JPS61182134A - Adding device - Google Patents

Adding device

Info

Publication number
JPS61182134A
JPS61182134A JP60021564A JP2156485A JPS61182134A JP S61182134 A JPS61182134 A JP S61182134A JP 60021564 A JP60021564 A JP 60021564A JP 2156485 A JP2156485 A JP 2156485A JP S61182134 A JPS61182134 A JP S61182134A
Authority
JP
Japan
Prior art keywords
exponent
mantissa
data format
addition
output
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
JP60021564A
Other languages
Japanese (ja)
Inventor
Takashi Kanazawa
金澤 敬
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 Corp
Original Assignee
NEC Corp
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 Corp filed Critical NEC Corp
Priority to JP60021564A priority Critical patent/JPS61182134A/en
Publication of JPS61182134A publication Critical patent/JPS61182134A/en
Pending legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F7/00Methods or arrangements for processing data by operating upon the order or content of the data handled
    • G06F7/38Methods or arrangements for performing computations using exclusively denominational number representation, e.g. using binary, ternary, decimal representation
    • G06F7/48Methods or arrangements for performing computations using exclusively denominational number representation, e.g. using binary, ternary, decimal representation using non-contact-making devices, e.g. tube, solid state device; using unspecified devices
    • G06F7/483Computations with numbers represented by a non-linear combination of denominational numbers, e.g. rational numbers, logarithmic number system or floating-point numbers
    • G06F7/485Adding; Subtracting
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F7/00Methods or arrangements for processing data by operating upon the order or content of the data handled
    • G06F7/38Methods or arrangements for performing computations using exclusively denominational number representation, e.g. using binary, ternary, decimal representation
    • G06F7/48Methods or arrangements for performing computations using exclusively denominational number representation, e.g. using binary, ternary, decimal representation using non-contact-making devices, e.g. tube, solid state device; using unspecified devices
    • G06F7/499Denomination or exception handling, e.g. rounding or overflow
    • G06F7/49936Normalisation mentioned as feature only

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computational Mathematics (AREA)
  • Computing Systems (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Nonlinear Science (AREA)
  • General Engineering & Computer Science (AREA)

Abstract

PURPOSE:To perform both adding operations of floating points and fixed points by means of the same hardware by providing a data type indicating means, an adding means for exponent part, a selecting means for exponent part, a digit matching means, an adding means for mantissa part, a normalizing means, a code processing means and a corresponding means to an adding device. CONSTITUTION:The data operands are set to input registers 10 and 20. Then 0 or 1 is set to a data type indicating means 30 according to a floating or fixed point input data type. In the case of the floating point, an exponent difference is obtained by an adding means for exponent part. A digit matching means 60 obtains a mantissa having the smaller value according to the value of an exponent or a mantissa shifted by an exponent difference as the addend/augend of the mantissa part. An exponent switching means selects a larger exponent. An adding means 70 for mantissa part adds the mantissa addends and augends. A normalizing means 80 shifts the result of addition toward the highest bit. A correcting means 100 adds the overflow given from the means 70 to an exponent part switching means 50. Then the output of the means 80 is reduced.

Description

【発明の詳細な説明】 産業上の利用分野 本発明は、加算装置に関し、%に1浮動小数点データ形
式のオペランドと固定小数点データ形式のオペランドに
対し、それぞれ浮動小数点加算、固定小数点加算を行な
う加算装置に関するものである。
DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention relates to an addition device that performs floating-point addition and fixed-point addition on operands in a floating-point data format and operands in a fixed-point data format, respectively. It is related to the device.

従来の技術 従来、浮動小数点データ形式のオペランドと固定小数点
データ形式のオペランドに対してそれぞれ浮動小数点加
算、固定小数点加算を実行する場合には、第4図に示す
ような、第1オペランドを4納する入力レジスタ605
、第2オペランドを格納する入力レジスタ610、第1
及び第2オペランドの符号から代数的に求まる符号を生
成する符号生成手段620、第1及び第2オペランドの
指数部の大小関係により第1″またけ第2オペランドの
指数部を選択する指数部選択手段6301第1及び第2
オペランドの指数差を生成する指数差生成手段640、
指数差生成手段640で生成された指数差に応じて第1
及び第2オペランドの仮数部の桁合せを行なう桁合せ手
段650、桁合せ手段650により桁合せされた第1及
び第2オペランドの仮数部を加算する加算手段660、
加算手段660の加算結果を正規化する正規化手段67
01指数部選択手段630から送られてくる指数の補正
を正規化手段670の指示により行なう補正手段680
、結果レジスタ690から構成される浮動小数点加算器
と、第5図に示すような、第1オペランドを格納する入
力レジスタ710、第2オペランドを格納する入力レジ
スタ7201第1オペランドと第2オペランドを加算す
る加算手段730、結果レジスタ740から構成される
固定小数点加算器を、例えば、第3図て示すような構成
で備える必要があった。
2. Description of the Related Art Conventionally, when performing floating-point addition and fixed-point addition on operands in floating-point data format and fixed-point data format, respectively, the first operand is input register 605
, an input register 610 storing the second operand;
and a code generating means 620 that generates a code algebraically determined from the sign of the second operand, and an exponent part selection unit that selects the exponent part of the second operand across the first and second operands based on the magnitude relationship between the exponent parts of the first and second operands. Means 6301 first and second
exponent difference generation means 640 for generating an exponent difference between operands;
The first index difference generated by the index difference generation means 640
and digit alignment means 650 for aligning the mantissa parts of the second operand; addition means 660 for adding the mantissa parts of the first and second operands whose digits have been aligned by the digit alignment means 650;
Normalizing means 67 for normalizing the addition result of the adding means 660
A correction means 680 that corrects the index sent from the 01 exponent part selection means 630 according to instructions from the normalization means 670.
, a floating point adder consisting of a result register 690, an input register 710 that stores the first operand, and an input register 7201 that stores the second operand, as shown in FIG. It was necessary to provide a fixed-point adder consisting of an adding means 730 and a result register 740, for example, as shown in FIG.

発明が解決しようとする問題点 しかしながら、従来の上記構成によれば、金物量が増大
するという欠点があり、また、ベクトル処理装置に組み
込む場合には、通常、浮動小数点加算器と固定小数点加
算器の演算時間が異なるために命令制御が複雑になると
いう欠点があった。
Problems to be Solved by the Invention However, the conventional configuration described above has the disadvantage that the amount of hardware increases, and when incorporated into a vector processing device, it usually requires a floating point adder and a fixed point adder. The disadvantage is that instruction control becomes complicated because the computation times are different.

本発明は従来の技術に内在する上記諸欠点を解消する為
罠なされたものであり、従って本発明の目的は、同一の
金物で浮動小数点及び固定小数点加算が容易にでき、簡
単な命令制御の新規な710算装置を提供することにあ
る。
The present invention has been devised to eliminate the above-mentioned drawbacks inherent in the prior art, and therefore, an object of the present invention is to easily perform floating point and fixed point addition using the same hardware, and to provide simple instruction control. The purpose of this invention is to provide a new 710 calculation device.

問題点を解決するだめの手段 上記目的を達成する為に、本発明の加算装置は、データ
形式の種類を示すデータ形式指示手段と、前記データ形
式指示手段の指示により浮動小数点モードでは第1及び
第2オペランドの指数部の指数差を生成し、固定小数点
モードでは前記第1及び第2オペランドの仮数部の加算
を行なう指数部加算手段と、前記データ形式指示手段の
指示により浮動小数点モードでは前記第1及び第2オペ
ランドの大小関係により定まる前記第1あるいは第2オ
ペランドの指数部を選択し、同定小数点モードでは前記
指数部加算手段の加算結果を選択する指数部選択手段と
、前記データ形式指示手段の指示と前記指数部の大小関
係により浮動小数点モードでのみ前記指数部加算手段に
より生成された指数差に応じて前記第1及び第2オペラ
ンドの仮数部の桁合せを行ない仮数部被加数と仮数部加
数な生成する桁合せ手段と、前記桁合せ手段により生成
された仮数部被加数と仮数部加数を加算する仮数部加算
手段と、前記データ形式指示手段の指示により浮動小数
点モードでのみ前記仮数部加算手段の加算結果を正規化
する正規化手段と、前記データ形式指示手段の指示によ
り浮動小数点モードでは前記第1及び第2オペランドの
符号ピットから代数的に定まる符号を生成し、固定小数
点モードでは第1及び第2オペランドの符号ピットの加
算を行なう符号処理手段と、前記指数部選択手段により
生成された指数と前記符号処理手段により生成された符
号を前記データ形式指示手段と前記正規化手段と前記仮
数部加算手段の指示により補正する補正手段から構成さ
れる。
Means for Solving the Problems In order to achieve the above object, the addition device of the present invention includes a data format indicating means for indicating the type of data format, and a first exponent part adding means for generating an exponent difference between the exponent parts of the second operand and adding the mantissa parts of the first and second operands in the fixed point mode; exponent part selecting means for selecting an exponent part of the first or second operand determined by the magnitude relationship between the first and second operands, and selecting an addition result of the exponent part adding means in the identified decimal point mode; According to the instruction of the means and the magnitude relationship of the exponent part, the mantissa parts of the first and second operands are digit-aligned according to the exponent difference generated by the exponent part adding means only in floating point mode, and the mantissa part summand is digit alignment means for generating a mantissa addend and a mantissa addend; a mantissa addition means for adding the mantissa augend and the mantissa addend generated by the digit alignment means; normalizing means for normalizing the addition result of the mantissa adding means only in the floating point mode; and generating a code algebraically determined from the code pits of the first and second operands in the floating point mode according to instructions from the data format indicating means. In the fixed-point mode, a code processing means adds the code pits of the first and second operands, and the exponent generated by the exponent selection means and the code generated by the code processing means are sent to the data format indicating means. and a correction means for performing correction according to instructions from the normalization means and the mantissa addition means.

発明の実施例 次に本発明をその好ましい一実施例について図面を参照
して詳細に説明する。
Embodiment of the Invention Next, a preferred embodiment of the present invention will be described in detail with reference to the drawings.

第1図は本発明の一実施例を示すブロック構成図である
。第1図において、本発明に係る加算装置の一実施例は
、第1オペランドを格納するレジスタ10、第2オペラ
ンドを格納する入力レジスタ加、データ形式指示手段(
9)、指数部切換手段切、夕11切撲十段(資)、桁会
せ手段60、仮数部加算手段70、正規化手段80、符
号処理手段90.補正手段1oo、g来しジスタ110
から構成される。
FIG. 1 is a block diagram showing one embodiment of the present invention. In FIG. 1, one embodiment of the adding device according to the present invention includes a register 10 for storing a first operand, an input register for storing a second operand, data format indicating means (
9), exponent part switching means off, evening 11 kiributsu 10 dan (shi), digit alignment means 60, mantissa addition means 70, normalization means 80, code processing means 90. Correction means 1oo, g coming register 110
It consists of

指数部加算手段40のa、b、C人力には、それぞれ入
力レジスタ10の一部、入力レジスタ頭の一部、データ
形式指示手段Iの出力が入力される。
A part of the input register 10, a part of the head of the input register, and the output of the data format indicating means I are input to input terminals a, b, and C of the exponent part adding means 40, respectively.

指数部切換手段(資)のa、b、c、C人力にはそれぞ
れ入力レジスタ10の一部、入力レジスタ加の一部、指
数部加算手段のd出力、データ形式指示手段Iの出力が
供給される。桁合せ手段60のa、b。
A part of the input register 10, a part of the input register adder, the d output of the exponent part addition means, and the output of the data format indicating means I are supplied to the a, b, c, and C manuals of the exponent part switching means (equipment), respectively. be done. a and b of the digit alignment means 60;

c、C人力には、それぞれ入力レジスタ10の一部、入
力レジスタ加の一部、指数部加算手段(社)のd出力、
データ形式指示手段(資)の出力が入力される。
C and C manually include a part of the input register 10, a part of the input register addition, the d output of the exponent addition means (company),
The output of the data format indicating means (material) is input.

仮数部加算手段70のa、b、C人力には、それぞれt
行合せ手段のe、f出力、データ形式指示手段間の出力
が供給される。正規化手段80のa、b入力には、それ
ぞれ仮数部加算手段700d出力、データ形式指示手段
間の出力が入力される。符号処理手段90のa、b、C
人力には、それぞれ入力レジスタ10の一部、入力レジ
スタ加の一部、データ形式指示手段間の出力が供給され
る。補正手段100のa、b、c、d、C人力には、そ
れぞれ符号処理手段90のd出力、指数切換手段間のC
出力、データ形式指示手段(9)の出力、仮数部加算手
段70のd出力、正規化手段80のC出力が供給される
。結果レジスタ110には、補正手段100ので出力と
正規化手段80のd出力が供給される。
The mantissa adding means 70 has t, b, and C manpower, respectively.
The e and f outputs of the aligning means and the outputs between the data format indicating means are supplied. The output of the mantissa adding means 700d and the output between the data format indicating means are input to the a and b inputs of the normalizing means 80, respectively. a, b, C of the code processing means 90
Each of the input terminals is supplied with a part of the input register 10, a part of the input register, and an output between the data format indicating means. For a, b, c, d, and C of the correction means 100, d output of the code processing means 90 and C between the index switching means are respectively used.
The output, the output of the data format indicating means (9), the d output of the mantissa addition means 70, and the C output of the normalization means 80 are supplied. The result register 110 is supplied with the output of the correction means 100 and the d output of the normalization means 80.

以下に動作を示す。本実施例の加算装置は、第2図に示
す(a)浮動小数点データ形式と(b)固定小数点デー
タ形式のデータを扱うものとする。まず、入力レジスタ
10,20に第2図に示される伝)あるいは(b)のい
ずれかのデータ形式のオペランドがセットされる。それ
と同時にデータ形式指示手段美には、入力レジスタ10
及び加に(a)のデータ形式のオペランドがセットされ
た場合には@0”がセットされ、入力レジスタ10及び
20 K (b)のデータ形式のオペランドがセットさ
れた場合には11”がセットされる。ここで、入力レジ
スタ10及び加のピット0をそれぞれSa、Sb、 ピ
ット1からピット7なそれぞれga、F、b、ピット8
からピット63をそれぞれMa、Mbとする。次に、指
数部加算手段40では、データ形式指示手段Iの出力が
10″の場合、ii:a、Ebをそれぞれa、b入力で
受は取り、IEa−Eblを実行してd出力から出力す
る。データ形式指示手段(9)の出力が@1”の場合に
は、Ea + Ebが実行され、Ea + Ebの桁上
シと共にd出力から出力される。桁合せ手段60では、
データ形式指示手段(9)の出力が0″の場合、Ea 
> Ebのとき、b入力から入力されるMbを、Ea 
< gbのときに、C人力から入力されるMaを指数部
算手段加のd出力からC人力に供給される指数差11E
a−Eblだけ最下位ピット方向にシフトし、それぞれ
仮数部被加数A、仮数部加数Bとしてe、  f出力か
ら出力し、一方データ形式指示手段蜀の出力が”1″の
場合、C入力には、指数加算手段蜀のd出力Ea + 
Ebが供給されるが、それを無視してa、b入力から入
力されるMa、Mbは、e、f出力から仮数部破開dA
、仮故部加fiBとしてそのまま出力される。これと並
行して、指数部切換手段美では、データ形式指示手段I
の出力が°O″の場合にはa、b入力Ka、Wbの大き
い方を選択し、データ形式指示手段Iの出力が1″の場
合。
The operation is shown below. The addition device of this embodiment handles data in (a) floating point data format and (b) fixed point data format shown in FIG. First, operands in either the data format shown in FIG. 2 or (b) are set in the input registers 10 and 20. At the same time, the data format indicating means includes the input register 10.
In addition, if the operand in the data format of (a) is set, @0'' is set, and if the operand in the data format of input registers 10 and 20K (b) is set, 11'' is set. be done. Here, input register 10 and add pit 0 are Sa, Sb, pit 1 to pit 7 are ga, F, b, pit 8, respectively.
Let the pits 63 be Ma and Mb, respectively. Next, in the exponent adding means 40, when the output of the data format indicating means I is 10'', ii: a and Eb are received as a and b inputs, respectively, and IEa-Ebl is executed and output from d output. When the output of the data format indicating means (9) is @1'', Ea + Eb is executed and outputted from the d output together with the carry bit of Ea + Eb. In the digit alignment means 60,
If the output of the data format indicating means (9) is 0'', Ea
> When Eb, Mb input from b input is Ea
When < gb, Ma input from C human power is added to the index difference 11E supplied to C human power from the d output of the exponent division calculation means.
It is shifted by a-Ebl toward the lowest pit and outputted as mantissa addend A and mantissa addend B from e and f outputs, respectively. On the other hand, when the output of data format indicating means Shu is "1", C The input is the d output Ea + of the exponent addition means Shu.
Eb is supplied, but ignoring it, Ma and Mb, which are input from the a and b inputs, are the mantissa burst dA from the e and f outputs.
, is output as is as a temporary failure part addition fiB. In parallel with this, in the exponent part switching means beauty, the data format indicating means I
If the output of is °O'', select the larger one of inputs a, b, Ka and Wb, and if the output of data format indicating means I is 1''.

C入力から入力される指数部切換手段40のd出力であ
るに+Wb及び桁上シを選択してC出力から出力する。
+Wb and digit C are selected from the d output of the exponent part switching means 40 inputted from the C input, and outputted from the C output.

続いて、仮数部加算手段70では、桁合せ手段60から
送られてくる仮数部被加数Aと仮数部加613をそれぞ
れa、b入力から受は取シ加算しC出力から出力する。
Subsequently, the mantissa adding means 70 receives and adds the mantissa augend A and the mantissa addend 613 sent from the digit alignment means 60 from the a and b inputs, respectively, and outputs them from the C output.

また、仮数部加算手段7゜は、データ形式指示手段Iの
出力が“0”の場合には、仮数部オーバーフローを検出
してd出力から出力し、データ形式指示手段(資)の出
力が@1″の場合には、加算術の最上位ピットからの桁
上シをd出力から出力する。次いで、正規化手段80で
は、データ形式指示手段(9)の出力が°O″の場合に
は、C人力から入力される仮数部加算手段70の加算結
果から最上位ピット方向にシフトされるべきシフトカウ
ントSCを生成し、C出力から出力すると共に、シフト
カウントSC桁分だけ、加算結果を最上位ピット方向だ
シフトし、仮数部加算手段70で仮数部オーバーフロー
が生じた場合、−桁最下位ピット方向にシフトしてd出
力から出力する。
Further, when the output of the data format indicating means I is "0", the mantissa addition means 7° detects the mantissa overflow and outputs it from the d output, and the output of the data format indicating means (material) is @ 1'', the carry shift from the most significant pit of addition is output from the d output.Next, in the normalizing means 80, when the output of the data format indicating means (9) is °O'', , C generates a shift count SC to be shifted toward the most significant pit from the addition result of the mantissa addition means 70 inputted manually, and outputs it from the C output, and also shifts the addition result to the maximum by the shift count SC digits. If the mantissa overflow occurs in the mantissa addition means 70, the signal is shifted in the direction of the lowest pit by -digits and output from the d output.

データ形式指示手段加の出力が@1″の場合には、a入
力から入力される加算結果がそのままd出力から出力さ
れ、シフトカラン)SGは0としてC出力から出力され
る。次に、符号処理手段90では、データ形式指示手段
父の出力が10″の場合、a。
When the output of the data format indicating means is @1'', the addition result input from the a input is output as is from the d output, and the shift column (SG) is set to 0 and output from the C output. In the processing means 90, when the output of the data format indicating means is 10'', a.

b入力から供給されるSa、Sbと仮数部加算手段70
の実質的な減算モード時の符号とから代数的に求まる符
号を生成してd出力から出力する。データ形式指示手段
加の出力が@1”の場合には、Sa + Sbの結果が
符号処理手段90d出力から出力される。続いて、補正
手段100では、データ形式指示手段(資)の出力が1
0”の場合、指数部加算手段閏のC出力からb入力に入
力されるEaあるいはgbに仮数部加算手段70のd出
力からd入力に供給される仮数部オーバーフローが加算
(仮数部オーバーフローが発生した場合+1)され、さ
らに、正規化手段80のC出力からe入力に供給される
SCが減じられf出力から出力され、また、データ形式
指示手段(資)の出力が01”の場合、指数部加算手段
菊からb入力に入力されるEa + Ebと仮数部加算
手段70のd出力からd入力に供給される桁上・りと、
正規化手段80のC出力から供給されるsc (20)
が加算され、その後、Ea + Wb十桁上夛の最上位
ピットからの桁上りと符号処理手段90のd出力からa
入力に入力されるSa + Sbが加算されてf出力か
ら出力される。最後に補正手段100のf出力が結果レ
ジスタ110の上位16ビツトとして、また、正規化手
段80のd出力が結果レジスタ110の下位56ビツト
として格納される。
Sa, Sb supplied from b input and mantissa addition means 70
A code determined algebraically from the code in the substantial subtraction mode is generated and output from the d output. When the output of the data format indicating means is @1'', the result of Sa + Sb is output from the code processing means 90d output.Subsequently, in the correction means 100, the output of the data format indicating means (equipment) is outputted from the code processing means 90d. 1
0'', the mantissa overflow supplied from the d output of the mantissa addition means 70 to the d input is added to Ea or gb input from the C output of the exponent addition means leap to the b input (mantissa overflow occurs). In addition, the SC supplied to the e input is subtracted from the C output of the normalization means 80 and output from the f output, and when the output of the data format indicating means (material) is 01'', the exponent is Ea + Eb inputted from the part addition means 70 to the b input, and the carry bit supplied from the d output of the mantissa addition means 70 to the d input,
sc (20) supplied from the C output of the normalization means 80
is added, and then a is obtained from the carry from the most significant pit of Ea + Wb 10-digit addition and the d output of the code processing means 90.
Sa + Sb input to the inputs are added and output from the f output. Finally, the f output of the correction means 100 is stored as the upper 16 bits of the result register 110, and the d output of the normalization means 80 is stored as the lower 56 bits of the result register 110.

以上説明したように、本発明の加算装置はデータ形式指
示手段によりデータ形式の種数を変えることにより、同
一の金物で浮動小数点加算と固定小数点加算が可能とな
る。
As explained above, the adding device of the present invention enables floating point addition and fixed point addition with the same hardware by changing the genus of the data format using the data format indicating means.

発明の効果 本発明の加算装置はデータ形式が指示手段と、指数部加
算手段と、指数部選択手段と、桁合せ手段と、仮数部加
算手段と、正規化手段と、符号処理手段と、補正手段で
構成することにより、浮動小数点加算と固定小数点加算
を同一の金物で実現でき、金物量を削減できると共に、
ベクトル処理装置に組み込む場合、浮動小数点加算と固
定小数点加算が同じ実行時間で実行できるために、演算
装置の制御を簡単化できるという効果がある。
Effects of the Invention The addition device of the present invention has a data format that includes an instruction means, an exponent addition means, an exponent selection means, a digit alignment means, a mantissa addition means, a normalization means, a code processing means, and a correction means. By configuring it with a means, floating point addition and fixed point addition can be realized with the same hardware, reducing the amount of hardware, and
When incorporated into a vector processing device, floating point addition and fixed point addition can be executed in the same execution time, which has the effect of simplifying the control of the arithmetic device.

【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構成図、第2
図は本発明の一実施例による装置の扱うデータ形式を示
す図、#I、3図は従来における装置の一構成例を示す
ブロック図、第4図は第3図に示す構成側中の浮動小数
点加算装置の一例を示すブロック図、第5図は第3図に
示す構成側中の固定小数点加算装置の一例を示すブロッ
ク図である。 10、20.605 、610 、710 、720・
・・入力レジスタ、(資)・・・データ形式指示手段、
40,640・・・指数部加算手段、 50.630・
・・指数部切換手段、60,650・・・桁合せ手段、
70.660・・・仮数部加算手段、80゜670・・
・正規化手段、90.620・・・符号処理手段、10
0 、680・・・補正手段、110 、690 、9
80・・・結果レジスタ、730・・・加算器 第1図 (a)S:Mの符g E;指数  M:イ反委丈(b)
S:符号 第2図
FIG. 1 is a block diagram showing one embodiment of the present invention, and FIG.
The figure is a diagram showing a data format handled by a device according to an embodiment of the present invention, #I, 3 is a block diagram showing an example of the configuration of a conventional device, and FIG. 4 is a floating diagram in the configuration side shown in FIG. 3. A block diagram showing an example of a decimal point addition device. FIG. 5 is a block diagram showing an example of a fixed point addition device in the configuration shown in FIG. 3. 10, 20.605, 610, 710, 720・
...input register, (capital)...data format instruction means,
40,640...Exponent part addition means, 50.630.
... exponent part switching means, 60,650... digit alignment means,
70.660...mantissa addition means, 80°670...
- Normalization means, 90.620... code processing means, 10
0, 680...correction means, 110, 690, 9
80...Result register, 730...Adder Figure 1 (a) S: M sign g E: Exponent M: I anti-committee length (b)
S: Code Figure 2

Claims (1)

【特許請求の範囲】[Claims] 浮動小数点データ形式のオペランドと固定小数点データ
形式のオペランドに対してそれぞれ浮動小数点加算、固
定小数点加算を行なう加算装置において、データ形式の
種類を示すデータ形式指示手段と、前記データ形式指示
手段の指示により浮動小数点モードでは第1及び第2オ
ペランドの指数部の指数差を生成し固定小数点モードで
は前記第1及び第2オペランドの指数部の加算を行なう
指数部加算手段と、前記データ形式指示手段の指示によ
り浮動小数点モードでは前記第1及び第2オペランドの
指数部の大小関係により定まる前記第1あるいは第2オ
ペランドの指数部を選択し固定小数点モードでは前記指
数部加算手段の加算結果を選択する指数部選択手段と、
前記データ形式の指示と前記指数部の大小関係により浮
動小数点モードでのみ前記指数部加算手段により生成さ
れた指数差に応じて前記第1及び第2オペランドの仮数
部の桁合せを行ないそれぞれ仮数部被加数、仮数部加数
を生成する桁合せ手段と、前記桁合せ手段により生成さ
れた仮数部被加数と仮数部加数を加算する仮数部加算手
段と、前記データ形式指示手段の指示により浮動小数点
モードでのみ前記仮数部加算手段の仮算結果を正規化す
る正規化手段と、前記データ形式指示手段の指示により
浮動小数点モードでは前記第1及び第2オペランドの符
号ピットから代数的に定まる符号を生成し固定小数点モ
ードでは前記第1及び第2オペランドの符号ピットの加
算を行なう符号処理手段と、前記指数部選択手段により
生成された指数と前記符号処理手段により生成された符
号を前記データ形式指示手段と前記正規化手段と前記仮
数部加算手段の指示により補正する補正手段とを備える
ことを特徴とする加算装置。
In an adder that performs floating-point addition and fixed-point addition on operands in floating-point data format and operands in fixed-point data format, respectively, data format indicating means for indicating the type of data format; exponent addition means for generating an exponent difference between the exponent parts of the first and second operands in floating point mode and adding the exponent parts of the first and second operands in fixed point mode; and instructions for the data format indicating means. In floating point mode, the exponent part of the first or second operand determined by the magnitude relationship between the exponent parts of the first and second operands is selected, and in the fixed point mode, the exponent part is selected as the addition result of the exponent part adding means. selection means,
According to the data format instruction and the magnitude relationship of the exponent part, the mantissa parts of the first and second operands are aligned in accordance with the exponent difference generated by the exponent part adding means only in floating point mode, and the mantissa parts of each mantissa part are adjusted. a digit alignment means for generating a summand and a mantissa addend; a mantissa addition means for adding a mantissa augend and a mantissa addend generated by the digit alignment means; and an instruction for the data format instruction means. normalizing means for normalizing the provisional calculation result of the mantissa addition means only in floating point mode; code processing means that generates a fixed code and adds the code pits of the first and second operands in the fixed point mode; An addition device comprising: a data format designation means; a correction means for performing correction according to instructions from the normalization means and the mantissa addition means.
JP60021564A 1985-02-06 1985-02-06 Adding device Pending JPS61182134A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP60021564A JPS61182134A (en) 1985-02-06 1985-02-06 Adding device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP60021564A JPS61182134A (en) 1985-02-06 1985-02-06 Adding device

Publications (1)

Publication Number Publication Date
JPS61182134A true JPS61182134A (en) 1986-08-14

Family

ID=12058514

Family Applications (1)

Application Number Title Priority Date Filing Date
JP60021564A Pending JPS61182134A (en) 1985-02-06 1985-02-06 Adding device

Country Status (1)

Country Link
JP (1) JPS61182134A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH02239323A (en) * 1989-01-27 1990-09-21 Hughes Aircraft Co Register logic computation unit
JPH04205559A (en) * 1990-11-30 1992-07-27 Nec Corp Vector arithmetic unit

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH02239323A (en) * 1989-01-27 1990-09-21 Hughes Aircraft Co Register logic computation unit
JPH04205559A (en) * 1990-11-30 1992-07-27 Nec Corp Vector arithmetic unit

Similar Documents

Publication Publication Date Title
US11347511B2 (en) Floating-point scaling operation
US7236995B2 (en) Data processing apparatus and method for converting a number between fixed-point and floating-point representations
US6233597B1 (en) Computing apparatus for double-precision multiplication
JPH09269891A (en) Method and device for adding partial product, method and device for multiplying floating point, and method and device for floating point product-sum operation
JPS62191926A (en) Arithmetic unit
US10310809B2 (en) Apparatus and method for supporting a conversion instruction
JPS59117637A (en) Floating point multiplier
JPH01302425A (en) Floating decimal point adder/subtracter circuit
JPS61182134A (en) Adding device
US20200183650A1 (en) Radix-1000 decimal floating-point numbers and arithmetic units using a skewed representation of the fraction
US6202078B1 (en) Arithmetic circuit using a booth algorithm
JPH0540605A (en) Floating point multiplier
JPS63186329A (en) Pre-processor for trigonometric function
JPH0251732A (en) Floating point computing element
JPH01282633A (en) Non-normalized number processing system
JPS60186932A (en) Arithmetic device
JPH04160533A (en) Microprocessor
JPS6224322A (en) Floating point arithmetic unit
JPH0323937B2 (en)
JPS6194144A (en) Floating point adder circuit
JPH1185471A (en) Operation method and arithmetic unit
JPH10333886A (en) Multiplying circuit for floating point
JPS6149234A (en) Multiplier circuit for floating point
JPH08263269A (en) Floating-point arithmetic circuit
JPH03150629A (en) Floating point data expression format converting system