Information processing

Electrical computers and digital processing systems: multicomput – Computer-to-computer data routing – Least weight routing

Reexamination Certificate

Rate now

  [ 0.00 ] – not rated yet Voters 0   Comments 0

Details

C709S241000, C709S241000, C710S005000

Reexamination Certificate

active

06292821

ABSTRACT:

BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to methods and systems for processing information.
2. Related Art
The majority of present day computers process information using the von Neumann information processing model where a set of instructions is stepped through sequentially, each instruction being initiated by a pulse from a regular train of clock pulses. Such computer systems have proved very useful in situations where information is predictable and the same well defined set of operations needs to be carried out repetitively to process the information. Another known method for processing information is parallel processing where there are multiple processors which concurrently process information. In one such architecture multiple identical processors are used to carry out the same operation on many items of data concurrently, rather than passing the data sequentially through a single processor, thereby processing the information at a greater rate.
In both sequential and parallel architectures the processing elements are typically connected in a fixed or relatively inflexible configuration with the various component parts being controlled by regular clock pulses to ensure that the information remains synchronised as it is passed from one component to another within the system.
Information processing systems and methods based on the traditional von Neumann model or using existing parallel processing architectures have proved useful in domains where the information processing task can be easily reduced to a series of highly repetitive, well ordered and relatively simple steps. They have proved far less successful in problem domains which involve significant amounts of pattern matching and classification where data may be noisy and unpredictable. In such domains algorithms are often brittle, causing failures when unanticipated data is provided as input. As systems become larger and more complex the likelihood that an unanticipated state may occur increases, similarly increasing the probability of the system failing or giving incorrect results.
SUMMARY OF THE INVENTION
In accordance with a first aspect of the present invention there is provided an information processing system comprising:
a first data pool for containing information in the form of data items;
a plurality of processing elements disposed within said first data pool for processing data items; and
a system controller arranged to control the processing elements to take selectively at least one data item from a plurality of data items in said first data pool; in which system, each of the processing elements is arranged to perform, autonomously and asynchronously, a respective operation on its selected at least one data item and produce a resultant data item, and to place the resultant data item into said first data pool.
Preferably, said system controller is arranged such that said selective taking of at least one data item from a plurality of data items in said first data pool is on a random basis.
Preferably, systems of the present invention comprise a second data pool for containing input information and output information in the form of data items, said second data pool being adjacent to said first data pool and separated from it by a boundary; and further comprise a transfer means for importing data items of input information from said second data pool to said first data pool and for exporting data items of output information from said first data pool to said second data pool.
The transfer means may be constituted by one or more of the processing elements.
Preferably, at least one of said processing elements constituting the transfer means is arranged to import data items of input information from said second data pool to said first data pool, such a processing element being referred to as an importing processing element; and
at least one of said processing elements constituting the transfer means is arranged to export data items of output information from said first data pool to said second data pool, such a processing element being referred to as an exporting processing element.
The importing processing element may be a different processing element from the exporting processing element. Alternatively, a single processing element may constitute both said importing processing element and said exporting processing element.
Preferably, said importing processing element is arranged to select a data item of input information from said second data pool in accordance with a first predetermined criterion, to create a copy of said selected data item of input information and place said copy into said first data pool, and to destroy said selected data item of input information.
The exporting processing element may be arranged to select a data item of output information from said first data pool in accordance with a second predetermined criterion, to create a copy of said selected data item of output information and place said copy into said second data pool, and to destroy said selected data item of output information.
Preferably, each of the processing elements is arranged to distinguish between data items that it is allowed to operate upon and data items that it is not allowed to operate upon.
Preferably, at least one of said processing elements is arranged to perform a join operation on two or more selected data items.
Preferably, at least one of said processing elements is arranged to perform a break operation on a selected data item.
Preferably, said system controller is arranged such as to determine the rate at which at least one of the processing elements performs its selection from said first data pool.
In accordance with a second aspect of the present invention, there is provided a method of processing information comprising the steps of:
(i) populating a first data pool with a plurality of data items and a plurality of processing elements, and, for each of the processing elements,
(ii) selecting, in accordance with a predetermined function, at least one of the data items currently in the first data pool and taking the selected data item(s) from the first data pool,
(iii) performing, autonomously and asynchronously, by the associated processing element, a respective operation on its selected data item(s) and producing a resultant data item, and
(iv) placing the resultant data item in said first data pool.
Preferably, said predetermined function of the selecting and taking step randomises the selection of said at least one data item.
Preferably, the selecting and taking step includes a substep of distinguishing between data items that the processing element is allowed to operate upon and data items that it is not allowed to operate upon.
Preferably, said step of populating said first data pool with a plurality of data items comprises transferring data items which meet a first predetermined criterion relating to input information from a second data pool to said first data pool.
There may be included a step of transferring data items which meet a second predetermined criterion relating to output information from said first data pool to said second data pool.
Preferably, the performing step comprises a join operation on two or more selected data items.
The performing step may comprise a break operation on a selected data item.
Preferably, said predetermined function defines the rate of selection of said at least one data item.
Systems of the present invention are highly parallel, with constituent processing elements operating asynchronously on the data. The behavior of the information processing system can be controlled not only by the specific operations performed by the processing elements, but also by the dynamics which affect the rate at which particular processing elements operate.
A unit of information is typically represented by multiple data items, or a “concentration” of data items, and this concentration can also be modified so as to affect the results of the processing carried out by the system. The transfer of data between processing elements is a result of each p

LandOfFree

Say what you really think

Search LandOfFree.com for the USA inventors and patents. Rate them and share your experience with other people.

Rating

Information processing does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Information processing, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Information processing will most certainly appreciate the feedback.

Rate now

     

Profile ID: LFUS-PAI-O-2485930

  Search
All data on this website is collected from public sources. Our data reflects the most accurate information available at the time of publication.