JPS6349986A - Data processing system and method - Google Patents

Abstract

(57) [Abstract] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to data processing systems and methods, and more particularly, to
The present invention relates to, but is not limited to, a system and method for processing data obtained during a logging operation for investigating geological formations through a borehole.

BACKGROUND OF THE INVENTION Traditional Techniques Borehole logging operations are an important step in discovering and developing underground resources of hydrocarbons and other valuable raw materials. Typically, after a borehole is drilled, an elongated logging instrument, or sonde, is lowered to the bottom of the borehole and electrical power and signals are transmitted between surface equipment and the sonde via armored cables containing electrical conductors. so that The sonde is then pulled upward along the borehole by hoisting the cable, and sensors within the sonde are used to measure electrical, acoustic and/or nuclear parameters of the formation through which the borehole has been passed. . Measurements are sent upwards via cables and recorded on devices on the surface. Analysis and interpretation of these measurements at the borehole site during the logging operation or later at a remote location can provide information regarding the presence and productivity of hydrocarbons and other feedstocks.

As hydrocarbon resources are being depleted, commercial efforts are being made to exploit increasingly difficult-to-extract resources. Accordingly, the logging instruments used to investigate such resources have become increasingly complex and sophisticated;
It is common practice to provide a computer on ground equipment at the location of the borehole to monitor and control the operation of the sonde and to collect and analyze measurements obtained by the sonde. Such computers typically encode and send timing and other control signals to the sonde to initiate measurement cycles, and receive and decode signals from the sonde to provide data regarding measurements and sonde operation. retrieve and analyze the data to monitor the sonde's operation, generate and send control signals to the sonde to optimize the sonde's operation based on the conditions of the borehole and formation, and monitor the depth of the sonde in the borehole; processing the data to obtain the required measurements, processing the measurements to derive a value for the formation parameter, and forming a real-time display representing the selected operating parameters of the sonde, the measurements, and the formation parameter. Perform various functions. These various functions place a considerable burden on the computer system. It has proven very difficult and time consuming to implement a computer system to perform all necessary tasks quickly and accurately enough to obtain acceptable logging speeds.

The data collection and control needs described above are also found in other environments, such as avionics and aerospace.

Problems to be Solved by the Invention A particular problem here is that computers must perform data processing functions that are essentially independent of processing time and duration.

Process control functions, which are influenced by the time and duration of the process, must also be performed.

Therefore, for example, processing data obtained from a sonde using a Fourier transform is a time-independent task, and the same result as the transform output is always obtained from the same set of data values, regardless of the time and duration of the processing. It will be done. In contrast, the collection of data from the sonde takes an hour, and the actual value found by examining the last of the series of received data items (this is the next control to send to the sonde). (affecting the signal) is based on the rate at which data items are sent up the borehole and the rate at which successive data items from the sonde are examined.

Variations in the throughput speed of the equipment being raised and lowered along the borehole will affect the results. Furthermore, the rate at which signals are transmitted up the borehole is typically based primarily on the characteristics of the sonde, and therefore the reception time of each data item at the surface cannot be controlled by surface computer equipment.

Similarly, the characteristics of the sonde require that it be commanded down the borehole at sufficient speed to ensure optimal operation. In order to quickly change the sonde's behavior in response to changing conditions, time-independent actions should not interfere with co-occurring time-based functions to avoid data loss or defects. There must be. Also, similarly.

It is also important to ensure that time-based functions do not introduce errors into time-insensitive processing.

These characteristics may be affected by system changes in hardware (e.g., changing the clock speed of a processor) or system changes in software (e.g., adding a program module).
must be ensured regardless of the

Techniques such as interrupt handling are known for coordinating the execution of multiple functions. However, these techniques must be applied very carefully to ensure sufficient operating speed and to avoid contention in the use of processor resources such as random access memory. This is especially true for modern borehole logging operations, where advanced measurement techniques generate large amounts of data using sondes, which require precise control and , is expensive due to search demands to complete logging tasks as quickly as possible. Therefore, providing a reliable, efficient, and effective data processing system that operates in such an environment is extremely difficult and costly.

The problem becomes even more serious when multiple tasks are to be performed simultaneously, either by time-sharing a processor or by operating multiple processors in parallel.

Various means are known for subdividing the functions performed by a computer and for coordinating the operation of multiprocessor systems. for example.

Information Processing 77, International Federation on Information Processing Society (Informati
on Processing 77, Interna
tional Federation of Info
ra+ation Processing 5ocie
G. Kahn and D.B. Tias (1977), pp. 993-998.
MacQueen (D, B, MacQueen) ``Coroutin and network of parallel processing''
es and networks of parallel
The paper entitled I process) J discloses a system for performing multiple processes that can be executed simultaneously and communicate with each other via "channels."

These channels act as first-in, first-out buffers that carry information in only one direction from the process that generates the data to the process that requests the data. You can access the channel you connect to and receive the next following item in the sequence stored in the channel. Such a system is an example of a so-called data flow architecture.

Another type of data flow system was introduced in the 1976 Proceedings on the 1976 International
Conference on Parallel Processing (P
roceedings of the Interna
tional Conference on para
100-105 of llel Processing)
D, P, Misnath (D, P, Misnath) published on the page
``Performance analysis of data flow processors (Pe
rformance analysis of ad
ATA-Flow Processor) J and Byte, McGraw-Hill.
aw-Hill) Volume 10, No. 5 (May 1985)
W.G. published on pages 201-214 of
Paceman (υ, G.

``Applying data flow in t
It is described in a paper entitled he real tiorld)j. Such systems communicate between processes by data packets of fixed size, and a process does not execute until all data items required to perform its function are completely obtained.

Once all required data items are obtained, execution of the process begins, with no control required as to whether or to what extent the process is executed. A similar data flow system is available at Cornish.
ish et al., U.S. Pat. No. 4,197.589; a concurrent task and instruction processor is disclosed in U.S. Pat.
, No. 790.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a data processing system and method that easily performs time-independent and time-based operations together.

Another object of the present invention is to provide such a system and method that facilitates processing of data by multiple concurrent processes.

Yet another object of the present invention is to provide a system and method for processing data using multiple data processors.

According to one feature of the invention, input means for receiving sequentially occurring data items to be processed.

processor means for simultaneously executing a plurality of processes to produce further sequentially occurring data items, each data item based on a source of each data item and a sequential position relative to each other received from the same source; means for associating a selected data item from said storage means with a unique index value determined during execution of the process; and data storage means for storing said data item. A data processing system is provided, comprising means for supplying selected data items to said processor means for use in execution of a process, and output means for supplying data items resulting from execution of said process.

According to another feature of the invention, input means for receiving sequentially occurring data items to be processed; processor means for simultaneously executing a plurality of processes to produce further sequentially occurring data items; means for associating the received data item and the further data item with each unique index value based on the data item's source and its sequential position relative to each other received from the same source; and means for providing selected data items from said storage means to said processor means for use in the execution of a process, said data items having an association determined during execution of said process. The data item is selected based on the index value, and all exchanges of data items between the processes are performed through the supply means, and further the selected data item obtained from the execution of the process is supplied. A data processing system is provided, comprising output means, wherein the data item is selected based on its associated index value.

According to still further features of the invention, receiving sequentially occurring data items to be processed, executing multiple processes simultaneously to create further sequentially occurring data items, and determining the source and source of each data item. associating the data items with each unique index value based on their sequential positions relative to each other received from the same source, storing said data items, and storing the selected stored data items in the process. A data processing method is provided comprising providing selected data items for use in executing a process based on associated index values determined during execution.

According to still further features of the invention, receiving sequentially occurring data items to be processed, executing multiple processes simultaneously to create further sequentially occurring data items, and determining the source and source of each data item. associating said received data item and said further data item with each unique index value based on their sequential positions relative to each other received from the same source, storing said data item, and storing said data item in selected storage; providing a data item for use in the execution of a process, said data item being selected based on an associated index value determined during the execution of that process, and in this way all exchanges of data items between processes. A data processing method is provided, further providing selected data items resulting from execution of the process as output of the data processing, the data items being selected based on associated index values.

Further objects and features of the present invention will become apparent from the following detailed description, taken in conjunction with the accompanying drawings.

A data processing system and method according to the present invention.

It is not limited to any particular type of data processing. However, they perform both time-independent functions, such as processing data obtained from measurement operations to derive further information, and time-based functions, such as real-time control of measuring equipment. It can be effectively used for simultaneous execution of processes that must be executed simultaneously. An example that includes both of these types of functionality is the operation of a borehole logging instrument as schematically illustrated in FIG.

To explain FIG. 1, an elongated logging device or sonde 1
0 is suspended from an armored conductor cable 12 in a borehole 14 penetrating a geological formation 16. The sonde 10 spins the cable 12 through a pulley 20 and a depth gauge 22 by a winch forming part of the surface equipment 24 and (
The cable is moved up and down within the borehole 14 by hoisting the cable (while actually performing logging measurements). Depth gauge 22 measures the displacement of cable 12 through pulley 20 and thus the depth of sonde 10 in borehole 14 .

Instrument 10 typically includes one or more transducers for measuring parameters of formation 16 by electrical, acoustic or nuclear techniques. These transducers are
Electrical signals related to the required parameters are generated, these signals are suitably conditioned by processing and interface circuits within the sonde 10, and transmitted via cable 123 upwards to the surface equipment 24. Device 24 includes a data processor 26 that typically receives, decodes, amplifies and transmits the depth signals generated by depth gauge 22 to a chart and/or magnetic tape recorder. Record as a function. Additionally, device 24 processes the data represented by these signals and generates indications of the necessary formation parameters (which are also recorded). By further processing these and other signals from the sonde 10, the surface device 24 can monitor the operation of the sonde 10 to, for example, synchronize the operation of the circuitry of the sonde 10 or change the gain of an amplifier. generates a signal to
Via the cable 12 it can be transmitted downward to control the sonde 10.

Details of the construction and operation of the sonde 10 and the equipment for communicating signals between the sonde 1o and the surface device 24 do not form part of the present invention.

It is believed to be obvious to those skilled in the art and need not be explained here. Similarly, specific algorithms for processing data received from the sonde 10 to derive values for formation parameters or for deriving control signals to be sent to the sonde 10 do not form part of the invention and are therefore does not explain.

Data processor 26 performs two different types of processing. The first type is essentially independent of the time at which it is executed. Therefore, when deriving the values of the necessary information parameters from the data received from the sonde 10, the same result does not always occur regardless of when this derivation is performed.

The second type of processing is time-based. Therefore, sensing the appearance of data items on cable 12 to be transmitted up the borehole by sonde 10 and ensuring that this data is captured and recorded relies on processor 26 monitoring cable 12 in a timely manner. . Similarly, the generation of command signals sent down the borehole to the sonde 10 must occur at a rate sufficient to properly operate the sonde 10 and sufficiently responsive to changing conditions. .

These two types of processing tend to conflict with each other. Accurately deriving parameter values requires significant processing time, can lead to delays and inaccuracies in the generation of command signals, and/or can result in failure to detect and record data sent up the borehole. There are things to do. Similarly, receiving data and generating command signals may interfere with evaluation of the received data by changing the data values such that additional calculations are required. In addition, even if Progera 11 continues to operate at the same speed as the IP, when migrating to another type of computer system, e.g. when using one or more new A system for operating a data processor that cannot continue to function properly after being modified to add program modules.A system for operating a data processor that can be easily and effectively and appropriately programmed to perform such tasks and alleviate these problems. and the method will be explained with reference to FIGS. 2 to 15.

In these figures, data processing is shown for convenience as being performed by multiple central processing units (CPUs) operating in parallel. It is assumed that each CPU is dedicated to only one task in all processing operations. However, the present invention also applies to cases where one or more CPUs are provided and their resources are commonly used among a number of different processes, each corresponding to a task performed by a dedicated CPU. The term concurrent processes, which should be understood to be equally applicable, shall hereinafter refer to two or more processes executed by any of these configurations. Such a process is itself quite complex, involving hundreds of instructions. If there are two or more CPUs, these may be distributed processing units connected by relatively low bandwidth data links;
Alternatively, it may be multiple processing units connected by relatively high bandwidth links (e.g., sharing the same main memory).

Similarly, the storage of data before and after processing by the CPU is shown as being divided among a number of separate storage devices for clarity. However, it is also possible to use one large storage device and divide its capacity appropriately for different data storage needs. Techniques for sharing processor and memory resources in this manner are well known and will not be described in detail here.

For example, FIGS. 2 and 3 relate to a configuration for calculating the roots of the following quadratic equation (ie, the value of X such that y=Q).

y==ax2+bx+c According to the law, the above formula becomes as follows.

-b+v'(b"-4a c) a However, a, b, and C are input as coefficients.

The number of CPUs, their inputs and outputs, and the configuration and connections of their associated storage devices are indicated where appropriate. Obviously, those details will vary depending on the particular application in which the processor is being installed.

Referring to FIG. 2, data processor 30 is shown as having three CPUs 32-36. Each CPU receives data items on one or more inputs, processes them, and then provides the data items on one or more outputs. To this end, the CPU includes well-known arithmetic logic circuits, registers, and instruction decoding circuits. It is further assumed that each CPU has its own memory for storing the program of instructions that control its operation and for storing the data processed in accordance with these instructions. It is clear that these CPUs can share common memory resources for these purposes.

Data processed by processor 30 (coefficients a, b
and the value of C) are received via input lines 40, 42 and 44, which lines are connected to storage devices 54-6
4 is connected to inputs 46, 48 and 5o of data storage 52, including 46, 48 and 5o. Each single input 66,68 and 70 of the storage devices 54,56 and 58 is connected to the inputs 46,48 and 50 for receiving input data to the processor 30. Each single human power of other storage devices 60 to 64 72.74
and 76 are used to store data obtained through processing by the CPUs 32 to 36, as explained below. For clarity of illustration, storage devices 54 to 6
The input line to 4 is shown in dotted lines to help distinguish it from other connections.

Data storage devices 54 to 64 may be one example.

A semiconductor random access memory, a magnetic disk memory, or a combination of both may be incorporated. As discussed below, a data storage device may not be able to write new data values to locations where data has already been written. Therefore, it is possible to incorporate a memory that is only capable of performing a single write operation at a given location.

An example of such a memory is a laser disk storage device, in which a laser beam is used to create a permanent pattern of microscopic pits on the surface of a metal or metallized plastic disk for later reading. Data is stored using

Such storage devices are available from, for example, Optimemof, Redondo Beach, California, USA.
5unnyvaLe, California and
0ptical DiskManufacturin
g Co., a division of Alcatel Thomson. Laser disk storage devices have the advantage of being able to store very large amounts of data in a relatively small physical space.

Each storage device 54-64 has its respective input 66-7.
6 and is configured to receive data and store this data as a stream of successive data items. Each further data item is added at the end of this stream after the data item already received. The length of the flow of data items is
Although in principle it could be extended indefinitely, it is clear that the finite physical storage capacity of the storage device imposes a constraint on the amount of data that can be stored at any one time.

However, each new data stream is treated as increasing the length of the stream.

The term data stream is used herein to refer to a discrete information entity that is to be accessed without regard to other items. Thus, depending on the application, a data item may be, for example, a single binary digit (1 bit), an 8-bit byte, a 16-bit or 32-bit word, a multibyte value representing a value in integer or decimal format, a sequence of alphabets. A sequence of bytes that encodes a character, or a full array of such values. The data item itself is
A value that serves as a pointer to the location of another data item in storage.

When a data item is added to a stream in storage, it is associated with a unique index value based on its position within the stream. The index value of each data item is one greater than the index value of the immediately preceding item in the same stream. In the simple case of Figure 1, where each stream has its own dedicated storage, this index value is
It is simply the number of data items previously received by the storage device (assuming the first item has index value 0). In these situations, no special action is required to assign a data item its index value.6 Furthermore, if the physical storage location in the storage device starts at address O, then the index value of the data item corresponds directly to the address of that location in storage. The storage device need only maintain an indication of the next free storage location into which to write the next received data item.

Rather than dedicating individual storage devices to each stream of data, multiple streams of data can also be stored on the same storage device. One possible solution is to treat the storage as partitioned into a number of segments, with each segment designated to one stream. In such cases, it is also necessary to maintain for each stream a record of the storage location containing the first data item of that stream.

Therefore, the storage location of another data item in the same stream can be identified by adding that item's index value to the address of the storage location containing the first item. Alternatively, data items from separate streams may be mixed. In this case, a table of storage locations associated with each stream could be maintained, or a linked list could be used, where each stored data item contains the location of the next data item in the same stream. The pointers pointing to are combined. These techniques for managing shared memory are known.

Storage devices 54-64 may incorporate memory to which new data can be written over existing data. If data items that are no longer needed can be identified and their locations reused, new data items can continue to be added to the stream indefinitely, but the amount of items in storage at any given time is It is clear that the capacity will never be exceeded. One idea by which such reuse can be accomplished is to limit the selection of data items to be retrieved from storage to data items received after the most recently retrieved data item. Therefore, assuming that numerically increasing index values are assigned to data items received by the storage device,
The index value of the item to be searched cannot be decreased. A location that contains a data item with an index value smaller than the index value of the last item retrieved from the storage device is a location that can be reused. Techniques for managing finite memory resources to achieve this result are known, for example the so-called "garbage collection".

Even if a new data item physically reuses the storage location of a previous, currently undesired data item, this new item is associated with its own unique index value and Index values for replaced items are not reused. Also, existing data items will not be changed. Replacing the value of a data item with another writes a new data item with its own unique index value and uses this new data item rather than the previous superseded item. It is done by

If a storage location is reused, it is necessary to maintain a record of the index value of the earliest data item still stored and the address of its physical location. The difference between this index value and the unique index values of other data items can be added to the address of the previous item to derive the address of the item.

If the storage device 52 incorporates a write-once memory such as the laser disk device, such reuse is not possible, and therefore no restrictions are placed on the index value of the data item to be searched. The index value for searching a data item can be increased or decreased arbitrarily. Although storage device 52 eventually becomes saturated, its very large storage capacity allows for a significant period of useful operating cycles before the storage medium must be replaced or expanded.

Each storage bag @54 to 64 has one or more outputs, and the devices 54 and 56 each have three output cuffs 8.80.82.
and 84.86.88, the device 58 has one output 9
o, the device 60 has two outputs 92 and 94,
Devices 62 and 64 each have one output 96 and 98. Each such output is configured to retrieve any data item in the stream stored in the associated storage by reference to the index value of the required item. If a data item is requested with an index value for which no data item has yet been received, the storage device waits until a data item for that index value is received and then returns the data item on request. It is composed of If the storage location is reused, the method for identifying the storage location for data retrieval is
The same method used to store data items is used, and the range of index values within which data can be retrieved is subject to the constraints described above.

Each of the storage devices 54, 56 and 60 is shown in FIG. 2 as having separate outputs that are also separate from the inputs for clarity of illustration. However, for example, in a magnetic disk storage device, a single read/write
It should be appreciated that input and output functions can be performed by appropriately controlling the write head. Similarly, when using semiconductor memories, these functions are performed by addressing the circuits in a known manner.

Input lines 4° to 44 by data processor 30
Each data item received via the storage device 54
to 58 are stored. All these data items (coefficients a, b and C) are used by the CPU 32, so that the three inputs 100 to 104 of the CPU are the outputs 8 of the storage devices 54, 56 and 58, respectively.
2.84 and 90. A dotted control line 106 extending from the CPU 32 to the output 82 causes the CPU 32 to specify the index value of the data item (value of coefficient A) retrieved from the storage device 54 and sent to the CPU by the output 82. be able to. Output 8
Similar control lines 108 and 110 to 4 and 90 allow CPU 32 to specify the desired data items (corresponding values of coefficients A and C) to be retrieved from stores 56 and 58. The CPU 32 is shown as generating a single stream of data items (intermediate result values known as radicals) which are supplied via the input cuff 2 to the storage device 6o.

With the output 92 of the storage device 60, the CPU 34 obtains a data item based on the specified index value via the control line 112, and also obtains the data item (corresponding values of coefficients a and b) from the storage table [t54 and 56]. and these are supplied by outputs 80 and 88 under control of control lines 114 and 116. The CPU 34 supplies the data item obtained by the process performed once through the input cuff 4 to the storage device 62. These data items are the values of one root of the quadratic equation for the corresponding values of coefficients a, b and C. CPU 36 similarly receives data items designated via control lines 118, 120 and 122 from storage devices 54, 56 and 60 via output ports 8, 86 and 94, and receives data items obtained from input ports 6. The data is supplied to the storage device 64 via the. These data items are the values of the other roots of the quadratic equation for these coefficient values.

Finally, the data items in storage devices 62 and 64 are provided by outputs 96 and 98 to output lines 124 and 126 of data processor 30 based on the index values specified via control lines 128 and 130.

The identification of the data stream to be retrieved may or may not be explicitly specified, depending on the configuration of the storage device and the format of the index value. Thus, in the configuration of FIG. 2, the required flow is uniquely identified by the physical selection of control lines 106 through 122, 128 or 130. When sharing memory between various streams, the stream identification can be encoded into the index values, giving each stream its own separate set of index values. Alternatively, the index value may be similar for each stream, and the identity of the stream may be provided as a parameter with the index value when a data item is requested.

The processes performed by each of CPUs 32-36 are summarized in FIG. 3 with particular reference to CPU 32. The operations of CPUs 34 and 36 are similar.

Referring to FIG. 3, the index values of the data items to be provided to CPtT 32 must be initialized when the process is first started.

This is represented by step 132.

CPU 32 then requests the data item of coefficient a from storage 54 based on the current index value at step 134 . In step 136, the CPU 32
checks whether the data item has been received. If not, the process repeats through step 136 until a data item is received. When a data item is received, the process continues to step 138, where the data item corresponding to the coefficient is requested from storage 56, and step 140 is repeated as necessary until the data item is received. This sequence of requests and repetitions is performed a third time in steps 142 and 144, and the data entry corresponding to coefficient C is obtained from storage 58.

When data items for all three coefficients are obtained, CPU3
2 proceeds to step 146, where the corresponding value of radical R is calculated based on the following equation.

R=v'(b''-4a c) This value is provided in step 148 as a data item to the storage device 6o, where the storage device corresponds to the data item retrieved in steps 134, 138 and 142. Receive index value.

The CPU 32 continues to step 150 and stores the storage device 54.
The index value to be specified when retrieving data items from 58 to 58 is increased, and then again step 1
34 and repeat the sequence for the next set of coefficients a, b and C.

Although the operation of CPUs 34 and 36 is very similar, the data item requested in step 142 is obtained from storage 60. In the case of the CPU 34, the calculation performed in step 146 is based on the following equation.

-b+R a The result is provided to storage 62 as a data item in step 148. In the case of CPU36,
The calculation performed in step 146 is based on the following equation.

b-R a The result is provided to storage 64 as a data item in step 148 .

As shown in FIG. 2, each process executed by any of the CPUs 32-36 provides data items to a storage device dedicated to storing data items from that process. Each process shown in Figure 2 has only one output, but of course
More than one output can be generated by a single process, in which case each output will have its own dedicated storage. Each process receives data items from one or more storage devices 54-60, but any one storage device may supply data items to more than one process. Therefore, the only means of communication between processes is through storage devices 54-60.

This feature has a significant effect. All data traffic to and from a process is buffered in storage by a stream of data items. Therefore, synchronization between processes is inherently provided by flows, and there is no need for explicit configuration for this in the processes themselves.

A process can add data items to a flow,
No need to wait for other processes to read it.

Each process can request data items as needed, and when a data item arrives, the process using that item is not forced to run. Because the index value is never reassigned to a new data item and the data item is never changed, the data item is always available to the process within the constraints imposed by finite storage capacity. Within this same constraint, a process may request any data item in the stream and is not limited to receiving items in a predetermined sequence. However, there is no risk that a process will receive data items out of order.

Further, in the embodiment shown in FIG. 2, the only means for connecting the inputs and outputs of all data processors 30 and the processes executed by CPUs 32 to 36 is a storage device in storage section 52 (i.e., storage device 54). .56.58.
62 and 64). Although this is desirable,
not important.

The process functions shown in FIGS. 2 and 3 are purely time-independent. In particular, storage devices 54 to 6
The value obtained when searching for a data item from 4 is independent of the time of occurrence of the item being searched for. When implementing time-independent functions, additional search functions are involved, as shown in FIGS. 4 and 5.

FIG. 4 shows a storage device 150 comprising further circuitry to enable retrieval of the most recently received data item in the stored stream, in this case the storage device 150. The index value of an item is not specified by the CPU or process that requires the item, control line 152 from the cPU register 15
4, this register is updated with the index value of the most recently received data item from storage 150 via connection 156 as each data item is received via input line 158. Register 154 is then connected to output 160 of storage device 150. Thus, when a request for a data item is received via line 152, this is set by register 154 to the index value of the most recently received item;
is sent to output 160 along with the value currently held in . The received data item is provided via output line 162.

FIG. 5 shows the storage device i! 164, the storage device has further circuitry capable of attempting to retrieve data items. In this case.

A predetermined limit is imposed on the amount of time the storage device waits if there is no data item with a specified index value. Control line 166 connects output 168 of storage device 164 to indicate the index value of the data item to be retrieved.
and the trigger input T of the monostable circuit 170.
is also connected to.

Output line 172 provides the retrieved data item and is connected to the clear input of monostable circuit 170. The set output of monostable circuit 170 is connected to time-out signal line 174.
and output 168. When an item of data is requested via control line 166, the request is sent to storage 164 and at the same time monostable circuit 170 is triggered and enters its quasi-stable state for a predetermined period of time. If the requested data item is in storage 164 or is received before the predetermined time has elapsed, it is provided via output line 172 and is output to monostable circuit 170 before reaching the end of the pseudo-stable time. Clear. However, if the requested data item is not available before the end of the predetermined time, the monostable exit v! 1170 times out and generates a signal on timeout line 174. This signal indicates to the requesting CPU or process that time has expired and cancels the outstanding request at output 168 of storage 164. If desired, the monostable circuit can have a variable pseudo-stable time determined based on a value specified on control line 166 by the CPU issuing the request when a data item is requested. The design of circuits to perform such functions will be apparent to those skilled in the art.

A second data processor according to the invention will be described with reference to FIG. The processor uses the search operations described with respect to FIGS. Perform time-based functions in the form of data collection. This sensor generates a data item once every 2 seconds, and the rate is 1
A signal is transmitted every /10 seconds. The processor.

It checks whether data items continue to arrive from the sensor and terminates sensor operation if the data is interrupted for a time that exceeds a preset limit.

Referring to FIG. 6, data items from the sensors are received via line 176 by a processor, indicated generally by reference numeral 175, and provided to data storage 178 of data storage section 180. A data item representing the speed of the sensor is received on line 182 and provided to another data storage device 184. The storage device has an output 186 configured to provide a refresh function as described with respect to FIG. This output displays the latest data item representing the speed of the sensor.
PU188, this CPU.

It also receives data items from the output 190 of the storage device 178 and stores the data items obtained through the processing in the storage device 19.
Supply to 2. Another output 194 from storage 178 with a time-out function as described with respect to FIG.
Rd supply 1, G (7') Out: 'J If 3Q f
The outputs 200 and 202 of the storage devices 192 and 198 are connected to the third CPU 20
4, which supplies data items to storage table gi206. The fourth and final CPU 208 is the storage device 17
8 via its output 210 and provides the data items to a storage device 212 . Storage devices 206 and 212 provide data items as output signals from data processor 175 via outputs 214 and 216 and output lines 218 and 220.

The operation of CPU 188 is shown in the flowchart of FIG. When the operation of processor 175 is started,
The index value of the data item to be retrieved from storage devices 178 and 184 is initialized in step 222. CPU 188 then enters a loop beginning at step 224 where sensor data items are requested from storage 178. Since the sensor only supplies a data item once every two seconds, CPU1
Initially, there is no data item at the position corresponding to the index value specified by 1 by 8R. Therefore, when CPU 188 proceeds to step 226 to determine whether a data item has been received from storage device 178, the result will initially be negative.

Therefore, CPU 188 proceeds back to step 226 and continues operation until a data item is received.

Once the data item is received, CPU 188 proceeds to step 228 and requests the latest speed data item from storage 184. Since velocity data items are received 10 times per second, even when the first request is made, the storage 18
4 already has a data item. Therefore, the latest values are always available and requests are met immediately. CPU 188 continues to step 230 and determines whether the speed of the sensor needs to be adjusted. This determination may be made on the basis of any suitable algorithm, which does not form part of the present invention and will not be described here. If no change is necessary, the CPU 188 proceeds to step 232.
to repeat the loop starting at step 224, where the index value used to request data items from storage 178 is incremented. If a change is required, the appropriate cloak is sent to storage 192 in step 234 and the CPU proceeds to step 232 as described above.

The operation of the CPU 196 is shown in the flowchart of FIG. When processor 175 begins operation, index values are initialized in step 236. The process then proceeds to step 238, where data items are requested from storage 178 at the specified time limit. 7th
Similar to the CPU 188 retrieving the data item in step 224 of the figure, there is an interval between obtaining the data item in response to the request. During this interval, the CPU
96 passes through two decision steps 240 and 242.

In the first of these steps, CPU 196 checks whether a data value has been received. If so, CPU 196 proceeds to step 244 and increments the index value to retrieve from storage 178 and then returns to step 238.

If the result of step 240 is negative, CPU1
96 proceeds to step 242 and outputs the output 1 of the storage device 178.
Check if a timeout signal is received from 94. If such a signal does not exist, it is based on the time limit for interruption of data from the sensor,
Therefore, CPU 196 performs decision steps 240 and 242.
repeat. However, if a timeout signal is not received, CPU 196 proceeds to step 246 and provides a command to memory 198 to terminate sensor operation.

As shown in FIG.
Request data items from both. Then step 2
At 52, a check is made to see if any items have been received from storage 192. If so, in step 254 the received data item is sent to storage device 206 and in step 256
resume requesting data items from 92, step 252
Return to. If not, proceed to step 258 and check for receipt of the data item from storage device 198. If there is no data item, step 252
The check loop is repeated by returning to . Otherwise, the received data item is sent to the second storage device 206 in step 260 and the request for the data item from the storage device 198 is resumed in step 262.

Return to step 252. The action of the process carried out by CPU 204 is to merge the commands generated by CPUs 188 and 196 in the order in which they were generated and transmit them to storage 206 for provision on line 218 as the output of processor 175. be. This process is only one example of a time-based process, as the exact sequence of commands that appear in storage 206 will vary with the conditions to which the sensor is subjected (eg, wellbore conditions) and the resulting effects.

Another way to merge two sets of commands in a time-based manner is to send two inputs to a storage device and add each series of commands to the flow in the storage device through each input.

CPU 208 simply retrieves the sensor data items from storage 178 and performs appropriate processing on these data items to form the desired information, which is then output to storage 212.

This processing can take various forms such as arithmetic operations and combinations with other data. The details of such processing are not relevant to the present invention and need not be described in detail.

As with processor 30 of FIG. 2, all data communications between processes being executed by the CPU are
This is achieved by a stream of data items held in storage, each data item being associated with a unique index value in the stream.

When using separate CPUs for each process, this configuration provides the following effects. That is, there are two types of operations: time-independent operations (e.g., processing of sensor data by CPU 208) and time-based operations (e.g., processing of sensor data by CPU 208).
8-8) can be performed in separate processes without incurring possible mutual adverse effects between the processes. This makes it efficient and
It is possible to facilitate the design and writing of effective and reliable programs and their processing. Furthermore, there may be no need to consider the specific details of the hardware on which the program will run. Similarly, no special configuration is required for the possibility of subsequent changes to the hardware, the use of different hardware, or changes to the program executing the process.

Therefore, data processing by the CPU 208 is performed by the CPU 208.
Depending on the complexity and speed of the 08 operation, it can proceed at a reasonable speed without fear that pending sensor data items will be lost or overwritten by execution of other processes.

These data items are made available in storage 178 and sent to output 210 when requested by CPU 208, whether or not they have already been retrieved by CPUs 188 and 196 via outputs 190 and 194.
Searched through. Similarly, output data obtained by such processing can be generated by CPU 208 at a rate independent of the rate at which the data is extracted via output line 218 without risk of losing data before extraction.

A time-based process executed by CPU 188 may be executed to monitor and control the speed of the sensor at a rate deemed desirable, and accordingly transfer data items from storage 178 via output 190. can be extracted.

This process functions independently of the performance of data processing functions in CPU 208, and does not introduce delays or timing issues.

Therefore, there is no need to consider timing conflicts when programming a process.

Even when processes are executed with a single CPU shared among multiple processes, there is no risk that the quality of data required by time-independent processes will be degraded by the execution of time-based processes. Analysis of the behavior of such processors by partitioning processes such that communication between processes occurs only through a flow of data items, at the risk of delays for time-based processes; Identification and reconciliation of potential conflicts is facilitated. Additionally, this architecture is believed to provide an effective structure for implementing techniques that can account for timing constraints and avoid timing conflicts.

Additionally, the present invention facilitates system designs that can be implemented initially with a single CPU and migrated to multi-CPU systems without significant changes.

Although it is not necessary to separate time-based and time-insensitive functions between various processes, it is generally considered good practice to do so, reducing timing conflicts and the difficulty of subsequent changes. Helpful.

Preferably, all communication between processor 175 and the outside world is by a further stream of data items, as shown in FIG. However, if desired, data may be exchanged directly with other devices in one or more processes. Such exchanges are typically time-based, ie.

The result depends, for example, on execution speed.

Processors 30 and 175 of FIGS. 2 and 6 have individual connections between each CPU and storage for communicating data items. Alternatively, as shown in Figure 10,
Such connections can be multiplexed. Referring to FIG. 10, the storage section 270 has six storage devices 272-282, each having a respective input 284-294 and a respective single output 296-306. are doing. Each of these inputs and outputs is connected to an input/output circuit 308 and to a respective control line 310-320 associated with each output 296-306. The three CPUs 322-326 also have input/output lines 340, output lines 342, and their associated control lines 344 connected to the input/output circuit 308 by respective control lines 328-332 and three-way data lines 334-338. Connected to circuit 308.

In operation, CPUs 322-326 control line 32.
Data items are obtained from storage devices 272-282 by sending signals to input/output circuitry 308 via 8-332. These signals identify the storage device from which the data item was requested and the index value for the requested item. Input/output circuit 308 sends requests containing index values to appropriate control lines 310-320.
The data is then supplied to the thus identified storage device. The requested data item is stored in each output 2 by the storage device.
96-306 back to the input/output circuit 308, which sends it on the appropriate data lines 334-338 to the CPU that issued the request. CPU32
Data items generated by 0 through 326 are input via a respective one of data lines 334 through 338, with corresponding control line signals 334 through 332 identifying which of storage devices 272 through 282 should receive the data item. / is supplied to the output circuit 308. This human power/
Output circuit 308 inputs data items 284 to 29
4, the data item is sent to the storage device thus identified.

Similarly, input/output circuit 308 transfers input data received via input line 340 to storage devices 272-28.
2 and provides data items to output line 344 based on requests sent via control line 344.

As shown in FIG. 10, each storage device 272 to 282
outputs one output 296 to 3, regardless of how many CPUs request data items from one storage device.
It is sufficient to have only 06. Input/output circuit 308 provides information to each storage device 272-282 about the length of the flow of data items in that storage device and the most recently retrieved data for each CPU requesting data items from that storage device. It is carried out to maintain a record of the item's index value. If storage unit 270 shares memory between multiple streams of data items as described above, input/output circuitry 308 may maintain a record of the physical location of the data items that make up each stream. maintain. Similarly, when using read/write memory, the input/output circuitry controls the reuse of a storage location when the data item contained therein becomes undesirable for further processing. do. It will be readily apparent to those skilled in the art that known techniques for managing memory resources may be used to perform these functions.

The input/output circuit comprises, in one possible embodiment, two blocks of random access memory and two registers. One register is used to communicate control and status information between the input/output circuits and the CPU, and the other register is used to transfer data items and index values. The first block of random access memory is organized as a set of tables, each table having 62 pairs of 16-bit inputs and two 32-bit inputs, with up to 62 pairs of data items from one storage device. Used to hold details of pending requests. One of the 32-bit inputs is used to store the index value of the most recently added data item to the stream of data items in that storage device. The other 32-bit input stores the physical address of the storage location containing the data item. The first entry in each pair of 16-bit inputs contains the offset from the index value of the most recently added data item to the index value of the data item requested by CPtJ. The second input of the pair is the C requesting the data item.
Identify the PU or process. 1 configured like this
Megabytes of random access memory serve to form up to 62 outputs for each of the 4096 storage devices.

The second block of random access memory is organized in 1024 groups of 4 people each, and the CPU
or as a first-in, first-out buffer for data items retrieved upon request from a process. The first input of each group is a 16-bit value that identifies the CPU or process issuing the request. The second input is a 12 bit value identifying the storage device or stream for which the value is requested.

The third input is used to store the data item itself, in this case 32 bits long. The fourth input is a 4-bit state value. In this configuration, the buffer requires 8 kilobytes of random access memory. The storage itself can be implemented to store up to 512 data items, each 32 bits in length, for each of the 4096 streams, requiring another 8 megabytes of random access memory. .

The CPU performs a write operation by placing the next data item to be stored in the data transfer register of the input/output circuit and placing the identification of the destination stream in the control and status register along with a code indicating that a write operation is requested. be able to.

The input/output circuit retrieves the value of the data transfer register and associates it with the stream of storage data items identified in the control and identification registers.The input/output circuitry then associates the pending request table for that stream with the most recently added data item. The data item is updated by incrementing the index value of the data item and the input for the physical location of that item. Additionally, the offset entry for the requested data item in each of the 62 pairs of entries of this same table is reduced 71
1. If offseller1 becomes zero and pairwise1 or process identifier is non-zero, then the data item just added has the index value of the requested data item. In this case, the data item, process identifier and flow identification are added to the buffer of retrieved data items;
The process identifier in the pending request table is set to zero to indicate that the request has been fulfilled.

A CPtJ or process requests a data item from the input/output circuit by placing the identification of the stream, including the item and its own identity, in the control and status registers and the index value of the requested item in the data transfer register. An unused pair of entries in the pending request table for an identified flow is updated with the item's offset from the index value of the most recently added item to that flow and the identity of the process issuing the request. The next data item to be retrieved in the buffer is placed in the data transfer register, and the process and flow identifiers at the relevant inputs of the buffer are placed in the control and status 1 registers.
IL. In this way, the identified process can retrieve the data item from the input/output circuit. 1
If more than one CPU is shared by processes, the process that added the request to the table of outstanding requests is suspended, as described below, and the process identified in the control and status registers is restarted.

Note that in this configuration, a pending request for a data item can be canceled by setting the process identification entry of the pending request table to zero.

Apart from using specialized hardware.

A data processor according to the invention may also be implemented by suitably programming a general purpose computer to perform the storage and processing functions described above. Figures 11-15 are flowcharts illustrating, by way of example, how to perform the various read and write functions described above.

These figures assume that basic memory management and process execution operations exist, such as writing and reading from memory, garbage collection, suspending and reactivating process execution. These operations are performed using, for example, the Interlisp-D programming environment available from Xerox Corporation, Pasadena, California, USA.
rammingenvironment), as described in the Interlisp Reference 1 Manual, published by Xerox Corporation in October 1983.
sp Reference Manual) J.

Each function is described in the form of a subroutine that can be called by the process that needs to perform it. When such a call occurs, control is temporarily transferred to the subroutine by the calling process. Thus, if a subroutine contains steps that suspend execution until reactivated by another event, execution of the calling process is also suspended.

FIG. 11 shows the procedure for reading a data item from a stream for a specified index value in response to a request from a process. The procedure begins at 350 and first checks at step 352 whether a data stream is available for the requested index value.

This can be done, for example, by comparing the index value to the current length of the stream. If the data item exists, the procedure proceeds directly to step 354, where the data item is obtained from memory and provided to the requesting process, and then the procedure continues to step 356.
Returns to the calling process. If the requested data item is not available (ie, a data item for that index value has not yet been received), the procedure proceeds to step 358. In this step,
The identity of the process making the request and the index value are
A data item is added to the table associated with the stream to hold details of requests pending against the stream. This table essentially consists of a group of memory locations reserved for this purpose, e.g. They are similar.

The procedure then proceeds to step 360, where execution of the procedure is suspended until the data item is added to the flow (e.g.,
(using the process suspension behavior of Interlisp-D above)
, once the data item has been added, execution of the procedure resumes, as described below.

The procedure then proceeds to step 354 as described above.

The procedure for writing data items into a flow is shown in FIG. Referring to FIG. 12, the procedure is entered at step 362 and begins at step 364, where the data stream is stored in memory. This is performed in Interlisp-O by the C0N5 operation. Next step 366
At , a counter indicating the length of the stream of data items is incremented by one, and then, at step 368, the table of pending requests for that stream is checked for requests for an index value equal to the length of the new stream. be done. If there is no such request, the procedure proceeds directly to step 3'70, where the earliest data item read from the stream is identified, and then proceeds to step 372, where the earliest index value is determined. All data items that have are deleted. This step can be performed in Interlisp-D by the RPLACD operation. These two garbage collection stages de-reuse memory locations occupied by undesired data items, making them suitable when read/write memory of limited capacity is used. It will be appreciated that these can be omitted if write-once memory is used. After step 372, the procedure returns to the calling process at 374.

If one or more pending requests are found in step 368, the process immediately proceeds to step 376, where the input for each found request is examined. If this includes a reference to a timer function, the timer function is canceled, as discussed below with respect to FIG. Similarly, if this involves identifying another stream with pending requests, the first
As discussed below with respect to Figure 4, the pending requests for that flow are removed. The entry for each request found in step 368 is then removed from the table. Then, continuing to step 378, the process or processes that requested the data item with that index value are resumed (
In Interlisp-D, WAKE, PROCES
This is done with the S operation). The procedure then exits garbage collection at steps 370 and 372 and continues to step 374.

Referring to FIG. 13, step 380 enters a procedure for performing a read function that includes a timeout. In step 382, a check is made as to whether a data item is already available for the requested index value. If so, the data item is obtained from memory in step 384 and sent to the requesting process, while the procedure returns to step 386. Otherwise, the index value and the identity of the process issuing the request are added to the pending request table for that flow in step 388. In this regard, the procedure is the same as steps 350-358 of the normal reading procedure shown in FIG. However, in this case the timer function is also set during the required time period before generating the time-out signal. That identity is added to the pending request table, and another table is combined with a timer to determine the time interval that must elapse before a time-out signal is generated, the identity of the process issuing the request, and the data item. The identity of the requested stream is indicated. In step 390, a timer is started and in step 392 the procedure is suspended. Thereafter, when the procedure is restarted, a check is made in step 394 to see if the timer has generated a signal indicating that the specified time interval has elapsed, thereby resuming operation. If so, the request added to the pending request table in step 388 is removed in step 395, a time-out signal is generated in step 396, and the procedure returns to the calling process in step 398. Otherwise, restart must occur by receiving the data item for the requested index value, in which case the procedure proceeds to step 384 to obtain and provide the data item as described above.

FIG. 14 is a flowchart illustrating the execution of a read that includes an arbitration function that takes data items from multiple streams and combines them based on a predetermined arbitration algorithm.

This is 1, for example, the data processor 17 shown in FIG.
Memory'! The flow of data items in A locations 192 and 198 are merged.

Referring to FIG. 14, the procedure is entered in step 400 and begins in step 402, attempting to obtain data items from each of the streams identified by the requesting process. Step 40
4, it is checked whether the data item has been found.

If so, the procedure proceeds to step 406 in which one data item is selected based on a predetermined arbitration function, e.g., prioritizing data items from one stream over another. and can be selected. In this procedure, the selected data item is provided to the calling process at step 408 and control is transferred to it at 410. If the data item is not available in step 402, the procedure continues to step 412;
The index value and identification of each calling process are added to each associated flow's reservation table.

Additionally, each table entry includes the identification of each other stream for which a request is pending. The procedure then suspends operation in step 41.1 until the requested data item is obtained and the procedure resumes as described above. The procedure then returns to step 11Q2 to obtain the then available data item and continues to step 406 as described above.

Figure 15 shows the procedure for reading the latest value received via the flow. The procedure is entered in step 41G and checked in step 418 to see if the stream length is zero. If not, a data item for an index value equal to the length of the stream is obtained in step 420 and sent to the calling process before control is returned in step 422. Otherwise, after an indication is sent in step 424 that the data item is not available, step 4
Control is returned at 26.

Other reading functions can be provided. One idea is to read the next data item following the most recently obtained data item through the same output. This simply requires that a record of the index value at the time the last data item was provided be kept at the output in storage. Another idea is a move command that specifies the index value where the output should be set without the data item being read. This is useful when preparing to use the next item's read operation.

A set of data items having adjacent index values in the same data item stream may be requested by determining the index value of the first item in the set and the number of items required. Yes, but no data items are provided until all items in the set are obtained. Data items can also be requested from many different streams. However, in this case, it is assumed that the actual index value for each flow has a predetermined relationship (which differs for each flow) with respect to a common reference index value.

This facilitates, for example, retrieval of corresponding data items generated simultaneously from different sensors.

The data processor of the present invention has fewer CPUs than processes.
When implemented on a computer, especially when implemented on a single CPU, a supervisor or scheduler program is required to coordinate the allocation of CPU resources among the various processes. Such schedule programs simply maintain a first list of processes that are ready to run and a second list of processes that are held pending waiting for data items. Each process is assigned a priority rating, which allows the scheduling program to determine which processes in the first list should be executed in preference to other processes in that list. When a process requests a data item, it does so by invoking a read procedure as described above. If a data item is available, it is provided immediately and execution of the process continues. Otherwise, the reader 7 suspends its own execution and therefore suspends the calling process. When a process is suspended, it is moved from the first list to the second list. When restarted, the process is returned to the first list. When the running process moves to the second list, the schedule program scans the first list and
Identify the next process to run based on priority rating. If two or more CPUs are used, each
U executes a process and upon its completion begins execution of the highest priority process waiting in the first list of processes.

Although nine specific embodiments of the present invention have been described above in detail regarding the data processing method and system according to the present invention, the present invention is not limited thereto. The present invention has also been described with respect to processing data items that occur sequentially or serially. With appropriate buffering techniques, in parallel,
It is clear that even the data items received can easily be converted into serial ones. Also, the sequence does not necessarily have to be a time-based sequence. Although all processes are shown as receiving a data item ``1'' from at least one stream, they do not receive data items; they only generate data items, such as time and date information or waveform information. It is also possible to think of a process that provides an indication.

Therefore, it will be apparent to those skilled in the art that various modifications may be made to the invention within the spirit and scope of the claims.

[Brief explanation of the drawing]

FIG. 1 is a schematic diagram showing a logging operation using a borehole in which the present invention can be conveniently utilized; FIG. 2 is a block diagram showing a first embodiment of a data processor according to the present invention; FIG. FIG. 4 is a block diagram of a storage device used in implementing the present invention; FIG. 5 is a block diagram of a storage device used in implementing the present invention; FIG. FIG. 6 is a block diagram showing a second embodiment of a data processor according to the invention; FIGS. 7 to 9 are block diagrams of a central processing unit forming part of the data processor of FIG. 10 is a block diagram illustrating an alternative implementation of the data processor shown in FIGS. 2 and 6; and FIGS. 11-15 are flowcharts illustrating the operation of the unit; FIG. 1 is a flowchart illustrating procedures useful in implementing the data processor of the present invention. DESCRIPTION OF SYMBOLS 10...Logging equipment or sonde 12...Sheathed wire cable 14...Drilling hole 16...Geological formation 20...Pulley 22...Depth gauge 24...
Surface equipment 26...Data processor 30...Data processors 32-36...CPU 4o, 42.44...Input line 52...Data storage section 54-64...Storage device Drawing cleaning (No change in content) 162 156 Fig・4 Fig, 8 Fig, 9 σ398 1. Indication of case Patent application No. 193865 of 1985 2, title of invention Data processing system and method 3
, Relationship with the case of the person making the amendment: Applicant 4, engraving of the drawing originally attached to the representative application, as shown in the attached sheet (changed in content 4L)

Claims (18)

[Claims] (1) input means for receiving sequentially occurring data items to be processed; processor means for executing multiple processes simultaneously to produce further sequentially occurring data items; and the source of each data item and the same. means for associating data items with each unique index value based on sequential positions relative to each other received from a source; data storage means for storing said data items; and selected data items from said storage means. means for supplying to said processor means for use in executing a process, said data item having an index value corresponding to the number of data items received from the source of said item; and output means for supplying the obtained data items. (2) input means for receiving sequentially occurring data items to be processed; processor means for simultaneously executing a plurality of processes to produce further sequentially occurring data items; and a source of each data item and the same. means for associating data items with each unique index value based on sequential positions relative to each other received from a source; data storage means for storing said data items; and selected data items from said storage means. If the data item is obtained within the specified time period, this data item is provided to the processor means for use in executing the process; and if the data item is not obtained within the specified time period, the data item is obtained. means for providing an indication that the data item is selected based on an associated index value determined during execution of the process; A data processing system characterized by comprising an output means for supplying. (3) A system according to claim 1 or 2, wherein the data storage means is configured to prevent modification of stored data items. (4) The system according to claim 3, wherein the data storage means is a one-time write type storage device. 5. A system as claimed in any one of the preceding claims, wherein said data items are provided to said processor means solely through said storage means based on associated index values. (6) A system according to any of the preceding claims, wherein the supply means is arranged to supply data items based on index values that vary in only one direction. 7. A system according to any preceding claim, wherein the processor means comprises a single processing unit configured to execute the process on a time-sharing basis. (8) The system according to any one of claims 1 to 6, wherein the processor means includes a plurality of processing units. (9) The processing units are smaller in number than the number of processes, and the processing units are configured to execute the processes based on a predetermined priority assigned to each process. The system described in Section. (10) Receive sequential data items to be processed, run multiple processes simultaneously to create further sequential data items, and process the source of each data item and each other received from the same source. associate a data item with each unique index value based on its sequential position with respect to the item, store said data item, and store the selected stored data item in accordance with the number of data items received from the source of the item. A data processing method characterized in that data items having corresponding index values are provided for use in the execution of a process. (11) Receive a sequentially occurring data item to be processed, and create another sequentially occurring data item by executing multiple processes simultaneously. associating data items with each unique index value based on the source of each data item and its sequential position relative to each other received from the same source; storing said data items; and storing the selected stored data items. providing this data item for use in the execution of the process if it is obtained within a particular time period, and providing an indication that this data item is not available if said data item is not obtained within said time period; A data processing method, characterized in that the data items are selected based on associated index values determined during execution of the process. (12) A method according to claim 10 or 11, in which modification of stored data items is prevented. (13) The method according to claim 12, wherein the data item is stored in a write-once storage device. 14. The method of claims 10-13, wherein only stored data items are provided for use in executing a process based on their associated index values. (15) The data item is provided based on an index value that changes only in one direction.
The method described in Items 1 to 14. (16) The above processing is executed on a time-sharing basis by one processing unit.
The method described in Section 5. (17) The method according to any one of claims 10 to 15, wherein the process is executed by a plurality of processing units. (18) The method of claim 17, wherein the processing units are fewer than the number of processes, and the processes are executed based on a predetermined priority assigned to each of them.