JPH02288975A - Vector arithmetic unit - Google Patents

Vector arithmetic unit

Info

Publication number
JPH02288975A
JPH02288975A JP3801889A JP3801889A JPH02288975A JP H02288975 A JPH02288975 A JP H02288975A JP 3801889 A JP3801889 A JP 3801889A JP 3801889 A JP3801889 A JP 3801889A JP H02288975 A JPH02288975 A JP H02288975A
Authority
JP
Japan
Prior art keywords
address
register
vector
data
address register
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
JP3801889A
Other languages
Japanese (ja)
Inventor
Yukihiro Fujino
藤野 幸広
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.)
NEC Engineering Ltd
Original Assignee
NEC Engineering 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 NEC Engineering Ltd filed Critical NEC Engineering Ltd
Priority to JP3801889A priority Critical patent/JPH02288975A/en
Publication of JPH02288975A publication Critical patent/JPH02288975A/en
Pending legal-status Critical Current

Links

Landscapes

  • Complex Calculations (AREA)

Abstract

PURPOSE:To decrease the number of operations and to shorten a whole operation time by executing no unnecessary operations when a mask bit shows '1.' CONSTITUTION:When the mask bit read from a memory 7 by an address, for which '1' is added to the output of an address register 5, is '1,' '1' and '1' are added to the output of the address register 5 by an adder 8 and stored to the address register 5. Samely, the data of vector registers 1 and 2 are read by the address of the address register 5 and calculated by an arithmetic part 3. At such a time, the output of the address register 5 is increased by '2.' Namely, at such a time, when the address is stored to a vector register 4, data to be abandoned when the mask bit of the memory 7 is referred are stored. Accordingly, the arithmetic is not executed but operation is skipped. Thus, when the mask bit is '1,' the number of the unnecessary operations is decreased and a whole processing time is shortened.

Description

【発明の詳細な説明】 技術分野 本発明はベクトル演算装置に関し、特に演算結果を記憶
手段に格納する場合に格納指示手段の指示内容に従って
格納するようにしたベタ1〜ル演算装置に関する。
TECHNICAL FIELD The present invention relates to a vector arithmetic device, and more particularly to a vector arithmetic device in which arithmetic results are stored in storage means in accordance with instructions from storage instructing means.

従来技術 一般に、ベクトル演算装置は1命令で複数の演算を処理
するベタ1〜ル命令をもつ「1算機に使われ、1ステツ
プの長さが短く1ステツプ内に処理できる内容が少ない
。そのためステップ数を増やしたり処理を単純化したも
のが多い。
PRIOR ART In general, vector arithmetic units are used in arithmetic machines that have basic instructions that process multiple operations with one instruction, and the length of one step is short, so there is little content that can be processed within one step. Many of them increase the number of steps or simplify processing.

従来、この種のベクトル演算装置は第2図に示すような
構成になっている。演算データを格納するベクトルレジ
スタ1.2と、演算器3と、演算結果を格納するベタ1
〜ルレジスタ4と、ベタ1−ルレジスタ12の読出しア
ドレスおよびベクトルレジスタ4の書込みアドレスを格
納するアドレスレジスタ5と、アドレスを作成するカウ
ンタ6と、ベタ1〜ルレジスタ4に書込むか否かを示す
マスクピッ1〜を格納するメモリ7とから構成される。
Conventionally, this type of vector calculation device has a configuration as shown in FIG. A vector register 1.2 that stores calculation data, a calculation unit 3, and a vector register 1 that stores calculation results.
an address register 5 that stores the read address of the flat register 12 and the write address of the vector register 4, a counter 6 that creates an address, and a mask pick that indicates whether or not to write to the flat register 4. The memory 7 stores 1 to 1.

ベクトルレジスタ1,2に格納されたデータは、アドレ
スレジスタ5で示されるアドレスにしたかって読出され
、演算器3によって演算されてその結果かベクトルレジ
スタ4の7゛ドレスレジスタ5で示されたアドレスへ格
納される。この場合、メモリ7のアドレスレジスタ5に
示されたアドレスで読出された値が1″であった場合、
マスクがかかっているのてベクトルレジスタ4への格納
を中止し演算結果は捨てられる。
The data stored in vector registers 1 and 2 is read out according to the address indicated by address register 5, and is computed by arithmetic unit 3, and the result is sent to the address indicated by address register 5 of vector register 4. Stored. In this case, if the value read at the address indicated in the address register 5 of the memory 7 is 1'',
Since it is masked, storage into the vector register 4 is stopped and the operation result is discarded.

なお、ベクトルレジスタとは1つのベクl〜ル命令によ
りレジスタ内の全データを演算できる構成にしたレジス
タである。
Note that a vector register is a register configured such that all data in the register can be operated on by one vector instruction.

」一連した従来のベタ1ヘル演算装置は、マスクビット
が“1″でマスクがかかっており、演算結果が捨てられ
る場合があっても、全ての演算を行うので、全体の演算
時間が長くなる、という欠点がある。
A series of conventional solid-state arithmetic units are masked with a mask bit of "1", and even if the result of an operation is discarded, all operations are performed, resulting in a longer overall operation time. , there is a drawback.

発明の目的 本発明は以−Hの欠点を解決することを課題とし、マス
クのかかっているアドレスを含む場合は少なくともマス
クのかかっているアドレスの演算を1回以干、はとはし
て無用な演算時間を減らして、全体の演算時間を短縮す
ることのできるベクI〜ル演算装置を提供することを目
的とする。
Purpose of the Invention It is an object of the present invention to solve the following drawbacks: When a masked address is included, the masked address must be operated at least once, which is completely unnecessary. An object of the present invention is to provide a vector calculation device capable of reducing the calculation time and the overall calculation time.

発明の構成 本発明によるベタ1〜ル演算装道“は、複数のデータを
格納するベクI−ル記憶f段と、この記憶手段からの読
出しデータを用いて演算を行う演算手段と、この演算結
果を格納する演算結果記憶手段と、この演算結果記憶手
段に前記演算結果を書込むか否かを指示する指示手段と
を含むベタ1〜ル演算装置であって、前記ベタ1ヘル記
憶手段の読出し時にこの読出しアドレスに対応する前記
指示手段の内容を読出すと同時に次のアドレスに対応す
る内容をも読出ず手段と、この読出された前記衣のアド
レスに対応する内容に応じて前記続出しアドレスを生成
する手段とを含むことを特徴とするベクl〜ル演算装置
か得られる。
Structure of the Invention The vector I-L calculation device according to the present invention includes a vector storage f stage for storing a plurality of data, a calculation means for performing calculations using data read from the storage means, and a calculation means for performing calculations using data read from the storage means. A solid 1 to 3 calculation device comprising a calculation result storage means for storing a result, and an instruction means for instructing whether or not to write the calculation result in the calculation result storage means, At the time of reading, the content of the instruction means corresponding to this readout address is read out, and at the same time, the content corresponding to the next address is also read out. There is obtained a vector arithmetic device characterized in that it includes means for generating an address.

実施例 以下、本発明の詳細をその実施例につき図面を参照して
説明する。
EXAMPLES Hereinafter, the details of the present invention will be explained by way of examples with reference to the drawings.

第1図は本発明の一実施例の11772図である。FIG. 1 is a diagram 11772 of one embodiment of the present invention.

ベクトルレジスター、2は演算データを複数個格納して
いる6演算器3はベタ1〜ルレジスター、2のデータを
使って演算を行い、演算結果をベクトルレジスタ4に格
納する。指示手段としてのメモリ7は2つの7′ドレス
により各々同時に異なるデータを読出ずことのできるレ
ジスタファイルであり、ベタ1〜ルレジスタ4に演算結
果を書込むか否かを示すマスクピッ1−が格納されてい
るアドレス作成手段としての加算器8およびアドレスレ
ジスタ5があり、加算器8はアドレスレジスタ5の出力
と、メモリ7からアドレスレジスタ5に示されるアドレ
スにカウンタ9に1′”を加えたアドレスで読出された
値と、” 1 ”とを加算する加算器である。アドレス
レジスタ5はペターへルレシスター、2の読出しアドレ
スおよびヘクl−ルレジスター1のぜ)込みアドレスを
格納するレジスタである。
A vector register 2 stores a plurality of pieces of operation data.The arithmetic unit 3 performs an operation using the data in the registers 1 to 2, and stores the result of the operation in a vector register 4. The memory 7 serving as an instruction means is a register file in which different data can be read out at the same time by two 7' addresses, and mask picks 1-, which indicate whether or not to write operation results in the registers 1-4, are stored. There is an adder 8 and an address register 5 as means for creating an address. This is an adder that adds "1" to the read value.Address register 5 is a register that stores the read address of register register 2 and the insertion address of register register 1.

次に、以上の構成をもつ本実施例の動作について説明す
る。ベクトルレジスター、2に格納されているデータを
アドレスレジスタ5で示されるアドレスて続出し、演算
83で演算を行う。ここで、メモリ7に格納されている
マスクピッl−のうち、アドレスレジスタ5て示されて
いるアドレスて読出したマスクビットが0″であれば、
ベクトルレジスタ4に演算器3の結果をアドレスレジス
タ5で示されるアドレスに書込み、同時に、メモリ7か
らアドレスレジスタ5の出力に1′″を加えたアドレス
で読出したマスクピッ1〜が′0″てあれば、加算器8
により、アドレスレジスタ5の出力に1″と” o ”
を加算し、アドレスレジスタ5へ格納する。
Next, the operation of this embodiment having the above configuration will be explained. The data stored in the vector register 2 is sequentially outputted to the address indicated by the address register 5, and an operation is performed in operation 83. Here, if the mask bit read out at the address indicated by the address register 5 among the mask bits stored in the memory 7 is 0'',
Write the result of the arithmetic unit 3 to the vector register 4 at the address indicated by the address register 5, and at the same time, write the mask bits 1 to 1 read from the memory 7 at the address obtained by adding 1''' to the output of the address register 5 to be '0''. For example, adder 8
As a result, the output of address register 5 is 1″ and “o”.
are added and stored in address register 5.

次のスデップで、同様にしてベクトルレジスタ1.2の
データをアドレスレジスタ5のアドレスで読出して演算
器3で演算する。ここでメモリ7に格納されているマス
クピッ1〜のうちアドレスレジスタ5て示されているア
ドレスで読出したマスクピッ1〜が“0′″であれは、
ベタ1〜ルレジスタ4に演算器3の結果をアドレスレジ
スタ5て示さねるアドレスに書込み、同時に、メモリ7
からアドレスレジスタ5の出力に” ] ”を加えたア
ドレスで読出したマスクビットが′1′″であれは、加
算器8によりアドレスレジスタ5の出力に1′″と“1
′°を加算し、アドレスレジスタ5へ格納する。
In the next step, the data in the vector registers 1.2 is similarly read out using the address in the address register 5, and the arithmetic unit 3 calculates it. Here, if the mask bit 1~ read out at the address indicated by the address register 5 among the mask bits 1~ stored in the memory 7 is "0'", then
The results of the arithmetic unit 3 are written to the address registers 5 and 4, and at the same time, the memory 7
If the mask bit read at the address obtained by adding " ] " to the output of the address register 5 is '1', the adder 8 adds 1' and '1' to the output of the address register 5.
'° is added and stored in address register 5.

次のステップで、同様にしてベクトルレジスタ1.2の
データをアドレスレジスタ5のアドレスで読出して演算
器3で演算する。この時アドレスレジスタ5の出力は前
ステップの時より“2″増えている。つまりこのときの
アドレスはベクトルレジスタ4に格納する際、メモリ7
のマスクビットを参照した時に捨てられるデータが格納
されているので、演算を行わずにとばしたのである。従
って、このようにしたことによりマスクビットか“1″
の場合の無用な演算回数を減らし全体の処理時間を短く
することができる。
In the next step, the data in the vector registers 1.2 is similarly read out using the address in the address register 5, and the arithmetic unit 3 calculates it. At this time, the output of the address register 5 has increased by "2" from the previous step. In other words, when storing the address in the vector register 4, the address at this time is
Since the data that would be discarded when the mask bit of is referenced is stored, it was skipped without performing any calculations. Therefore, by doing this, the mask bit becomes "1".
In this case, the number of unnecessary operations can be reduced and the overall processing time can be shortened.

発明の効果 蒸上の如く、本発明によれは、マスクビットか“′1′
”を示す場合には無用な演算を行わないようにしている
ので、演算回数がそれたけ減少して、全体の演算時間を
短縮できるという効果かある。
Effects of the Invention As described above, according to the present invention, the mask bit or "'1"
”, unnecessary calculations are not performed, so the number of calculations is reduced by that amount, which has the effect of shortening the overall calculation time.

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

第1図は本発明の実施例のブロック図、第2図は従来の
ベクトル演算装置のブロック図である。 主要部分の符号の説明 1.2.4・・・・・・ベクトルレジスタ3・・・・・
・演算器 5・・・・・・アドレスレジスタ 7・・・・・・マスクピッ1〜メモリ 8・・・・・・加算器
FIG. 1 is a block diagram of an embodiment of the present invention, and FIG. 2 is a block diagram of a conventional vector calculation device. Explanation of symbols of main parts 1.2.4...Vector register 3...
- Arithmetic unit 5...Address register 7...Mask pin 1 to memory 8...Adder

Claims (1)

【特許請求の範囲】[Claims] (1)複数のデータを格納するベクトル記憶手段と、こ
の記憶手段からの読出しデータを用いて演算を行う演算
手段と、この演算結果を格納する演算結果記憶手段と、
この演算結果記憶手段に前記演算結果を書込むか否かを
指示する指示手段とを含むベクトル演算装置であって、
前記ベクトル記憶手段の読出し時にこの読出しアドレス
に対応する前記指示手段の内容を読出すと同時に次のア
ドレスに対応する内容をも読出す手段と、この読出され
た前記次のアドレスに対応する内容に応じて前記読出し
アドレスを生成する手段とを含むことを特徴とするベク
トル演算装置。
(1) vector storage means for storing a plurality of data; calculation means for performing calculations using data read from the storage means; calculation result storage means for storing the calculation results;
A vector arithmetic device comprising: instruction means for instructing whether or not to write the arithmetic result in the arithmetic result storage means,
means for reading the contents of the instruction means corresponding to the read address when reading the vector storage means, and at the same time reading the contents corresponding to the next address; A vector arithmetic device comprising means for generating the read address accordingly.
JP3801889A 1989-02-17 1989-02-17 Vector arithmetic unit Pending JPH02288975A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP3801889A JPH02288975A (en) 1989-02-17 1989-02-17 Vector arithmetic unit

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP3801889A JPH02288975A (en) 1989-02-17 1989-02-17 Vector arithmetic unit

Publications (1)

Publication Number Publication Date
JPH02288975A true JPH02288975A (en) 1990-11-28

Family

ID=12513835

Family Applications (1)

Application Number Title Priority Date Filing Date
JP3801889A Pending JPH02288975A (en) 1989-02-17 1989-02-17 Vector arithmetic unit

Country Status (1)

Country Link
JP (1) JPH02288975A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0877143A (en) * 1994-09-02 1996-03-22 Kofu Nippon Denki Kk Vector data processor

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0877143A (en) * 1994-09-02 1996-03-22 Kofu Nippon Denki Kk Vector data processor

Similar Documents

Publication Publication Date Title
JPH0319986B2 (en)
US5142630A (en) System for calculating branch destination address based upon address mode bit in operand before executing an instruction which changes the address mode and branching
JPS6041769B2 (en) Addressing method
JPH02288975A (en) Vector arithmetic unit
JPS61110240A (en) Optimization compiler
JPS5890244A (en) Data processor
JPS60129853A (en) Address generating device
JPH06124207A (en) Sub-routine branching instruction execution method
KR930003410B1 (en) Data processing apparatus having high-sped reference of stack data
JPS59201144A (en) Decimal arithmetic circuit
JPS622332B2 (en)
JPS59111533A (en) Digital data arithmetic circuit
JP3115428B2 (en) Pipeline control method
JP2842930B2 (en) Instruction readout circuit used in test processor of semiconductor integrated circuit test equipment
JP2926951B2 (en) Save / restore register address generation circuit
JP3523407B2 (en) Information processing equipment
JP3124361B2 (en) Memory data loading device
JP2989830B2 (en) Vector processing method
JP3705263B2 (en) Sequencer
JPS62295138A (en) Address forming system
JPS63276126A (en) Instruction decoding circuit
JPH05250156A (en) Risc processor
JP2005107926A (en) Vector processor, digital data processing method and digital data processing program
JP2002304293A (en) Processor device, method of updating address, and method of updating repeat counter
JPS59188900A (en) Data processor