Boots – shoes – and leggings
Patent
1985-07-25
1989-03-14
Shaw, Gareth D.
Boots, shoes, and leggings
G06F 928
Patent
active
048129704
DESCRIPTION:
BRIEF SUMMARY
CROSS REFERENCE TO RELATED APPLICATIONS
This application is related to U.S. Applications having Ser. Nos. 755,321, 758,665 and 752,190.
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a microprogram control system, and particularly to a microprogram control system in a data processing unit employing a pipeline control system, wherein flexibility is imparted to the system by reading microinstructions and suppressing a hardware control mode to execute processings with microinstructions corresponding to each flow when executing an instruction which results in processing several flows.
2. Description of the Related Art
As shown in FIG. 1, for example, prefetched instructions are executed in each stages D, A, T, B, E and W by a pipeline processor. In FIG. 1, 1 is an instruction word register for an instruction word. 2 is a control storage or memory storing microinstructions. 3 to 7 are microinstruction field registers which issue instructions for processing to the hardware corresponding to stages D to W. 8 is a general purpose register group. 9 is a base register. 10 is an index register. 11 is a displacement register. 12 is an address calculator or three input adder. 13 is a logical address register. 14 is a TLB (Address Translation Buffer). 15 is an address buffer. 16 is a buffer memory. 17 is an operand register. 18 to 20 are operand buffer registers. 21 is an operand register. 22 is an operation unit such as an adder. 23 is a result register. The general purpose register group 8 is a register group to be used in the three stages D, B and W.
When an instruction 1 is supplied to the pipeline indicated in FIG. 1, a microinstruction required for executing the pertinent instruction 1 is read and loaded into the register 3 in stage D (decode). Simultaneously, a value is loaded into the registers 9, 10 and 11 to calculate the address for the operand (Opr 2 in the figure) which is required for executing the pertinent instruction 1. The address of the operand Opr 2 is calculated by the calculator 12 in stage A (address) and the result is loaded into the logical address register 13. The TLB 14 and buffer memory 16 are indexed through the stages T (translate) and B (buffer access) and a value of the operand is loaded into the register 17. At this time, a value of operand Opr 1 is also loaded into the register 21. At this point, the specified operation is carried out by operation unit 22 in the stage E (execution) and the operation result is loaded into the general purpose register 8 in the stage W (write).
The pipeline control is conducted as explained above and the processings for each stage advances in such a way that the microinstructions are read from the control memory 2 shown in the figure and these instructions are executed sequentially. On the one hand, when of executing a given instruction 1, the necessary processings are sometimes not completed by only the processing flow executed along the pipeline processing stages from stage D to stage W in FIG. 1 (called the processing flow). In general, a plurality of processing flows are required for execution of some instructions and the respective processing flows are desired to be completely executed by microinstructions. In such a case, the bit groups of microinstructions corresponding to one flow of processing read from the control memory 2 are forced to correspond to the controls for respective stages of the pipeline and thereby the respective corresponding bit groups instruct the processing in respective stages from the registers 3, 4, 5, 6, 7 shown in FIG. 1.
As explained above, instructions have been developed for a plurality of flows and processing has been conducted by reading microinstructions corresponding to respective flows from the control memory 2. But, there is a restriction in the time (a minimum available or necessary) for accessing the control memory 2 and as a result the controls in stage D of each flow has been performed by hardware and it has been difficult to employ a control method using microprograms
REFERENCES:
patent: 3800293 (1974-03-01), Enger et al.
patent: 4179736 (1979-12-01), Wilhite
patent: 4187539 (1980-02-01), Eaton
patent: 4197578 (1980-04-01), Wada et al.
patent: 4228497 (1980-10-01), Grupta
patent: 4245302 (1981-01-01), Amdahl
patent: 4251859 (1981-02-01), Momose et al.
patent: 4373180 (1983-02-01), Linde
patent: 4390946 (1983-06-01), Lane
patent: 4439827 (1984-03-01), Wilkes
patent: 4631672 (1986-12-01), Sakamoto
IBM Technical Disclosure Bulletin, vol. 25, No. 8, Jan. 1983, pp. 4300-4301, New York, US; D. W. Finney et al.: "NLM ROS Bit Sets Up Unique First Microcycle".
Kitamura Toshiaki
Oinaga Yuji
Onishi Katsumi
Chun Debra A.
Fujitsu Limited
Shaw Gareth D.
LandOfFree
Microprogram control system does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Microprogram control system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Microprogram control system will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-898086