Browse Prior Art Database

Overlapped Cycle Technique

IP.com Disclosure Number: IPCOM000080047D
Original Publication Date: 1973-Oct-01
Included in the Prior Art Database: 2005-Feb-26
Document File: 2 page(s) / 28K

Publishing Venue

IBM

Related People

Dumstorff, EF: AUTHOR [+4]

Abstract

As shown in Fig. 1, this system uses an address to ROS 5 on line 6 to generate an instruction on line 7. The address includes a first portion from instruction address register (IAR) 8 and a second portion from one of page registers 10 and 11. The page register used in forming a particular address is selected by latch 13, which influences which of AND blocks 14 and 15 is currently active.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 72% of the total text.

Page 1 of 2

Overlapped Cycle Technique

As shown in Fig. 1, this system uses an address to ROS 5 on line 6 to generate an instruction on line 7. The address includes a first portion from instruction address register (IAR) 8 and a second portion from one of page registers 10 and 11. The page register used in forming a particular address is selected by latch 13, which influences which of AND blocks 14 and 15 is currently active.

Each cycle or instruction time, as seen in Fig. 2, is partitioned into R, B and A times. One instruction requires four partitions, which is accomplished by overlapping the cycles during R time to reduce the cycle time by 25%. In the overlapped cycle technique employed herein, the instruction for cycle 2 is latched during the first half of the initial R time and the result from cycle 1 (the new value for the paging register) is latched during the second half of the final R time of cycle 1.

During the first half of the final R time of cycle 1 (the load instruction) the selected page register contains the value used to specify the preceding instruction. During the second half of R time, when the branch is being executed, the branch location is loaded into the selected page register from bus 17. The above sequence results in getting the actual jump instruction used to branch to the new page of data from the same page the preceding instruction was on, even though that preceding instruction is used to reload the selected page register with the new page number...