Electrical computers and digital processing systems: multicomput – Computer-to-computer data routing – Least weight routing
Reexamination Certificate
1997-04-04
2001-03-27
Banankhah, Majid (Department: 2151)
Electrical computers and digital processing systems: multicomput
Computer-to-computer data routing
Least weight routing
Reexamination Certificate
active
06209041
ABSTRACT:
BACKGROUND THE INVENTION
1. The Field of the Invention
The field of the present invention is computer software driver development. The present invention relates to tools, software frameworks, conventions, etc. to simplify code generation of drivers and provide standardized access points. More specifically, the present invention is directed to methods and computer program products for managing buffers used by multiple entities, such as interconnected software drivers, so that a minimum amount of inter-buffer data transfer occurs between entities during processing by software drivers running in a relatively unprotected operating system mode.
2. Present State of the Art
When multiple components process a stream or amount of data, they will typically place a quantity of data, such as a frame, into a buffer, process the data, and copy the manipulated data into another buffer corresponding to another processing component. If there are multiple independent processing components, the sheer volume of data transfers between buffers can impact processor performance to the degree that it may inhibit the actual processing of the data itself in a timely fashion.
While this problem may occur with any type of data processing between different components, it is particularly acute with the media data found in multimedia applications. This is due, in part, to the large quantities information generated that is in many instances tied to a time frame to effectively produce a veritable “stream” of continuous data for processing. Such media information may come from a live feed over a network, phone line, digitizing hardware, etc. or some other source such as a file and usually must be processed in “real time” so as to be coherent to a viewer or listener.
In a multimedia environment, when processing a stream of media data, the processor performance impact due to data transfers between different buffers creates a latency problem that reduces overall media processing and rendering performance of a hardware/software system. A latency problem may be so bad as to limit what a particular system may handle in terms of multimedia capabilities. The situation becomes exacerbated on general purpose computer systems such as a processor running Windows NT™ by Microsoft® because of the already existing demands on the processor. Reducing latency problems opens many opportunities to provide low-cost multimedia applications that would otherwise be unavailable. If latency is sufficiently reduced by increased system processing efficiency and other means, a general purpose computer system will be able to run multimedia applications that were previously not possible.
Another problem associated with multiple buffers corresponding to multiple processing components is the amount of system resources that may be used and therefore unavailable to other applications. In other words, if each processing component has its own buffer taken from system resources such as system memory, system resources become more scarce thereby causing inefficiencies that may result in degraded system performance.
Shared memory between processes and components attempts to address the resource allocation problem and allow shared buffers. A shared memory facility normally occurs at an operating system “user mode” that has significant safety overhead to assure only authorized applications used designated address spaces in system memory and where user written applications are run. The safety mechanisms are for assuring that one application or process will not impact another application.
In a multimedia environment, it is advantageous to interconnect software drivers to that processing may occur in software drivers that run in a operating system mode with a much higher run priority and little security protection to allow access to actual hardware that the drivers, in many instances, directly manipulate. Many applications are benefited by running in this looser and more performance-oriented mode, generally referred throughout, in Windows NT terminology, as “kernel mode.” Other robust operating systems will have a functionally equivalent mode.
One prime example of a program currently incapable of easily using kernel mode drivers, used throughout this application, comprises graph building functionality that allows a user to select and connect together different processing blocks, called filters, to successively manipulate a stream of multimedia data. The data typically is a series of samples representing sound or video and the processing blocks may include decompression processing for compressed data, special effects processing, CODEC functionality, rendering the data into analog signals, etc.
Such filters typically are located in user mode so that the graph builder portion of the program may interconnect and control their operation and be responsive to user input and rearrangement of processing blocks. Because of the consistent stream nature of multimedia data and the generation of large quantities of data, performance is a critical issue. In a general purpose operating system, system performance caused by repeatedly passing/switching back and forth between user mode and kernel mode can be so degraded as to prohibit certain multimedia applications as mentioned previously.
Furthermore, the processing blocks will many times have hardware associated therewith requiring multiple transitions between user mode and kernel mode components. Such transitions comprise another form of overhead that takes away from the overall multimedia processing system performance. When making transitions between user mode kernel mode, there may also be overhead associated with copying the data between different buffers that would be unnecessary if the processing remained in kernel mode.
What is needed is a mechanism for creating a shared buffer and communicating its existence between independent software drivers so that data may be processed in the same buffer by two or more components in an interconnected software driver environment. It would be advantageous for such a mechanism to also communicate buffer requirements of a software driver to allow a third party agent such as a graph builder dynamically interconnect different software drivers to accomplish a certain media stream processing sequence.
SUMMARY AND OBJECTS OF THE INVENTION
It is an object of the present invention to separate the buffers holding data for processing from a series of distinct processing components so that each processing component can sequentially process the data without necessarily transferring the data to a buffer associated with the processing component.
It is an object of the present invention to reduce the number of inter buffer data transfers associated with the processing of data by a plurality of separate and distinct processing components.
It is another object of the present invention to reduce latency effects in multimedia system.
It is a further object of the present invention to allow general purpose computing systems with the ability to handle data intensive multimedia applications with greater effectiveness than heretofore possible.
It is yet another object of the present invention reduce the need for system resources in processing applications having multiple processing components by providing a standardized way to communicate buffering needs and buffer allocation abilities of a particular component.
It is an object of the present invention to allow a third party component to create available buffer allocation mechanisms as part of interconnecting separate components as determined by first querying and receiving the requirements of all components prior to making the interconnections.
Additional objects and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the invention. The objects and advantages of the invention may be realized and obtained by means of the instruments and combinations particularly pointed out in the appended claims.
To achieve the foregoing objects, and in
O'Rourke Thomas J.
Shaw George H. J.
Woodruff Bryan A.
Banankhah Majid
Bullock, Jr. Lewis A.
Microsoft Corporation
Workman & Nydegger & Seeley
LandOfFree
Method and computer program product for reducing... 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 and computer program product for reducing..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and computer program product for reducing... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2508977