JPS6134643A - Buffer control system - Google Patents

Buffer control system

Info

Publication number
JPS6134643A
JPS6134643A JP15660984A JP15660984A JPS6134643A JP S6134643 A JPS6134643 A JP S6134643A JP 15660984 A JP15660984 A JP 15660984A JP 15660984 A JP15660984 A JP 15660984A JP S6134643 A JPS6134643 A JP S6134643A
Authority
JP
Japan
Prior art keywords
bank
buffer
word
identification number
address
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.)
Pending
Application number
JP15660984A
Other languages
Japanese (ja)
Inventor
Hideaki Fujimaki
藤巻 秀明
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.)
Fujitsu Ltd
Original Assignee
Fujitsu 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 Fujitsu Ltd filed Critical Fujitsu Ltd
Priority to JP15660984A priority Critical patent/JPS6134643A/en
Publication of JPS6134643A publication Critical patent/JPS6134643A/en
Pending legal-status Critical Current

Links

Landscapes

  • Memory System Of A Hierarchy Structure (AREA)

Abstract

PURPOSE:To control a substantial main storage access time of each virtual computer by providing a means for limiting the quantity of a buffer which each virtual computer can use, to a bank unit of the buffer. CONSTITUTION:When selecting a bank for writing a new data block, a control circuit 8 refers to not only control information 7 but also identification number indicating registers 11-1-11-n and an identification number register 12, and knows a bank which a virtual computer can use, by the registers 11-1-11-n. Also, the bak whose access is the oldest in its banks is determined based on the control information 7, and a word of its bank is informed to a buffer registration control circuit 9. The circuit 9 writes a data block and an address in the word of this bank. In this way, a word having a bank in which a tag part 6 coincides with a high rank address 3 at the time of read-out from the buffer, and also an identification number which coincides with a register 12 corresponding to the bank is selected as an object data block word.

Description

【発明の詳細な説明】 〔産業上の利用分野〕 本発明は仮想計算機を動作する処理装置のバッファの制
御方式に関する。
DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a buffer control method for a processing device that operates a virtual machine.

仮想計算機方式は、1計算機システム内に、1又は複数
の仮想的な計算機システム、即ち仮想計算機、を構成し
、該仮想計算機は互いに独立の計算機であるように稼動
することのできる、計算機システムの制御方式として広
く用いられる。
A virtual computer system is a computer system in which one or more virtual computer systems, that is, virtual computers, are configured within one computer system, and the virtual computers can operate as if they were mutually independent computers. Widely used as a control method.

複数の仮想計算機は、論理的には独立であるが、物理的
には同一の計算機ハードウェアを時分割的に共用するの
で、処理速度等の面でiJ和互に影響がある。
Although the plurality of virtual machines are logically independent, they physically share the same computer hardware in a time-sharing manner, so they mutually affect iJ in terms of processing speed and the like.

このために、各仮想言1算機で実行される業務等からの
要求に応じて、各仮想旧算機に割り当′ζる計算機資源
を加減するごとによって、仮想計算機の性能を1lil
J御する手段が必要とされる。
For this purpose, the performance of the virtual machine can be increased by 1 liter by increasing or subtracting the computer resources allocated to each virtual machine in accordance with the requests from the work etc. executed on each virtual machine.
A means of controlling J is required.

〔従来の技術と発明が解決しようとする問題点〕仮想計
算機の性能を制御する手段として、従来各仮想計算機に
対する上記10の割当■を制御する方法が使用されてい
る。
[Prior Art and Problems to be Solved by the Invention] Conventionally, as a means for controlling the performance of virtual machines, the method of controlling the above-mentioned 10 assignments (3) to each virtual machine has been used.

即ち、仮想計算機方式においては、主記憶装置を分割し
て各仮想計算機に割り当て、各仮想剖算機は割り当てら
れた上記1.0領域を使って、それぞれ自月の仮想計算
機を構成する主起俯、装置とみなした制御をする。
In other words, in the virtual computer method, the main storage device is divided and allocated to each virtual computer, and each virtual computer uses the above-mentioned 1.0 area allocated to it to store the main memory that constitutes its own virtual computer. Look down and control it as if it were a device.

従って、この割当量を加減ずれば、im常の計算機シス
テムの主記憶装置の大きさを増減した場合に類似した性
能制御の効果が得られる。
Therefore, by increasing or decreasing this allocation amount, a performance control effect similar to that obtained by increasing or decreasing the size of the main memory of a regular computer system can be obtained.

しかし、この効果はIと較的間接的であるので、更に単
独に、又は上記手段と併用して性能をより的確に制御す
る手段が要望されている。
However, since this effect is relatively indirect, there is a need for a means to more accurately control performance either alone or in combination with the above means.

公知のように、多くの計算機システムの中央処理装置等
には、主記憶装置の情報のコピーを保持する、比較的小
容量で高速なへソファを持つことにより、主記憶アクセ
スの実質的な高速化をはかる技術が使用されている。
As is well known, the central processing unit of many computer systems has a relatively small capacity and high speed memory that holds a copy of the information in the main memory, thereby making it possible to access the main memory at substantially higher speeds. Technology is being used to change the situation.

第2図はこのようなバッファの構成の一例である。公知
のバッファの制御を、主記憶読出しアクセス要求の場合
について以下に簡単に述べる。
FIG. 2 shows an example of the structure of such a buffer. Known buffer control will be briefly described below for the case of a main memory read access request.

アクセスを要求する記1aアドレスは、アドレスレジス
タ2に置かれるものとする。
It is assumed that the address 1a to which access is requested is placed in the address register 2.

通常、バッファは一定長のデータブロック部5と該ブロ
ックの上記1.1装置上の記憶アドレスの一部(例えば
アドレスレジスタ2の記憶アドレスの上位ビット列3)
を保持するタグ部6とを内容とする語を単位として、同
数の語からなる複数バンク1−1〜1−nから構成され
る。
Normally, the buffer includes a data block section 5 of a fixed length and a part of the storage address of the block on the device described in 1.1 (for example, the upper bit string 3 of the storage address of the address register 2).
Each bank is composed of a plurality of banks 1-1 to 1-n each having the same number of words, with each word containing the tag section 6 holding the same number of words as a unit.

記憶アドレスの下位ビット列は、バンク内の1語を選択
するアドレスとして使用され、へソファアクセスにおい
て、このアドレスにより選択される各バンク1−1〜1
−nの各1語が一斉に読み出され、その中からタグ部に
上位アドレス3と同一の内容を持つ語が目的のデータブ
ロックの語として選択される。
The lower bit string of the storage address is used as an address to select one word in the bank, and in the Sofa access, each bank 1-1 to 1 selected by this address is used.
-n words are read out all at once, and a word whose tag portion has the same content as the upper address 3 is selected as the word of the target data block.

もし、上記動作の結果、バッファにアドレスの−・致す
るものが無い場合には、主起m装置に実際にアクセスし
てデータを得る必要があるので、アクセス時間は長くな
る。
If, as a result of the above operation, there is no matching address in the buffer, it is necessary to actually access the host m device to obtain the data, which increases the access time.

この場合、上記10装置から読み出されたデータブロッ
クは、前記と同様にしてアドレスされる全バンクの詔の
中から選択される1語に、アドレスの」二位ヒント列と
ともに格納される。
In this case, the data blocks read from the ten devices described above are stored in one word selected from among the edicts of all banks addressed in the same manner as described above, together with the second-place hint string of the address.

この場合の選択のために、例えばア1ルス下位ビット列
4でアドレスされる語群に対して制御情報7が設けられ
、該情報はバッファアクセス中に常に制御回路8により
更新されていて、その情報により、同一語群の中の語で
、最近のアクセスが最も古い時刻に行われた詔を決定で
きるようにされている。
For selection in this case, for example, control information 7 is provided for the word group addressed by the alphanumeric lower bit string 4, and this information is constantly updated by the control circuit 8 during buffer access. This makes it possible to determine which edict was most recently accessed among words within the same word group.

この情報を制御回路8に入力して1語を選択し、バッフ
ァ登録制御回路9により、選択した語に書込みデータレ
ジスタ10−1〜10−〇のデータを書込む。
This information is input to the control circuit 8 to select one word, and the buffer registration control circuit 9 writes the data in the write data registers 10-1 to 10-0 to the selected word.

以上により明らかなように、バッファのバンクの数によ
って、目的のデータがバッファから得られる確率は変化
し、従って実質的な主記憶アクセス時間が変化するので
、計算機の性能に直接に影響するが、従来上記仮想計算
機の性能の制御を、バッファを制御することによって実
現する方式は無かった。
As is clear from the above, the probability that the desired data can be obtained from the buffer changes depending on the number of buffer banks, and therefore the actual main memory access time changes, which directly affects the performance of the computer. Conventionally, there has been no method for controlling the performance of a virtual machine by controlling a buffer.

〔問題点を解決するための手段〕[Means for solving problems]

前記の問題点は、主記憶装置の情報のコピーを保持し、
複数のバンクよりなるバッファを有する処理装置上で仮
想計算機を動作する場合において、上記バンクを個別に
仮想計算機に割り当てる本発明のバッファ制御方式によ
って解決される。
The above problem is that keeping a copy of the information in main memory,
In the case where a virtual machine is operated on a processing device having a buffer consisting of a plurality of banks, the problem is solved by the buffer control method of the present invention in which the banks are individually assigned to the virtual machine.

〔作用〕[Effect]

即ち、各仮想計算機が使用できるバッファの量を、バッ
ファのバンク単位に制限する手段が設けられるので、こ
の手段により各仮想計算機の実質的な主記憶アクセス時
間を制御することが可能となり、仮想計算機の性能の制
御をより直接的に行うことができる。
In other words, since a means is provided to limit the amount of buffer that each virtual machine can use in units of buffer banks, this means makes it possible to control the actual main memory access time of each virtual machine. performance can be controlled more directly.

〔実施例〕〔Example〕

第1図は本発明一実施例のブロック図である。 FIG. 1 is a block diagram of one embodiment of the present invention.

図において、第2図と同じ部分は同一の符号によって示
す。
In the figure, the same parts as in FIG. 2 are designated by the same reference numerals.

バッファのバンクを仮想計算機に割り当てる制御は、バ
ッファへのアクセスにおいて、該アクセスを要求した仮
想計算機に割り当てられているバンクのみに対して、読
出し及び書込みを行うように制御することにより実現さ
れる。
The control of allocating a buffer bank to a virtual machine is achieved by controlling, when accessing a buffer, reading and writing only to the bank that is allocated to the virtual machine that requested the access.

この制御のために、各仮想計算機には、それらを特定す
る識別番号を付し、現に実行中の仮想計算機の識別番号
が、識別番号レジスタ12に表示される。
For this control, each virtual machine is given an identification number to identify it, and the identification number of the virtual machine currently being executed is displayed in the identification number register 12.

又、例えば各バンクに対して1個づつ設ける識別番号指
示レジスタ11−1〜11−nを設け、それぞれのバン
クの使用を許す仮想d1算機識別番号を表示する。
Further, for example, identification number instruction registers 11-1 to 11-n are provided, one for each bank, to display virtual d1 computer identification numbers that permit use of each bank.

制御回路8は新たなデータブロックを書込むべきバンク
の選択において、制御情報7の他に識別番号指示レジス
タ11−1〜11−n及び識別番号レジスタ12を参照
し、識別番号レジスタ12に表示する仮想計算機が使用
できるバンクを識別番号指示レジスタ11−1〜11−
nで知り、そのバンクの中で最近のアクセスが最も古い
バンクを制御情報7に基づいて決定し、そのバンクの語
をバッファ登録制御回路9に通知する。
When selecting a bank to write a new data block, the control circuit 8 refers to the identification number instruction registers 11-1 to 11-n and the identification number register 12 in addition to the control information 7, and displays the information in the identification number register 12. Identification number instruction registers 11-1 to 11- indicate banks that can be used by the virtual machine.
n, the bank with the oldest recent access is determined based on the control information 7, and the word of the bank is notified to the buffer registration control circuit 9.

バッファ登録制御回路9は通知されたバンクの語に、前
記のようにデータブロック及びアドレスを書込む。
The buffer registration control circuit 9 writes the data block and address in the notified bank word as described above.

バッファよりの読出し時には、タグ部6と上位アト□レ
ス3が一致するバンクで、且つ識別番号指示レジスタ1
1−1〜11−nのうち、該バンクに対応するレジスタ
に識別番号レジスタ12に一致する識別番号がある語を
、目的のデータブロックの詔として選択する。
When reading from the buffer, the tag part 6 and the upper address □ match the bank, and the identification number instruction register 1
Among 1-1 to 11-n, a word whose register corresponding to the bank has an identification number matching the identification number register 12 is selected as the edict of the target data block.

以上の制御により、各仮想計算機の要求したデータブロ
ックは、識別番号指示レジスタ11−1〜11−nで指
定されているバンクにのめ保持されることになり、仮想
計算機に対するバッファのバンク割当が実現される。
With the above control, the data blocks requested by each virtual machine are held in the banks specified by the identification number instruction registers 11-1 to 11-n, and the buffer bank allocation to the virtual machine is changed. Realized.

〔発明の効果〕〔Effect of the invention〕

以」二の説明から明らかなように本発明によれば、仮想
計算機の性能の制御を従来より的確に行う手段が提供さ
れ、仮想計算機システムの利用性を向」ニするという著
しい工業的効果がある。
As is clear from the following explanation, the present invention provides a means for controlling the performance of a virtual machine more accurately than before, and has the remarkable industrial effect of improving the usability of the virtual machine system. be.

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

第1図は本発明一実施例の構成ブロック図、第2図は従
来のバッファの構成ブロック図である。 図おいて、 1=1〜1−nはバンク、  2はアドレスレジスタ、
7は制御情報、     8は制御回路、9はバッファ
登録制御回路、 10−1〜10−nは書込めデータレジスタ、11−1
〜11−nは識別番号指示レジスタ、12は識別番号レ
ジスタを示す。
FIG. 1 is a block diagram of the configuration of an embodiment of the present invention, and FIG. 2 is a block diagram of the configuration of a conventional buffer. In the figure, 1=1 to 1-n are banks, 2 is address register,
7 is control information, 8 is a control circuit, 9 is a buffer registration control circuit, 10-1 to 10-n are write data registers, 11-1
.about.11-n indicates an identification number instruction register, and 12 indicates an identification number register.

Claims (1)

【特許請求の範囲】[Claims] 主記憶装置の情報のコピーを保持し、複数のバンクより
なるバッファを有する処理装置上で仮想計算機を動作す
る場合において、上記バンクを個別に仮想計算機に割り
当てる手段を有することを特徴とするバッファ制御方式
In the case where a virtual computer is operated on a processing device that holds a copy of information in a main storage device and has a buffer consisting of a plurality of banks, the buffer control is characterized by having means for individually allocating the banks to the virtual computer. method.
JP15660984A 1984-07-27 1984-07-27 Buffer control system Pending JPS6134643A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP15660984A JPS6134643A (en) 1984-07-27 1984-07-27 Buffer control system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP15660984A JPS6134643A (en) 1984-07-27 1984-07-27 Buffer control system

Publications (1)

Publication Number Publication Date
JPS6134643A true JPS6134643A (en) 1986-02-18

Family

ID=15631476

Family Applications (1)

Application Number Title Priority Date Filing Date
JP15660984A Pending JPS6134643A (en) 1984-07-27 1984-07-27 Buffer control system

Country Status (1)

Country Link
JP (1) JPS6134643A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS62179663U (en) * 1986-04-30 1987-11-14
JP2008293472A (en) * 2007-04-24 2008-12-04 Fujitsu Ltd Computer device and its cache recovery method

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS62179663U (en) * 1986-04-30 1987-11-14
JP2008293472A (en) * 2007-04-24 2008-12-04 Fujitsu Ltd Computer device and its cache recovery method

Similar Documents

Publication Publication Date Title
US4511964A (en) Dynamic physical memory mapping and management of independent programming environments
US5778434A (en) System and method for processing multiple requests and out of order returns
US3979726A (en) Apparatus for selectively clearing a cache store in a processor having segmentation and paging
US6021462A (en) Methods and apparatus for system memory efficient disk access to a raid system using stripe control information
US6003112A (en) Memory controller and method for clearing or copying memory utilizing register files to store address information
GB1245601A (en) Data storage apparatus
JPH10247163A (en) Computer system and its memory managing method
JPH05225065A (en) Logical map of data object using data space
US3778776A (en) Electronic computer comprising a plurality of general purpose registers and having a dynamic relocation capability
US4737909A (en) Cache memory address apparatus
US6745292B1 (en) Apparatus and method for selectively allocating cache lines in a partitioned cache shared by multiprocessors
US5940868A (en) Large memory allocation method and apparatus
US5293622A (en) Computer system with input/output cache
US4964037A (en) Memory addressing arrangement
JPS6134643A (en) Buffer control system
US6433786B1 (en) Memory architecture for video graphics environment
EP0611462B1 (en) Memory unit including a multiple write cache
US7181587B1 (en) Mapping an arbitrary number of contiguous memory pages at an arbitrary alignment
US5926840A (en) Out-of-order fetching
KR920005296B1 (en) Information processing device
JP3456727B2 (en) Data processing device
KR20190115921A (en) Semiconductor memory module and memory system including semiconductor memory module
JPH1091527A (en) Storage device and storage medium
EP0377969B1 (en) I/O cached computer systems
JP2001075824A (en) Accessing method and recording medium with access processing program recorded thereon