (en) * 1998-08-31 2001-06-27 Rohm Co., Ltd. Semiconductor device and substrate for semiconductor device
US6717256B1 (en) 1998-08-31 2004-04-06 Rohm Co., Ltd. Mounting structure for semiconductor device having entirely flat leads
EP1111738A4 (en) * 1998-08-31 2006-01-11 Rohm Co Ltd Semiconductor device and substrate for semiconductor device

Also Published As

Publication number Publication date
JPH0519742B2 (en) 1993-03-17

Similar Documents

Publication Publication Date Title
US7137115B2 (en) Method for controlling multithreading
Kalyanasundaram et al. Speed is as powerful as clairvoyance
WO2017101475A1 (en) Query method based on spark big data processing platform
EP1916601A2 (en) Multiprocessor system
KR20020018309A (en) Cpu scheduling method
US20220083375A1 (en) Method and apparatus for scheduling task processing entity
CN112346834A (en) Database request processing method and device, electronic equipment and medium
JPS63238630A (en) Interruption controller for microprocessor
Nicol et al. Parallel simulation of timed petri-nets
RU2134446C1 (en) Method for controlling primitive program overflow by messages in multiprocessor control system (versions)
JPS59195757A (en) Load distribution control system
CN111158875B (en) Multi-module-based multi-task processing method, device and system
JP2001117786A (en) Process scheduling device and process scheduling method
Bampis et al. From preemptive to non-preemptive speed-scaling scheduling
JP2006277744A (en) Structuring staggered job termination pattern
Chretienne Tree scheduling with communication delays
Shah et al. Phase level energy aware map reduce scheduling for big data applications
CN110990490A (en) Method, device, equipment and medium for posting block chain in block chain network
JPH09179834A (en) Scheduling method of parallel system for process
Kobayashi Improved lower bounds for online scheduling to minimize total stretch
CN112559054B (en) Method and computing system for synchronizing instructions
Nakechbandi et al. Bounding the makespan of best pre-schedulings of task graphs with fixed communication delays and random execution times on a virtual distributed system.
Özkan et al. On the control of fork-join networks
JP2001282555A (en) Method for managing process of computer system
JPH04355867A (en) System for distributed processing