JPS6242234A - Arithmetic processing system for variable having bit train attribute - Google Patents

Abstract

PURPOSE:To perform the arithmetic processing of a variable having the bit train attribute with high efficiency by carrying out this arithmetic processing every 8 bytes. CONSTITUTION:The arithmetic processing is carried out in the following prescribed procedure for a variable having the bit train attribute described in a PL/I language. That is, a machine word instruction train needed for arithmetic processing and a machine work instruction train used for selection of the corresponding routine in response to the bit length are previously stored in a PL/I library and the linked to another object module when a load module is produced. Then the source data set is supplied to a PL/I complier and the output of this complier is stored in the object data set. Then the object data set is linked to the contents of the PL/I resident library and the load module data set is produced and executed.

Description

[Detailed Description of the Invention] [Summary] An instruction sequence for performing arithmetic processing on a variable with a bit string attribute written in the PL/I language in 8-byte units is registered in the PL/I library, and the bit string attribute is An object module obtained by translating a PL/r source program for performing arithmetic processing in the PL/I library is linked with an object module stored in the PL/I library.

[Industrial application field]

The present invention relates to an arithmetic processing method for variables having bit string attributes written in the PL/r language, which allows arithmetic processing of variables having bit string attributes to be performed in units of 8 bytes.

C. Prior Art and Problems] The CHILL language is known as a language having a bit string attribute, and in the CHILL language, arithmetic processing of variables having a bit string attribute is performed in units of focus. The PL/I language can also perform arithmetic processing on variables with bit string attributes, but conventional PL/I compilers also create source instructions for performing arithmetic processing on variables with bit string attributes written in the PL/I language. It converts a sequence into a sequence of machine language instructions that perform bit-by-bit processing. If the arithmetic processing of variables with bit string attributes is performed bit by bit, the disadvantage is that the processing time is long.

[Purpose of the invention]

The present invention is based on the above consideration, and an object of the present invention is to provide an arithmetic processing method for a variable having a bit string attribute, which allows arithmetic processing of a variable having a bit string attribute to be performed in units of 8 bytes.

[Means to achieve the purpose]

Therefore, the arithmetic processing method for variables with bit string attributes according to the present invention uses machine language instruction strings to execute operations on variables with bit string attributes written in PL/I language.
In a computer system in which a load module is created by registering the above machine language instruction string in the I library and linking it with other object modules, the above machine language instruction string is a variable with a bit attribute. This feature is characterized in that it is configured to process data in units of 8 bytes.

[Embodiments of the invention] Hereinafter, the present invention will be explained with reference to Examples.

An example of arithmetic processing of a variable with a focus attribute written in the PL/I language is shown below.

DECLARE I A.

2 8 B (T (68) ALIGEND.

2 CBIT (64) UNALIGEND.

2 D BIT (64) UNALIGEND.

D-B&C The above program is a program that performs an AND operation on each bit from the beginning of variables B and C, and assigns the result to each variable. Variable B in the above program.

C and D are allocated on the storage area as shown in FIG.

The processing procedure for performing arithmetic processing on variables having bit string attributes as described above is as follows.

(1) Assign variable B to consecutive registers XI and X2, and assign variable C to registers Yl and Y2.

1-11 bytes: ICMXl or Yl, 8. Variable 1-2
2 bytes: I CM Xi or Y2.12. Variable 1-
33 bytes: ICMXl or Y2, 14. Variable 1-44
Byte: LXl or Yl, variable 1-55 byte: LXl
or Yl, variable ICMX2 or Y2,8. Variable 1-66 bytes: LXl or Yl, variable ICMX2 or Y2.12. Variable 1-77 part: LXl or Yl, variable ICMX2 or Y2, 14. Variable 1-88 bytes: LMXl or Yl, variable (2) Perform shift processing, etc., taking into account when the variable is not at the beginning of the register (straddling between registers or not at a byte boundary).

(3) Execute the NR command.

(Substitute the result of 41(3) into the variable.

4-11 bytes: STCMXl or Yl, 8. Variable 4-
22 bytes: STCMXl or Y2.12. Variable 4-3
3 bytes: STCMXl or Y2.14. Variable 4-44
Byte: 5TX1 or Yl, variable 4-55 byte: 5T
X1 or Yl, variable STCMX2 or Y2.8. Variable 4-66 bytes: 5TX1 or Yl, variable STCMX2
Or Y2.12. Variable 4-77 bytes: 5TX1 or Yl, variable STCMX2
Or Y2.14. Variable 4-88 bytes: STMXl or Yl, variable (51 (1
1, (2], (41, (5)). Note that the ICM is 1, (2), (41, (5)).
ASK instruction, STCM is 5TORE CHRACTER
The S UNDERMASK instruction is a load instruction, LM is a double-length load instruction, ST is a store instruction, STM is a double-length store instruction, and NR is an AND instruction. Note that the calculation process for variables with bit attributes is as follows:
In addition to the AND operation, there are also the following: - OR or NOT operation - ALL or ANY built-in function - BOOLMi built-in function - INDEX, 5UBSTR, VERIFY built-in function - Concatenation operation, etc. FIG. 2 shows a flowchart of the bit string operation module.

The machine language instruction sequence for performing the processing in (1) to 5) above and the machine language instruction sequence for selecting the corresponding routine according to the bit length are stored in advance in the PL/I lamp library and loaded. - Linked to other object modules when creating a module. Figure 3 is a diagram showing the flow of PL/I language processing, where the source data set is P
It is input to the L/I compiler, its output is stored in an object datacent, the object data set and the contents of the PL/I resident library are ringed, a load module datacent is generated and executed.

〔Effect of the invention〕

As is clear from the above description, according to the present invention, operations on variables having bit string attributes can be executed in units of 8 bytes, so this type of operation can be processed efficiently.

[Brief explanation of drawings]

FIG. 1 is a diagram showing an example of the allocation of variables with bit string attributes on a storage area, FIG. 2 is a flowchart of a bit string operation module, and FIG. 3 is a diagram showing the flow of PL/1 language processing.

Claims (1)

[Claims] Registering and placing a machine language instruction sequence for executing operations on variables with bit string attributes written in the PL/I language in a PL/I library, and linking the machine language instruction sequence with other object modules. In a computer system in which a load module is created by the following, the machine language instruction string is configured to process a variable having a bit attribute in units of 8 bytes. calculation processing method.