Electrical computers and digital data processing systems: input/ – Input/output data processing – Input/output access regulation
Reexamination Certificate
1998-10-14
2001-10-16
Lee, Thomas (Department: 2182)
Electrical computers and digital data processing systems: input/
Input/output data processing
Input/output access regulation
C710S040000, C710S113000, C710S116000
Reexamination Certificate
active
06304923
ABSTRACT:
TECHNICAL FIELD
This invention relates generally to controlling data transfer between a memory and a device in a computer system, and more particularly, to prioritizing service of multiple data transfer requests.
BACKGROUND OF THE INVENTION
In a typical computer system, multiple devices are able to write data to and read data from a main memory. A memory controller couples these devices with the main memory and controls the timing and sequence of such data transfer operations. Referring to
FIG. 1
, a memory controller
10
includes arbiter circuitry
12
that receives a plurality of signals indicative of requests for data transfer operations—commonly known as priority requests PREQ ∅-PREQ M. Each of these priority request signals corresponds to a device requesting to write data to or read data from a main memory
14
. Example devices include a microprocessor and input/output (I/O) devices, such as peripheral component interconnect (PCI) bus devices, industry standard architecture (ISA) bus devices, integrated drive electronics (IDE) devices, accelerated graphics port (AGP) devices, small computer system interface (SCSI) devices, and universal serial bus (USB) devices, to name just a few examples. The arbiter
12
prioritizes the various priority requests, and the memory controller
10
correspondingly initiates and controls the data transfer operations.
Typically, the priority requests are simply queued by the arbiter
12
, with the requested data transfer operations occurring on a first-come first-served basis. Certain of today's computer systems can reorder some data transfer operations, such as by providing “read around write” capability. Also, certain computer systems allow various write operations to be reordered or combined to minimize the frequency of write operations to the memory subsystem. Some computer systems provide fixed service priorities, in which priority requests from some devices (e.g., the CPU) always receive service prior to priority requests from other devices.
Many devices included in today's computer systems have maximum latency requirements—i.e., a maximum time interval following the requested data transfer operation by which such operation must be effected. For example, a video display device typically requires regular refresh of the display. The display refresh must occur regularly and timely to avoid undesirable artifacts showing on the display. As another example, a software-based modem requires timely transfer of coder/decoder (CODEC) data in order to avoid corruption of the modem data. To satisfy such latency requirements, today's computer systems include separately dedicated memories or device buffers. For example, the video display in today's computer systems commonly has a separately dedicated video memory and associated video memory controller. As another example, software-based modems have a large CODEC data buffers for storing received data and transmit data.
Given the improved production economies and performance improvements offered by today's increasingly integrated computer systems, it is highly desirable to minimize the number of separately dedicated memory circuits and associated controller circuits. However, today's main memory access arbitration schemes cannot satisfy the latency requirements of certain devices commonly included in a computer system.
SUMMARY OF THE INVENTION
In accordance with the present invention, a method is provided for controlling data transfer operations between a memory and a device in a computer system. The method includes receiving a data transfer request and a latency identifier value corresponding with a maximum time interval for servicing the data transfer request. The method further includes waiting for a time interval no greater than the maximum time interval and then servicing the data transfer request. The latency identifier value may be successively modified and compared to a predetermined value, with servicing of the data transfer request then being initiated on the basis of that comparison.
Multiple latency identifiers corresponding with data transfer requests from multiple computer system devices may be received, with one or more of the multiple latency identifiers being successively modified as time goes on. Comparing the multiple latency identifiers may be performed to identify a most urgent data transfer request. The latency identifier value corresponding with the most urgent data transfer request may then be compared to a predetermined range of values. If it is determined that this latency identifier value falls within the predetermined range of values, then servicing of the most urgent data transfer request is initiated.
REFERENCES:
patent: 5278967 (1994-01-01), Curran
patent: 5394541 (1995-02-01), Chesley et al.
patent: 5506968 (1996-04-01), Dukes
patent: 5555383 (1996-09-01), Elazar et al.
patent: 5818464 (1998-10-01), Wade
patent: 5862353 (1999-01-01), Revilla et al.
patent: 5862355 (1999-01-01), Logsdon
patent: 5956493 (1999-09-01), Hewitt et al.
patent: 5996037 (1999-11-01), Emnett
patent: 6078976 (2000-06-01), Obayashi
patent: 97/29432 (1997-08-01), None
Dorsey & Whitney LLP
Lee Thomas
Micro)n Technology, Inc.
Park Ilwoo
LandOfFree
Method for prioritizing data transfer request by comparing a... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for prioritizing data transfer request by comparing a..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for prioritizing data transfer request by comparing a... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2617130