Claims Analysis
Total Claims: 15
Claim 1
Independent Claim
476 words
A parallel processor comprising: four or more processing units; edge crossing circuitry for receiving an item of data that includes two or more bits and for bit serially providing the item of data; the edge crossing circuitry comprising an edge crossing shift register; interconnecting circuitry for interconnecting a set of at least two of the processing units and the edge crossing circuitry; each processing unit in the set of at least two of the processing units comprising an internal shift register; the interconnecting circuitry interconnecting pairs of the processing units in the set to form an array of processing units that includes all the processing units in the set, the array including a first edge processing unit that is in the set and is at a first edge of the array; the interconnecting circuitry interconnecting the internal shift register of each processing unit in each pair of processing units to the internal shift register of the other processing unit in the pair so that an item of data from the internal shift register of either of the processing units in the pair can be transferred to the internal shift register of the other processing unit in the pair; the interconnecting circuitry further connecting the edge crossing circuitry and the first edge processing unit so that items of data from the edge crossing shift register can be transferred bit serially to the internal shift register of the first edge processing unit; the internal shift register of each of the processing units in the set and the edge crossing shift register together forming an extended shift register; data transfer circuitry connecting the array of processing units to the edge crossing circuitry so that the data transfer circuitry can transfer an item of data obtained from the array of processing units to the edge crossing circuitry; and control circuitry connected for providing control signals to the array of processing units, the data transfer circuitry, and the edge crossing circuitry; the control signals causing the data transfer circuitry to transfer a first item of data obtained from the array of processing units to the edge crossing circuitry, causing a shift of the extended shift register so that the edge crossing shift register provides a bit from the first item of data to the processing unit shift register of the first edge processing unit through the interconnecting circuitry, and causing the array of processing units to perform an operation; the first item of data relating to part of a first segment of an image, the image also including a second segment adjacent to the first segment; in performing the operation, one of the processing units using the bit from the first item of data to obtain a second item of data relating to a part of the second segment that is across an edge from the part of the first segment.
Claim 2
Dependent Claim
476 words
A parallel processor comprising: four or more processing units; edge crossing circuitry for receiving an item of data that includes two or more bits and for bit serially providing the item of data; the edge crossing circuitry comprising an edge crossing shift register; interconnecting circuitry for interconnecting a set of at least two of the processing units and the edge crossing circuitry; each processing unit in the set of at least two of the processing units comprising an internal shift register; the interconnecting circuitry interconnecting pairs of the processing units in the set to form an array of processing units that includes all the processing units in the set, the array including a first edge processing unit that is in the set and is at a first edge of the array; the interconnecting circuitry interconnecting the internal shift register of each processing unit in each pair of processing units to the internal shift register of the other processing unit in the pair so that an item of data from the internal shift register of either of the processing units in the pair can be transferred to the internal shift register of the other processing unit in the pair; the interconnecting circuitry further connecting the edge crossing circuitry and the first edge processing unit so that items of data from the edge crossing shift register can be transferred bit serially to the internal shift register of the first edge processing unit; the internal shift register of each of the processing units in the set and the edge crossing shift register together forming an extended shift register; data transfer circuitry connecting the array of processing units to the edge crossing circuitry so that the data transfer circuitry can transfer an item of data obtained from the array of processing units to the edge crossing circuitry; and control circuitry connected for providing control signals to the array of processing units, the data transfer circuitry, and the edge crossing circuitry; the control signals causing the data transfer circuitry to transfer a first item of data obtained from the array of processing units to the edge crossing circuitry, causing a shift of the extended shift register so that the edge crossing shift register provides a bit from the first item of data to the processing unit shift register of the first edge processing unit through the interconnecting circuitry, and causing the array of processing units to perform an operation; the first item of data relating to part of a first segment of an image, the image also including a second segment adjacent to the first segment; in performing the operation, one of the processing units using the bit from the first item of data to obtain a second item of data relating to a part of the second segment that is across an edge from the part of the first segment.
Claim 3
Dependent Claim
20 words
The parallel processor of claim 1 in which the first segment and the second segment are adjacent along the edge.
Claim 4
Dependent Claim
70 words
The parallel processor of claim 1 in which the first item of data includes a sequence of pixel values, each pixel value including K bits where K is greater than one; the interconnecting circuitry further interconnecting the edge crossing circuitry and the first edge processing unit so that a K-bit item of data from the edge crossing shift register can be transferred in parallel to the first edge processing unit.
Claim 5
Dependent Claim
56 words
The parallel processor of claim 1 in which the first item of data includes a sequence of pixel values, each pixel value including K bits where K is greater than one; the edge crossing shift register further comprising K-bit shifting circuitry for shifting an item of data by K bits within the edge crossing shift register.
Claim 6
Dependent Claim
76 words
The parallel processor of claim 1 in which the first item of data includes a sequence of pixel values, each pixel value including K bits where K is greater than one; the edge crossing circuitry further comprising K-bit shifting circuitry for receiving an item of data from the edge crossing shift register and for using the item of data to provide a K-bit shifted version of the item of data to the edge crossing shift register.
Claim 7
Dependent Claim
20 words
The parallel processor of claim 1 in which the first item of data includes a sequence of binary pixel values.
Claim 8
Dependent Claim
31 words
The parallel processor of claim 1 in which the control signals cause the edge crossing shift register and the internal shift register of the first edge processing unit to shift concurrently.
Claim 9
Dependent Claim
105 words
The parallel processor of claim 7 in which the interconnecting circuitry interconnects the edge crossing shift register to the internal shift register of the first edge processing unit so that items of data can be transferred bit serially from the internal shift register of the first edge processing unit to the edge crossing shift register; the control signals further causing the edge crossing shift register and the internal shift register of the first edge processing unit to shift concurrently so that a bit of data provided by the internal shift register of the first edge processing unit is received by the edge crossing shift register.
Claim 10
Dependent Claim
60 words
The parallel processor of claim 1 in which the array further includes a second edge processing unit at a second edge of the array; the interconnecting circuitry further connecting the edge crossing circuitry and the second edge processing unit so that items of data from the edge crossing circuitry can be transferred bit serially to the second edge processing unit.
Claim 11
Dependent Claim
178 words
The parallel processor of claim 9 in which the first edge processing unit comprises a first internal shift register and the second edge processing unit comprises a second internal shift register; the edge crossing shift register having a first end and a second end; the interconnecting circuitry interconnecting the edge crossing shift register and the first and second internal shift registers so that items of data from the second internal shift register can be transferred bit serially to the second end of the edge crossing shift register and so that items of data from the first end of the edge crossing shift register can be transferred bit serially to the first internal shift register; the control signals further causing the edge crossing shift register and the first and second internal shift registers to shift concurrently so that when a bit from the second internal shift register is transferred to the second end of the edge crossing shift register, another bit from the first end of the edge crossing shift register is transferred to the first internal shift register.
Claim 12
Dependent Claim
143 words
The parallel processor of claim 10 in which the interconnecting circuitry further interconnects the edge crossing shift register and the first and second internal shift registers so that items of data from the first internal shift register can be transferred bit serially to the first end of the edge crossing shift register and so that items of data from the second end of the edge crossing shift register can be transferred bit serially to the second internal shift register; the control signals further causing the edge crossing shift register and the first and second internal shift registers to shift concurrently so that when a bit from the first internal shift register is transferred to the first end of the edge crossing shift register, another bit from the second end of the edge crossing shift register is transferred to the second internal shift register.
Claim 13
Dependent Claim
32 words
The parallel processor of claim 9 in which the control signals further cause the second edge processing unit to provide the first item of data for transfer to the edge crossing circuitry.
Claim 14
Dependent Claim
81 words
The parallel processor of claim 1 in which each processing unit comprises processing circuitry for responding to instructions by performing operations on items of data; the control circuitry further providing a series of items of instruction data to the processing circuitry of each of the processing units; the items of instruction data indicating a sequence of instructions; the control circuitry being connected for providing each item in the series to the processing circuitry of all of the processing units in parallel.
Claim 15
Dependent Claim
98 words
The parallel processor of claim 1 in which the array further includes a second edge processing unit at a second edge of the array; the interconnecting circuitry further connecting the edge crossing shift register and the internal shift register of the second edge processing unit so that items of data from the second edge processing unit can be transferred bit serially to the edge crossing circuitry; the control signals causing the shift of the extended shift register so that the internal shift register of the second edge processing unit provides a bit to the edge crossing shift register.