JPH03257621A - Floating point multiplying circuit - Google Patents

Floating point multiplying circuit

Info

Publication number
JPH03257621A
JPH03257621A JP5708490A JP5708490A JPH03257621A JP H03257621 A JPH03257621 A JP H03257621A JP 5708490 A JP5708490 A JP 5708490A JP 5708490 A JP5708490 A JP 5708490A JP H03257621 A JPH03257621 A JP H03257621A
Authority
JP
Japan
Prior art keywords
digit
multiplication
circuit
normalization
amount
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
JP5708490A
Other languages
Japanese (ja)
Inventor
Takeshi Amamiya
雨宮 毅
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 Computertechno Ltd
Original Assignee
NEC Computertechno 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 Computertechno Ltd filed Critical NEC Computertechno Ltd
Priority to JP5708490A priority Critical patent/JPH03257621A/en
Publication of JPH03257621A publication Critical patent/JPH03257621A/en
Pending legal-status Critical Current

Links

Abstract

PURPOSE:To improve the throughput of multiplication by selecting any one of the data of a multiplied result, data shifting one digit to left in the matissa part of the multiplied result and the output data of an arithmetic result register storing data, for which the multiplied result is normalized by a normalizing shifter, according to a normalizing amount deciding signal. CONSTITUTION:Based on floating point data 10 and 20 to be inputted to a multiplier 1, a normalizing amount deciding circuit 6 calculates the normalizing amount of a multiplied result 100 from the multiplier 1. The circuit 6 outputs 00 and a circuit 8 selects the output of an intermediate result register 2. When the normalizing amount is one digit on the left side, the circuit 6 outputs 01 and the circuit 8 applies the mantissa part of an output 200 from the register 2 to a left-side one digit shifter 5 and normalizes the mantissa part by shifting one digit to the left side. Then, normalized data 500 is selected and outputted as an output signal 800. When the normalizing amount is neither 0 digit nor 1 digit, the circuit 6 outputs 1X and the circuit 8 selects normalized data 400 from an arithmetic result register 4. Thus, the through-put of a floating point multiplying instruction can be improved.

Description

【発明の詳細な説明】 技術分野 本発明は浮動小数点乗算回路に関し、特に情報処理装置
に用いられる浮動小数点乗算回路に関するものである。
DETAILED DESCRIPTION OF THE INVENTION Technical Field The present invention relates to a floating point multiplication circuit, and more particularly to a floating point multiplication circuit used in an information processing device.

従来技術 従来、この種の浮動小数点乗算回路は、仮数部の乗算を
行なう乗算器、この乗算器から出力される仮数部の乗算
結果を格納する中間結果レジスタ、およびこの中間結果
レジスタの正規化されていないデータを正規化する正規
化シフタを含んで構成されている。
Prior Art Conventionally, this type of floating-point multiplication circuit consists of a multiplier that multiplies the mantissa, an intermediate result register that stores the multiplication result of the mantissa output from this multiplier, and a normalized state of this intermediate result register. It consists of a normalization shifter that normalizes data that is not normalized.

このような乗算回路では、入力データのほとんどが正規
化されたデータであるので、中間結果レジスタに格納さ
れた正規化処理前の乗算結果の正規化量はほとんどの場
合“O゛桁か左1桁である。
In such a multiplication circuit, most of the input data is normalized data, so in most cases the normalized amount of the multiplication result before normalization stored in the intermediate result register is "0" digit or 1 left It is a digit.

しかし、中間結果レジスタに格納された乗算結果は正規
化シフタを用いてまず正規化処理され、演算結果レジス
タに一担格納される。その後で浮動小数点乗算回路の外
部レジスタに転送される構成となっている。このため、
正規化量が少なくても正規化量に関係なく正規化処理の
ために1クロツクサイクルかかる。この結果、処理性能
が向上しないという欠点かある。
However, the multiplication result stored in the intermediate result register is first normalized using a normalization shifter, and then stored in the operation result register. The configuration is such that the data is then transferred to an external register of the floating-point multiplication circuit. For this reason,
Even if the amount of normalization is small, it takes one clock cycle for the normalization process regardless of the amount of normalization. As a result, there is a drawback that processing performance is not improved.

発明の目的 本発明の目的は乗算処理性能の向上を図った浮動小数点
乗算回路を提供することである。
OBJECTS OF THE INVENTION An object of the present invention is to provide a floating point multiplication circuit with improved multiplication processing performance.

発明の構成 本発明による浮動小数点乗算回路は、浮動小数点データ
の乗算を行う乗算手段と、この乗算手段からの乗算結果
を正規化する正規化シフタ手段と、前記乗算手段からの
乗算結果の仮数部を1桁シフトさせる1桁シフタ手段と
、前記乗算手段に入力される浮動小数点データの仮数部
から乗算結果の正規化量を求め、0桁か、左1桁かまた
はそれ以外かの正規化量判定信号を出力する正規化量判
定手段と、この正規化量判定手段からの正規化量判定信
号が0桁の正規化量を示しているとき前記乗算手段から
の乗算結果を選択し、該正規化量判定信号が左1桁の正
規化量を示しているとき前記1桁シフタ手段の出力を選
択し、該正規化量判定信号が0桁および左1桁のいずれ
でもない正規化量を示しているとき前記正規化シフタ手
段で正規化されたデータを選択する選択手段とを含むこ
とを特徴とする。
Structure of the Invention A floating point multiplication circuit according to the present invention comprises a multiplication means for multiplying floating point data, a normalization shifter means for normalizing the multiplication result from the multiplication means, and a mantissa part of the multiplication result from the multiplication means. 1-digit shifter means for shifting by one digit, and a normalized amount of the multiplication result from the mantissa part of the floating-point data input to the multiplication means, and a normalized amount of 0 digit, left 1 digit, or other. A normalization amount determination means for outputting a determination signal; and when the normalization amount determination signal from the normalization amount determination means indicates a 0-digit normalization amount, the multiplication result from the multiplication means is selected, and the multiplication result from the multiplication means is selected; selecting the output of the one-digit shifter means when the normalization amount determination signal indicates the normalization amount of the first digit on the left; and selecting means for selecting the data normalized by the normalization shifter means when the data is normalized by the normalization shifter means.

実施例 次に、本発明の一実施例について図面を参照して説明す
る。
Embodiment Next, an embodiment of the present invention will be described with reference to the drawings.

第1図を参照すると、本発明の一実施例は、浮動小数点
データ10および20の乗算を実行する乗算器1と、こ
の乗算器1の出力データである乗算結果100を入力す
る中間結果レジスタ2と、中間結果レジスタ2の出力デ
ータ200の正規化処理を行う正規化シフタ3と、正規
化シフタ3において正規化処理されたデータ300を格
納する演算結果レジスタ4と、中間結果レジスタ2の出
力データ200の仮数部を左1桁シフタさせる左1桁シ
フタ5と、浮動小数点データ10および20の仮数部を
入力し、乗算器1から出力される乗算結果100の正規
化量が0桁なのか、左1桁なのか、またはそれ以外なの
かの正規化量判定信号600を出力する正規化量判定回
路6と、この回路6からの正規化量判定信号600を格
納し、乗算命令の終了まてこの信号600を保持するフ
リップフロップ(FF)7と、演算結果レジスタ4の出
力データ400、左1桁シフタ5の出力データ500お
よび中間結果レジスタ2の出力データ200のうちいず
れかのデータをFF7の出力信号700の制御により選
択し、その出力データ800に演算結果を出力する演算
結果選択回路8と、FF7の出力信号700を分岐アド
レス生成のための1つの要因として人力し、この実施例
である高速浮動小数点乗算回路の制御を行う演算制御回
路9とを含む。
Referring to FIG. 1, one embodiment of the present invention includes a multiplier 1 that performs multiplication of floating point data 10 and 20, and an intermediate result register 2 that receives the multiplication result 100 that is the output data of this multiplier 1. , a normalization shifter 3 that normalizes the output data 200 of the intermediate result register 2, an operation result register 4 that stores the data 300 normalized in the normalization shifter 3, and the output data of the intermediate result register 2. Is the normalized amount of the multiplication result 100 output from the multiplier 1 by inputting the left 1 digit shifter 5 which shifts the mantissa part of 200 by 1 digit to the left and the mantissa parts of floating point data 10 and 20? A normalization amount judgment circuit 6 that outputs a normalization amount judgment signal 600 indicating whether it is the left 1 digit or something else, and a normalization amount judgment signal 600 from this circuit 6 is stored, and the normalization amount judgment signal 600 is stored until the multiplication instruction ends. Flip-flop (FF) 7 that holds this signal 600, output data 400 of operation result register 4, output data 500 of left 1-digit shifter 5, and output data 200 of intermediate result register 2 are transferred to FF 7. In this embodiment, the calculation result selection circuit 8 selects by controlling the output signal 700 and outputs the calculation result as the output data 800, and the output signal 700 of the FF 7 is manually used as one factor for branch address generation. and an arithmetic control circuit 9 that controls a high-speed floating-point multiplication circuit.

次に本発明の一実施例の動作について図面を参照して詳
細に説明する。
Next, the operation of one embodiment of the present invention will be described in detail with reference to the drawings.

第1図を参照すると、正規化量判定回路6は、乗算器1
に入力すべき浮動小数点データ10および20に基づい
て乗算器1から乗算結果100の正規化量を求める。
Referring to FIG. 1, the normalization amount determination circuit 6 includes a multiplier 1
The normalized amount of the multiplication result 100 from the multiplier 1 is determined based on the floating point data 10 and 20 to be input to the multiplier 1.

正規化量が0桁であると判定されたとき、回路6は、“
00”を出力し、FF7に格納したあと演算結果選択回
路8に選択信号として与える。この選択信号により回路
8は中間結果レジスタ2の出力を選択する。
When it is determined that the normalized amount is 0 digits, the circuit 6 performs “
00'' is outputted, stored in the FF 7, and then given as a selection signal to the operation result selection circuit 8. The selection signal causes the circuit 8 to select the output of the intermediate result register 2.

一方、演算制御回路9は、「乗算ステップ」、「この乗
算回路の外部レジスタへの格納ステップ」の2ステツプ
により乗算命令を実行する。正規化量が左1桁であると
回路6で判定されたときも同じような動作となる。
On the other hand, the arithmetic control circuit 9 executes the multiplication instruction in two steps: a "multiplication step" and a "storage step to an external register of this multiplication circuit." A similar operation occurs when the circuit 6 determines that the normalized amount is one digit on the left.

正規化量か左1桁であると判定されたとき、回路6は“
01°を出力する、乗算ステップでFF7に“01”を
出力するとともに乗算器1からの正規化量産1桁の乗算
結果が中間結果レジスタ2に格納される。さらに、FF
7は出力信号700として“01”を出力するので演算
制御回路9は正規化のステップを省略する。
When it is determined that the normalized amount is the left 1 digit, the circuit 6 outputs “
In the multiplication step in which 01° is output, “01” is output to the FF 7, and the normalized mass-produced one-digit multiplication result from the multiplier 1 is stored in the intermediate result register 2. Furthermore, F.F.
7 outputs "01" as the output signal 700, so the arithmetic control circuit 9 omits the normalization step.

この回路9の格納ステップでは、FF7の出力信号7C
・0としての“01°に応答して回路8は、レジスタ2
の出力200の仮数部を左1桁シフタ5に与える。左1
桁シフタ5は仮数部を左1桁シフトすることにより正規
化する。選択回路8は正規化されたデータ500を選択
し出力信号800として出力する。
In the storing step of this circuit 9, the output signal 7C of the FF7
- In response to “01° as 0, circuit 8 sets register 2
The mantissa part of the output 200 is given to the left 1-digit shifter 5. left 1
The digit shifter 5 normalizes the mantissa by shifting it one digit to the left. The selection circuit 8 selects the normalized data 500 and outputs it as an output signal 800.

正規化量が0桁でも1桁でもないとき、回路6は“1X
′を正規化量判定回路600として出力する。X”はこ
のビットが任意であることを示す。
When the normalized amount is neither 0 digit nor 1 digit, the circuit 6
' is output as the normalization amount determination circuit 600. X” indicates that this bit is optional.

信号600の“IX”に応答して、演算結果選択回路8
は演算結果レジスタ4からの正規化されたデータ400
を選択する。すなわち、FF7の出力信号700か“I
X”のとき、「乗算ステップ」、「正規化ステップ」、
「この乗算回路の外部レジスタへの格納ステップ」の3
ステツプにより乗算命令を実行する。
In response to “IX” of the signal 600, the calculation result selection circuit 8
is the normalized data 400 from operation result register 4
Select. That is, the output signal 700 of FF7 or “I
X”, “multiplication step”, “normalization step”,
3 of “Step of storing this multiplication circuit in an external register”
Executes a multiplication instruction by stepping.

発明の詳細 な説明したように本発明は、乗算結果の正規化量が0桁
か、左1桁か、またはそれ以外かの正規化量判定信号を
出力する正規化量判定回路と、正規化量判定信号により
、乗算結果のデータと、乗算結果の仮数部を左1桁シフ
トしたデータと、乗算結果を正規化シフタで正規化処理
したデータを格納した演算結果レジスタの出力データと
のうちいずれかを選択する選択回路とを備えることによ
り、乗算結果の正規化量か0桁か左1桁のとき正規化処
理の動作と本乗算回路外部レジスタへの格納動作が1ス
テツプで実行でき、浮動小数点乗算命令の処理性能か上
かるという効果かある。
DETAILED DESCRIPTION OF THE INVENTION As described above, the present invention provides a normalization amount determination circuit that outputs a normalization amount determination signal indicating whether the normalization amount of a multiplication result is 0 digit, left 1 digit, or something else; Depending on the amount judgment signal, which of the data of the multiplication result, the data obtained by shifting the mantissa part of the multiplication result by one digit to the left, and the output data of the operation result register storing the data obtained by normalizing the multiplication result by the normalization shifter is selected. By providing a selection circuit that selects whether the normalized amount of the multiplication result is 0 digit or 1 digit on the left, the normalization processing operation and the storage operation to the external register of this multiplication circuit can be performed in one step, and the floating This has the effect of improving the processing performance of decimal point multiplication instructions.

【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図である。 主要部分の符号の説明 1・・・・・・乗算器  2・・・・・・中間結果レジ
スタ3・・・・・・正規化シフタ 4・・・・・・演算結果レジスタ 5・・・・・・左1桁シフタ 6・・・・・・正規化量判定回路 7・・・・・・フリップフロップ(F F)8・・・・
・・演算結果選択回路 9・・・・・・演算制御回路
FIG. 1 is a diagram showing an embodiment of the present invention. Explanation of symbols of main parts 1... Multiplier 2... Intermediate result register 3... Normalization shifter 4... Arithmetic result register 5... ...Left 1-digit shifter 6...Normalization amount judgment circuit 7...Flip-flop (F F) 8...
...Arithmetic result selection circuit 9...Arithmetic control circuit

Claims (1)

【特許請求の範囲】[Claims] (1)浮動小数点データの乗算を行う乗算手段と、この
乗算手段からの乗算結果を正規化する正規化シフタ手段
と、前記乗算手段からの乗算結果の仮数部を1桁シフト
させる1桁シフタ手段と、前記乗算手段に入力される浮
動小数点データの仮数部から乗算結果の正規化量を求め
、0桁か、左1桁かまたはそれ以外かの正規化量判定信
号を出力する正規化量判定手段と、この正規化量判定手
段からの正規化量判定信号が0桁の正規化量を示してい
るとき前記乗算手段からの乗算結果を選択し、該正規化
量判定信号が左1桁の正規化量を示しているとき前記1
桁シフタ手段の出力を選択し、該正規化量判定信号が0
桁および左1桁のいずれでもない正規化量を示している
とき前記正規化シフタ手段で正規化されたデータを選択
する選択手段とを含むことを特徴とする浮動小数点乗算
回路。
(1) Multiplication means for multiplying floating point data, normalization shifter means for normalizing the multiplication result from the multiplication means, and one-digit shifter means for shifting the mantissa part of the multiplication result from the multiplication means by one digit. and a normalization amount determination that calculates the normalization amount of the multiplication result from the mantissa part of the floating point data input to the multiplication means and outputs a normalization amount determination signal as to whether it is 0 digit, left 1 digit, or something else. and when the normalization amount judgment signal from the normalization amount judgment means indicates a normalized amount of 0 digit, selects the multiplication result from the multiplication means, and selects the multiplication result from the multiplication means, When indicating the normalized amount, the above 1
The output of the digit shifter means is selected, and the normalization amount determination signal is 0.
and selecting means for selecting data normalized by the normalization shifter means when the normalized amount indicates a normalized amount that is neither a digit nor a left digit.
JP5708490A 1990-03-08 1990-03-08 Floating point multiplying circuit Pending JPH03257621A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP5708490A JPH03257621A (en) 1990-03-08 1990-03-08 Floating point multiplying circuit

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP5708490A JPH03257621A (en) 1990-03-08 1990-03-08 Floating point multiplying circuit

Publications (1)

Publication Number Publication Date
JPH03257621A true JPH03257621A (en) 1991-11-18

Family

ID=13045625

Family Applications (1)

Application Number Title Priority Date Filing Date
JP5708490A Pending JPH03257621A (en) 1990-03-08 1990-03-08 Floating point multiplying circuit

Country Status (1)

Country Link
JP (1) JPH03257621A (en)

Similar Documents

Publication Publication Date Title
JPH0792739B2 (en) Floating point data normalization method
US10095475B2 (en) Decimal and binary floating point rounding
US9069686B2 (en) Digital signal processor having instruction set with one or more non-linear functions using reduced look-up table with exponentially varying step-size
JPH05250146A (en) Arithmetic operation circuit executing integer involution processing
US20100138465A1 (en) Digital Signal Processor With One Or More Non-Linear Functions Using Factorized Polynomial Interpolation
GB2549153B (en) Apparatus and method for supporting a conversion instruction
JPH10187416A (en) Floating point arithmetic unit
JPH03257621A (en) Floating point multiplying circuit
JPH10333883A (en) Multiplying method and circuit therefor
JP2575969B2 (en) Floating point multiplier / divider
JP3110072B2 (en) Pre-normalization circuit
JP2664750B2 (en) Arithmetic device and arithmetic processing method
JP3122622B2 (en) Division device
GB2581507A (en) Logarithmic number system
JPS59136849A (en) Dividing device
JP2615746B2 (en) Bit operation circuit
JPS6120132A (en) Processing unit
JPH02310618A (en) Information processor
JPH0553759A (en) Fixed decimal point arithmetic unit
JPS63279321A (en) Microprogram controller
JPH1185471A (en) Operation method and arithmetic unit
JPH01304535A (en) Division device
JPH0317738A (en) Arithmetic processor
JPH04117520A (en) Floating point computing device
JPS5972541A (en) Data processor