US5241652A
Expired / Lapsed
US
⏱️ 0.0 years remaining
System for performing rule partitioning in a rete network
Assignee: Hewlett Packard Development Co LP
•
Filing: 1992-03-31
•
Granted: 1993-08-31
•
Expires: 2012-03-31
Abstract
Abstract A rule-partitioning system for converting at least a portion of a target expert system program to a rule partitioned RETE network for execution on multiple processors, including a rule partitioning portion for assigning different rules of the target expert system program to different partitions on the basis of previously collected processing statistics and on the use of node sharing; and a compiler for converting the target expert system program to the RETE network, wherein the rules of the RETE network are assigned to the multiple processors in accordance with the partition assignments.
Inventors
- 👤 William Barabash
- 👤 William S. Yerazunis
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-06 06:18
CPC Classifications
G; G06; G06N; G06N5/00; G06N5/04; G06N5/046; G06N5/047; G06F; G06F9/00; G06F9/06; G06F9/44; G06F9/448; G06F9/4494
Claims Analysis
Total Claims: 51
Claim 1
Independent Claim
313 words
A system for building a reticular discrimination network (RETE network) capable of being partitioned across a set of multiple, interconnected computer processors that comprise a computer system, from at least a portion of a target expert system, the target expert system comprising a plurality of rules each of which comprise a plurality of conditions grouped in a sequence by a set of logical connectors and a result to occur when the conditions of the rule are satisfied, certain ones of the plurality of rules having conditions that are equivalent to conditions in other ones of the plurality of rules, each of the conditions in a rule being represented by a test node in the reticular discrimination network, the test nodes being connected through associated join nodes representing the logical connectors of the rule, the system for building a RETE network comprising: a. a memory configured to store the target expert system; b. a partitioner, coupled to the memory and configured to read each condition of each rule of the at least a portion of the target expert system, and partition the rules by associating each rule into one of a preselected set of partitions and storing partition information relating to the partitions in the memory, each partition representing one of the multiple, interconnected computer processors comprising the computer system, the partitioner including means to identify conditions shared by more than one rule and means to increase the likelihood that a first and second rule will be associated into a same one of the preselected set of partitions when the first and second rules share any of certain ones of the plurality of conditions; and c. a compiler configured to convert the rules of the at least a portion of the target expert system into a RETE network, each partitioned rule being compiled to execute on a corresponding processor based on the partition information.
Claim 2
Independent Claim
313 words
A system for building a reticular discrimination network (RETE network) capable of being partitioned across a set of multiple, interconnected computer processors that comprise a computer system, from at least a portion of a target expert system, the target expert system comprising a plurality of rules each of which comprise a plurality of conditions grouped in a sequence by a set of logical connectors and a result to occur when the conditions of the rule are satisfied, certain ones of the plurality of rules having conditions that are equivalent to conditions in other ones of the plurality of rules, each of the conditions in a rule being represented by a test node in the reticular discrimination network, the test nodes being connected through associated join nodes representing the logical connectors of the rule, the system for building a RETE network comprising: a. a memory configured to store the target expert system; b. a partitioner, coupled to the memory and configured to read each condition of each rule of the at least a portion of the target expert system, and partition the rules by associating each rule into one of a preselected set of partitions and storing partition information relating to the partitions in the memory, each partition representing one of the multiple, interconnected computer processors comprising the computer system, the partitioner including means to identify conditions shared by more than one rule and means to increase the likelihood that a first and second rule will be associated into a same one of the preselected set of partitions when the first and second rules share any of certain ones of the plurality of conditions; and c. a compiler configured to convert the rules of the at least a portion of the target expert system into a RETE network, each partitioned rule being compiled to execute on a corresponding processor based on the partition information.
Claim 3
Dependent Claim
85 words
In the system as defined in claim 1 the partitioner comprises: a. an iteration module coupled to the memory and configured to read a rule and temporarily assign the rule to each of the different partitions; b. an evaluation module for estimating a performance measure for each such temporary assignment of the rule, the performance measure being derived from predetermined processing statistics and condition sharing information; and c. an assignment module for permanently assigning the rule to the partition which yields the best performance measure.
Claim 4
Dependent Claim
40 words
In the system as defined in claim 2 the iteration module operates to temporarily assign the rules in an order by selecting for temporary assignment an unassigned rule which has a longest processing time according to the predetermined processing statistics.
Claim 5
Dependent Claim
17 words
The system of claim 2 further comprising a statistics generating module for generating the predetermined processing statistics.
Claim 6
Dependent Claim
50 words
In the system of claim 4 the evaluation module comprises means for calculating, as a function of the predetermined processing statistics and the condition sharing information, the time for the multiple processors to complete at least one recognize/act cycle for the target expert system with the rule as temporarily assigned.
Claim 7
Dependent Claim
45 words
In the system of claim 5 the means for calculating the time to complete at least one recognize/act cycle for the target expert system with the rule as temporarily assigned comprises means for computing a partition process time for each partition for each recognize/act cycle.
Claim 8
Dependent Claim
63 words
In the system of claim 6 the means for summing the number of activations in each cycle of each of the nodes representing the rules that are assigned to that partition comprises a means for including the number of activations of nodes that are part of conditions that are shared by more than one rule assigned to that partition in sums only once.
Claim 9
Dependent Claim
40 words
In the system of claim 5 the statistics generating module comprises means for generating the numbers of activations per recognize/act cycle for blocks of test and join nodes representing sequences of conditions shared by rules in the target expert system.
Claim 10
Dependent Claim
45 words
In the system of claim 8 the means for calculating the time to complete at least one recognize/act cycle for the target expert system with the rule as temporarily assigned comprises means for computing a partition process time for each partition for each recognize/act cycle.
Claim 11
Dependent Claim
46 words
In the system of claim 9 the means for computing the partition process time for each partition for each recognize/act cycle comprises means for summing the number of activations in each cycle of each of the blocks representing the rules that are assigned to that partition.
Claim 12
Dependent Claim
35 words
In the system of claim 10, the means for summing comprises means for including the number of activations of blocks that are shared by more than one rule in the partition in sums only once.
Claim 13
Dependent Claim
50 words
In the system of claim 6 the means for calculating the time to complete at least one recognize/act cycle for the target expert system with the rule as temporarily assigned comprises means for identifying the largest partition process time for each recognize/act cycle and summing the identified largest process times.
Claim 14
Dependent Claim
37 words
In the system of claim 6, the statistics generating module comprises means for generating the number of activations of conditions per recognize/act cycle for rules of the target expert system from runs of the target expert system.
Claim 15
Dependent Claim
46 words
In the system of claim 13, the means for computing the partition process time for each partition for each recognize/act cycle comprises means for summing the number of activations in each cycle of each of the nodes representing the rules that are assigned to that partition.
Claim 16
Dependent Claim
40 words
In the system of claim 1, the partitioner comprises means for enhancing the likelihood that the first and second rules will be associated into the same partition whenever the first and second rules share any of the plurality of conditions.
Claim 17
Independent Claim
382 words
A system for building a reticular discrimination network (RETE network) capable of being partitioned across a set of multiple, interconnected computer processors that comprise a computer system, from at least a portion of a target expert system program, the target expert system program comprising a plurality of rules, each of which comprise a plurality of conditions grouped in a sequence by a set of logical connectors and a result to occur when the conditions of the rule are satisfied, certain ones of the plurality of rules having conditions that are equivalent to conditions in other ones of the plurality of rules, each of the conditions in a rule being represented by a test node in the RETE network, the test nodes being connected through associated join nodes representing the logical connectors of the rule, the system for building a RETE network comprising: a. a memory configured to store the target expert system program; b. a uniprocessor compiler coupled to the memory and configured to compile the target expert system program to form a RETE network for execution on a single processor and store the compiled target expert system program in the memory, certain ones of the conditions represented by nodes of the uniprocessor RETE network being shared by more than one rule; c. a processor, coupled to the memory for executing the target expert system program, to obtain processing statistics in connection with each node of the uniprocessor RETE network during execution; d. a partitioner, coupled to the memory and configured to read each condition of each rule of at least a portion of the target expert system program, and partition by associating each rule into one of a preselected set of partitions, and store partition information relating to the partitioning in the memory, each partition representing one of the multiple processors comprising the computer system, the partitioning performed in a manner to identify conditions shared by more than one rule and partition the rules as a function of the identified shared conditions and the processing statistics; and e. a compiler, coupled to the memory and configured to convert the rules of the at least portion of the target expert system program into a RETE network, each partitioned rule being compiled to execute on a corresponding processor based on the partition information.
Claim 18
Independent Claim
382 words
A system for building a reticular discrimination network (RETE network) capable of being partitioned across a set of multiple, interconnected computer processors that comprise a computer system, from at least a portion of a target expert system program, the target expert system program comprising a plurality of rules, each of which comprise a plurality of conditions grouped in a sequence by a set of logical connectors and a result to occur when the conditions of the rule are satisfied, certain ones of the plurality of rules having conditions that are equivalent to conditions in other ones of the plurality of rules, each of the conditions in a rule being represented by a test node in the RETE network, the test nodes being connected through associated join nodes representing the logical connectors of the rule, the system for building a RETE network comprising: a. a memory configured to store the target expert system program; b. a uniprocessor compiler coupled to the memory and configured to compile the target expert system program to form a RETE network for execution on a single processor and store the compiled target expert system program in the memory, certain ones of the conditions represented by nodes of the uniprocessor RETE network being shared by more than one rule; c. a processor, coupled to the memory for executing the target expert system program, to obtain processing statistics in connection with each node of the uniprocessor RETE network during execution; d. a partitioner, coupled to the memory and configured to read each condition of each rule of at least a portion of the target expert system program, and partition by associating each rule into one of a preselected set of partitions, and store partition information relating to the partitioning in the memory, each partition representing one of the multiple processors comprising the computer system, the partitioning performed in a manner to identify conditions shared by more than one rule and partition the rules as a function of the identified shared conditions and the processing statistics; and e. a compiler, coupled to the memory and configured to convert the rules of the at least portion of the target expert system program into a RETE network, each partitioned rule being compiled to execute on a corresponding processor based on the partition information.
Claim 19
Dependent Claim
96 words
The system of claim 16 wherein the partitioner configured to partition the rules of the target expert system program comprises: a. an iteration module coupled to the memory configured to read the conditions comprising a rule for temporarily assigning a rule to each of the different partitions; b. an evaluation module for estimating a performance measure for each such temporary assignment of the rule, the performance measure being derived from the processing statistics and condition sharing information; and c. an assignment module for permanently assigning the rule to the partition which yields the best performance measure.
Claim 20
Dependent Claim
39 words
The system as defined in claim 17 wherein the iteration module operates to temporarily assign the rules in an order by selecting for temporary assignment an unassigned rule which has a longest processing time according to the processing statistics.
Claim 21
Dependent Claim
16 words
The system of claim 18 further comprising a statistics generating module for generating the processing statistics.
Claim 22
Dependent Claim
50 words
In the system of claim 19 the evaluation module comprises means for calculating, as a function of the processing statistics and the condition sharing information, the time for the multiple processors to complete at least one recognize/act cycle for the target expert system program with the rule as temporarily assigned.
Claim 23
Dependent Claim
46 words
In the system of claim 20 the means for calculating the time to complete at least one recognize/act cycle for the target expert system program with the rule as temporarily assigned comprises means for computing a partition process time for each partition for each recognize/act cycle.
Claim 24
Dependent Claim
62 words
In the system of claim 21 the means for summing the number of activations in each cycle of each of the nodes representing the rules that are assigned to that partition comprises means for including the number of activations of nodes that are part of conditions that are shared by more than one rule assigned to that partition in sums only once.
Claim 25
Dependent Claim
41 words
In the system of claim 20 the statistics generating module comprises means for generating the numbers of activations per recognize/act cycle for blocks of test and join nodes representing sequences of conditions shared by rules in the target expert system program.
Claim 26
Dependent Claim
46 words
In the system of claim 23 the means for calculating the time to complete at least one recognize/act cycle for the target expert system program with the rule as temporarily assigned comprises means for computing a partition process time for each partition for each recognize/act cycle.
Claim 27
Dependent Claim
46 words
In the system of claim 24 the means for computing the partition process time for each partition for each recognize/act cycle comprises means for summing the number of activations in each cycle of each of the blocks representing the rules that are assigned to that partition.
Claim 28
Dependent Claim
51 words
In the system of claim 21 the means for calculating the time to complete at least one recognize/act cycle for the target expert system program with the rule as temporarily assigned comprises means for identifying the largest partition process time for each recognize/act cycle and summing the identified largest process times.
Claim 29
Dependent Claim
39 words
In the system of claim 21, the statistics generating module comprises means for generating the number of activations of conditions per recognize/act cycle for rules of the target expert system program from runs of the target expert system program.
Claim 30
Dependent Claim
44 words
In the system of claim 27, the means for computing the partition process time for each partition for each recognize/act cycle comprises means for summing the number of activations in each cycle of each of the nodes representing the rules assigned to that partition.
Claim 31
Dependent Claim
43 words
In the system of claim 16, the partitioner comprises means for enhancing the likelihood that the first and second rules will be associated into the same partition whenever the first and second rules share any of certain ones of the plurality of conditions.
Claim 32
Dependent Claim
40 words
In the system of claim 16, the partitioner comprises means for enhancing the likelihood that the first and second rules will be associated into the same partition whenever the first and second rules share any of the plurality of conditions.
Claim 33
Independent Claim
335 words
A method for building a reticular discrimination network (RETE network) capable of being partitioned across a set of multiple, interconnected computer processors that comprise a computer system, from at least a portion of a target expert system program, the target expert system program comprising a plurality rules each of which comprise a plurality of conditions, grouped in a sequence by a set of logical connectors, and a result to occur when the conditions of the rule are satisfied, certain ones of the plurality of rules having conditions that are equivalent to conditions in other ones of the plurality of rules, each of the conditions in the a rule being represented by a test node in the reticular discrimination net, the test nodes being connected through associated join nodes representing the logical connectors of the rule, the method comprising the steps of: a. storing the target expert system program in memory; b. compiling the target expert system program to form a RETE network for execution on a single processor; c. executing the target expert system program on a single processor; d. obtaining processing statistics in connection with each node of the RETE network during execution of the target expert system program; e. partitioning each rule of at least a portion of the target expert system program by reading each condition of each rule from the memory and associating each rule into one of a preselected set of partitions, each partition representing one of the multiple processors comprising the computer system, the partitioning performed in a manner to identify conditions shared by more than one rule and partition the rules as a function of the identified shared conditions and the processing statistics; f. storing partition information relating to the partitioning in the memory; and g. compiling the at least a portion of the target expert system program by converting the rules of the target expert system program into a RETE network, each partitioned rule being compiled to execute on a corresponding processor based on the partitioning information.
Claim 34
Independent Claim
335 words
A method for building a reticular discrimination network (RETE network) capable of being partitioned across a set of multiple, interconnected computer processors that comprise a computer system, from at least a portion of a target expert system program, the target expert system program comprising a plurality rules each of which comprise a plurality of conditions, grouped in a sequence by a set of logical connectors, and a result to occur when the conditions of the rule are satisfied, certain ones of the plurality of rules having conditions that are equivalent to conditions in other ones of the plurality of rules, each of the conditions in the a rule being represented by a test node in the reticular discrimination net, the test nodes being connected through associated join nodes representing the logical connectors of the rule, the method comprising the steps of: a. storing the target expert system program in memory; b. compiling the target expert system program to form a RETE network for execution on a single processor; c. executing the target expert system program on a single processor; d. obtaining processing statistics in connection with each node of the RETE network during execution of the target expert system program; e. partitioning each rule of at least a portion of the target expert system program by reading each condition of each rule from the memory and associating each rule into one of a preselected set of partitions, each partition representing one of the multiple processors comprising the computer system, the partitioning performed in a manner to identify conditions shared by more than one rule and partition the rules as a function of the identified shared conditions and the processing statistics; f. storing partition information relating to the partitioning in the memory; and g. compiling the at least a portion of the target expert system program by converting the rules of the target expert system program into a RETE network, each partitioned rule being compiled to execute on a corresponding processor based on the partitioning information.
Claim 35
Dependent Claim
79 words
In the method of claim 31 the rule partitioning step comprises the steps of: a. selecting a rule from the memory; b. temporarily assigning, by temporary association, the rule selected to each of the different partitions; c. estimating a performance measure for each such temporary assignment of the selected rule, the performance measure being derived from the processing statistics and condition sharing information; and d. assigning permanently the selected rule to the partition which yields a best performance measure.
Claim 36
Dependent Claim
47 words
In the method of claim 32 the step of selecting a rule comprises the further step of selecting rules in an order by selecting for temporary assignment a rule that has yet to be assigned and which has a longest processing time according to the processing statistics.
Claim 37
Dependent Claim
31 words
In the method of claim 33 the step of obtaining processing statistics comprises obtaining the number of activations of conditions per recognize/act cycle for rules of the target expert system program.
Claim 38
Dependent Claim
59 words
The method of claim 34 wherein the performance measure is determined by the further step of calculating the time it would take for the multiple processors to complete at least one recognize/act cycle for the target expert system with the rule as temporarily assigned, the calculation being determined as a function of the processing statistics and condition sharing information.
Claim 39
Dependent Claim
63 words
The method of claim 35 wherein in the step of calculating the time to complete at least one recognize/act cycle is derived by computing a partition process time for each partition during each recognize/act cycle, the partition process time being computed by summing the number of activations for each of the nodes which make up the rules that are assigned to that partition.
Claim 40
Dependent Claim
48 words
The method of claim 36 wherein the calculation step to calculate the time to complete at least one recognize/act cycle is derived by adding, for each condition, the number of activations per recognize/act cycle only once, when that condition is shared by multiple rules assigned to that partition.
Claim 41
Dependent Claim
42 words
In the method of claim 34 the step of obtaining processing statistics further comprises obtaining the numbers of activations per recognize/act cycle for blocks of test and join nodes representing sequences of conditions shared by rules in the target expert system program.
Claim 42
Dependent Claim
58 words
The method of claim 38 wherein the step of calculating the time to complete at least one recognize/act cycle is derived by computing a partition process time of each partition during each recognize/act cycle, the partition process time being computed as a function of the sum of the number of activations of each shared block of condition sequences.
Claim 43
Dependent Claim
51 words
The method of claim 39 wherein the step of calculating the time to complete at least one recognize/act cycle is derived by adding for each block of conditions its number of activations per recognize/act cycle only once, when that block is shared by more than one rule assigned to that partition.
Claim 44
Dependent Claim
36 words
The method of claim 36 wherein the step of calculating the time to complete at least one recognize/act cycle is derived by summing the largest partition process times for each recognize/act cycle over all recognize/act cycles.
Claim 45
Dependent Claim
46 words
In the method of claim 31, the step of partitioning comprises the step of enhancing the likelihood that the first and second rules will be associated into the same partition whenever the first and second rules share any of certain ones of the plurality of conditions.
Claim 46
Dependent Claim
43 words
In the method of claim 31, the step of partitioning comprises the step of enhancing the likelihood that the first and second rules will be associated into the same partition whenever the first and second rules share any of the plurality of conditions.
Claim 47
Independent Claim
300 words
A method for building a reticular discrimination network (RETE network) capable of being partitioned across a set of multiple, interconnected computer processors that comprise a computer system, from at least a portion of a target expert system, the target expert system comprising a plurality of rules each of which comprise a plurality of conditions, grouped in a sequence by a set of logical connectors, and a result to occur when the conditions of the rule are satisfied, certain ones of the plurality of rules having conditions that are equivalent to conditions in other ones of the plurality of rules, each of the conditions in a rule being represented by a test node in the reticular discrimination network, the test nodes being connected through associated join nodes representing the logical connectors of the rule, the method comprising: a. storing the target expert system in memory; b. partitioning each rule of the at least a portion of the target expert system by associating each rule into one of a preselected set of partitions, each partition representing one of the multiple processors comprising the computer system, the partitioning step being performed in a manner to identify conditions shared by more than one rule, the partitioning step being further performed in manner to increase the likelihood that a first and second rule will be associated into a same one of the preselected set of partitions when the first and second rules share any of certain ones of the plurality of conditions; c. storing the partition information in the memory; and d. compiling the at least a portion of the target expert system in the memory to convert the rules of the target expert system into a RETE network, each partitioned rule being compiled to execute on a corresponding microprocessor based on the partition information.
Claim 48
Independent Claim
300 words
A method for building a reticular discrimination network (RETE network) capable of being partitioned across a set of multiple, interconnected computer processors that comprise a computer system, from at least a portion of a target expert system, the target expert system comprising a plurality of rules each of which comprise a plurality of conditions, grouped in a sequence by a set of logical connectors, and a result to occur when the conditions of the rule are satisfied, certain ones of the plurality of rules having conditions that are equivalent to conditions in other ones of the plurality of rules, each of the conditions in a rule being represented by a test node in the reticular discrimination network, the test nodes being connected through associated join nodes representing the logical connectors of the rule, the method comprising: a. storing the target expert system in memory; b. partitioning each rule of the at least a portion of the target expert system by associating each rule into one of a preselected set of partitions, each partition representing one of the multiple processors comprising the computer system, the partitioning step being performed in a manner to identify conditions shared by more than one rule, the partitioning step being further performed in manner to increase the likelihood that a first and second rule will be associated into a same one of the preselected set of partitions when the first and second rules share any of certain ones of the plurality of conditions; c. storing the partition information in the memory; and d. compiling the at least a portion of the target expert system in the memory to convert the rules of the target expert system into a RETE network, each partitioned rule being compiled to execute on a corresponding microprocessor based on the partition information.
Claim 49
Dependent Claim
43 words
In the method of claim 44, the step of partitioning comprises the step of enhancing the likelihood that the first and second rules will be associated into the same partition whenever the first and second rules share any of the plurality of conditions.
Claim 50
Independent Claim
363 words
A system for building a reticular discrimination network (RETE network) capable of being partitioned across a set of multiple, interconnected computer processors that comprise a computer system, from at least a portion of a target expert system, the target expert system comprising a plurality of rules each of which comprise a plurality of conditions grouped in a sequence by a set of logical connectors and a result to occur when the conditions of the rule are satisfied, certain ones of the plurality of rules having conditions that are equivalent to conditions in other ones of the plurality of rules, each of the conditions in a rule being represented by a test node in the reticular discrimination net, the test nodes being connected through associated join nodes representing the logical connectors of the rule, the system for building a RETE network comprising: a. a memory configured to store the target expert system; b. a partitioner, coupled to the memory and configured to read each condition of each rule of the at least a portion of the target expert system, and partition the rules by associating each rule into one of a preselected set of partitions and storing partition information relating to the partitions in the memory, each partition representing one of the multiple, interconnected computer processors comprising the computer system, the partitioning performed in a manner to identify conditions shared by more than one rule and partition the rules as a function of the identified shared conditions, the partitioner comprises: i. an iteration module coupled to the memory and configured to read a rule and temporarily assign the rule to each of the different partitions, ii. an evaluation module for estimating a performance measure for each such temporary assignment of the rule, the performance measure being derived from predetermined processing statistics and condition sharing information, and iii. an assignment module for permanently assigning the rule to the partition which yields the best performance measure; and d. a compiler configured to convert the rules of the at least a portion of the target expert system into a RETE network, each partitioned rule being compiled to execute on a corresponding processor based on the partition information.
Claim 51
Independent Claim
363 words
A system for building a reticular discrimination network (RETE network) capable of being partitioned across a set of multiple, interconnected computer processors that comprise a computer system, from at least a portion of a target expert system, the target expert system comprising a plurality of rules each of which comprise a plurality of conditions grouped in a sequence by a set of logical connectors and a result to occur when the conditions of the rule are satisfied, certain ones of the plurality of rules having conditions that are equivalent to conditions in other ones of the plurality of rules, each of the conditions in a rule being represented by a test node in the reticular discrimination net, the test nodes being connected through associated join nodes representing the logical connectors of the rule, the system for building a RETE network comprising: a. a memory configured to store the target expert system; b. a partitioner, coupled to the memory and configured to read each condition of each rule of the at least a portion of the target expert system, and partition the rules by associating each rule into one of a preselected set of partitions and storing partition information relating to the partitions in the memory, each partition representing one of the multiple, interconnected computer processors comprising the computer system, the partitioning performed in a manner to identify conditions shared by more than one rule and partition the rules as a function of the identified shared conditions, the partitioner comprises: i. an iteration module coupled to the memory and configured to read a rule and temporarily assign the rule to each of the different partitions, ii. an evaluation module for estimating a performance measure for each such temporary assignment of the rule, the performance measure being derived from predetermined processing statistics and condition sharing information, and iii. an assignment module for permanently assigning the rule to the partition which yields the best performance measure; and d. a compiler configured to convert the rules of the at least a portion of the target expert system into a RETE network, each partitioned rule being compiled to execute on a corresponding processor based on the partition information.
References Cited (Backward Citations)
JPS61128358A
Cited
US4761746A
Cited
US4752890A
Cited
US4837735A
Cited
JPS6482226A
Cited
JPS6482227A
Cited
US4884217A
Cited
US4849905A
Cited
US4924908A
Cited
US4937760A
Cited
US4956791A
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 1996-08-30 | 1997-03-01 | 1997-08-31 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2000-08-30 | 2001-03-01 | 2001-08-31 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2004-08-30 | 2005-03-01 | 2005-08-31 | $7700.00 | Window Closed / Lapsed |