JPS5856034A - Data processing unit - Google Patents

Abstract

PURPOSE:To decrease the number of steps of a microprogram greatly by specifying addresses of a constant ROM on indirect basis, and by using microinstructions for specifying arithmetic between a constant and a variable in common without reference to the contents of the constant. CONSTITUTION:Prior to the arithmetic of an expanded expression of sinX, x, s<2>, and 0 are set to resistances R1, R2, and R3 in a register group 4. Then, a constant ROM 2 uses the contents of a counter 8 as a ROM address for addressing to output a constant from an adress (k) of the ROM 2. Once a microinstruction is executed, 1 is added under the control of a microprogram storage/control part 9. Consequently, a new step is executed to perform correct arithmetic while terms are changed successively even when the same step is executed repeatedly.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a microprogram type data processing device that calculates polynomials using constants.

2. Description of the Related Art In a microprogram type data processing device, polynomial approximation expressions are generally used when executing high-performance instructions such as trigonometric functions. For example, if −lnx is expressed by a polynomial approximation formula (interlayer formula) and l is large, then 5lnx=a6x+ al x” + *@x
Art +-(1). Here&11oJeal...
are constants, and the constants for each of these terms are calculated in advance and stored in a constant ROM in order to shorten calculation time and improve calculation accuracy. Then, by specifying the ROM address of the constant ROM in the address field AD of the microinstruction (type 2) shown in Figure 1, a fixed constant is extracted into the corresponding term, and then the constant and variable are Perform calculations between. In addition, the first
In the figure 1. op is a function designation part of the instruction, and D and B are register designation parts.

This register designation part is an output optical register for the result of the operation designated by BK in the function designation part OP.

A register that holds the second instruction for the operation is specified.

Figure 2 shows a conventional data processing device that performs calculations such as polynomial approximation expressions when executing such a microinstruction, and 1 indicates a microprogram storage section in which various microprograms are stored and its control. This is a microglogram class consisting of parts (hereinafter referred to as MC8). 2 is a constant ROM, which outputs a corresponding constant from the address position specified by the address part AD of the microinstruction in FIG. 1 given from the MC81. The output of the constant ROM 2 is sent onto a first node (hereinafter referred to as A-4) 3. 4 is register R1, R2, R
This is a register group consisting of various registers such as 3 and R4. The register group 4 is connected to the A/J path 3, a second path (hereinafter referred to as B path) 5, and a third path (hereinafter referred to as D path) 6, and is controlled by MCs 8 and J. The data transferred from the D path 6 is held in the specified register, and the data held in the specified register is output to the A path 3 or the B path 5.

1 is an arithmetic unit. The calculation unit 1 performs calculations on the data I taken in from the A path 3 and the B path 5 under the control of the MC 8J, and outputs the result onto the D path 6.

In such a conventional data processing device, when calculating the expansion of a polynomial, for example, gin ROM of constant ROM each time
It is necessary to specify the attos and read the corresponding constant. Therefore, as is clear from the operation flowchart of lln x shown in FIG. 3, 1 al smX(RJ) → R4 (in the above explanation, 11 is replaced by ml).
The membership fee is the microinstructions that are different for each item, such as microinstructions replaced by microinstructions (i.e. instructions whose address part AD is different from the above microinstructions), etc., so the disadvantage is that the number of steps in the microprogram becomes extremely large. there were.

The present invention has been made in view of the above circumstances, and its purpose is to:
By using the indirect specification method for constant ROM addressing, the microinstructions that specify operations on constants and variables can be shared regardless of the contents of the constants, which significantly reduces the number of microinstruction steps. The goal is to provide a data processing method that can.

Hereinafter, one embodiment of the present invention will be described with reference to the drawings. Note that the same parts as in FIG. 2 are denoted by the same reference numerals, and the description thereof will be omitted. In FIG. 4, 8 is a counter (hereinafter referred to as CTR) that specifies the ROM address of constant ROM Z;
9 is a MOB having the same function as the MC 81 in FIG. C'rR& is controlled by MC89K. MC8
s M If a (indicating constant ROM 2) is specified in the register specification section of the microinstruction (tie fr) shown in Figure 5, CTR is set at the end of execution of the microinstruction.
For example, K is set to increment 8. Note that the register specification section A generally specifies the register that holds the first operand, and in this embodiment, when A = a, the constant ROM j is treated as a kind of first operand register. treated as shi. FIG. 6 shows an example of the storage specification of the constant ROM J, where the constant ROM j has addresses, k+1, k+2, etc.
The constants 1., '1, &@, etc. of the expansion formula of sln x in the above equation (1) are stored in advance in .

Next, an explanation will be given of the operation of the operation of the expansion formula of t'ginx according to an embodiment of the present invention with reference to the diagram in FIG. Now, the microprogram for calculating the expansion formula of 5inx is executed, and before calculating each term, register group 4 is executed.
X e X” in registers R1, R2, and R1 in
It is assumed that eO is set. Surikiwachi (RJ)
==X.

(RJ)=! ”, (RJ) = O. In such a state, the address k of the constant a at the beginning of the series of constants applied to the operation of the expansion of stn x is
Loaded into CTRl under the control of MCB 1 (7
Step S1), constant, ROMj is the content of CTRg t-R
The address is signaled as an OM address, and the constant R
A constant 10 is output from address OM j. Therefore, when the register specifying part A tea +, that is, the constant ROM I, of the next microinstruction (see Figure 5) is specified, and the register specifying part B specifies R1, that is, the register JRJ, the output of the constant ROM M Sunawa. That is, constant a is output to A path 3, and at the same time register R1 in register group 4
The contents (X in this case) are output on the B path 5.

Under the control of MC8J, the calculation unit 7 multiplies the data on the A/( path 3 and the data on the B path 5, that is, 1.×(R
1) and outputs the result on the D path 6. The storage destination of the data on this Dt4 path 6 is specified by the register specification part of the microinstruction mentioned above, and in this embodiment, the data on the D path 6, that is, the multiplication result of 1 x (RJ) (in this case, a6x) is stored in register R4 in register group 4. (Step 82) Then, when a microinstruction that specifies 1, that is, the constant lROM 2, is executed in the register specification part A described above, when the execution of the instruction ends, K C
TR# is incremented by 1 under the control of MC89. As a result, the content of C'rR8 becomes +1. When the operation (step, de s2) of the first term (first term in this case) is completed, the next microinstruction calculates the sum of the expansion expressions up to the first term (R4) + (RJ). (In this case, R6x is obtained), and the result is stored in register RjK in register group 4 (step 83). In the next microinstruction, the variable calculation (R2) x (R1) of the t+i term is performed in preparation for the operation of the 1+1 term (in this case, the 2nd term) (in this case, R3 is obtained), The result is stored in register RIK in register group 4 (step 8
4). It is determined whether or not the computation up to the number of terms specified in gray has been completed (step, fBB), and if not, steps fBJ to S4 are executed again. As mentioned above, the contents of KCTR# have already been incremented by +1 (in this case, it is +1), and the output inner address of constant ROM J is the address position next to the address in the previous step S2 (in this case, it is +1). Therefore, when a new step S2 is executed, the constant 1 is the aS in the previous step S2 (a- in this case). It is applied to the calculation of the next term instead of ``i-H (Jin), aS ), and even if you repeat the same step, f money repeatedly, the correct calculation will be performed while changing the terms one after another. You can. And,
Stay until the operation for the specified number of terms is completed! 82~S
By repeatedly executing step 4, the calculation of the expansion formula of sln x is saved.

In the above embodiment, the calculation of the expansion formula of elm x is explained, but it goes without saying that the present invention is not limited to this and can be applied to calculations of various polynomials.

As detailed above, according to the data processing device of the present invention,
Since the flow of the calculation execution part for each term of the polynomial can be looped, the number of steps in the microprogram can be significantly reduced.

[Brief explanation of drawings]

Figure 1 is a diagram showing the format of a microinstruction (tie 'f2), Figure 2 is a diagram showing the configuration of a conventional data processing device, and Figure 3 explains the operation of a conventional data processing device. 1. FIG. 4 is a flowchart showing an embodiment of the data processing device of the present invention; FIG. 5 is a diagram showing the format of microinstructions (type 1) applied in the above embodiment; FIG. 6 is a diagram showing an example of the contents of the constant ROM, and FIG. 7 is a flowchart for explaining the operation of one embodiment of the present invention. 1.9...Microglodalum case 1 part (MC8)
, 2... constant ROM, 4... register group, 7...
Arithmetic unit, 8... Counter (CTR).

Claims (1)

[Claims] A microprogram type data processing device is equipped with a constant ROM in which various constants are stored in advance and a counter that specifies the ROM address of this constant ROM, and performs operations on constants and variables for each term in a fixed order. When performing a polynomial calculation, after setting an initial address unique to the polynomial calculation to be applied to the counter, the contents of the counter are updated each time the constant ROM is used once by a microinstruction. Data processing equipment.