Particle interaction processing system

Claim 2

A data processor for processing data of a plurality of sites of a lattice to simulate an interactive process across the lattice, the data processor comprising: storage storing a state vector for each site of the lattice, each state vector including a like set of bits representing elements at the site of the lattice, each element having specific amplitude and specific direction of movement within the lattice; interaction processing means for processing, for each lattice site, the state vector to generate a second state vector in a time step; and move processing means for processing the state vectors across the lattice to generate with each time step new state vectors which reflect movement of elements in the lattice to new sites; the interaction processing means comprising logic which applies at least one rule to bits of the state vector to modify the state vector and means for selecting bits of the state vector and applying the bits to the logic such that the logic applies the at least one rule in successive cycles within each time step to different permutations of bits of the state vector as modified within the time step.

Claim 3

A data processor as claimed in claim 1 wherein the interaction processing means processes the state vector to cause a second state vector which reflects a transfer of energy between particles, different rates of energy being transferred in one direction then in an inverse direction.

Claim 4

A data processor as claimed in claim 1 further comprising a masking input to the logic to prevent performance of a collision rule.

Claim 5

A data processor as claimed in claim 1 further comprising local memory which is local to the logic and means for retrieving from storage full state vectors of plural adjacent sites of a volume of the lattice and storing the retrieved state vectors in local memory, and the means for selecting selects like subsets of bits from all state vectors of the volume and applies the bits in parallel to the logic for application of a common rule in parallel to plural sites, different subsets of bits being applied to the logic in sequence to apply to the common rule to different permutations of bits.

Claim 6

A data processor as claimed in claim 1 wherein the logic and the means for selecting process the state vector to cause a second state vector which reflects a transfer of energy between moving elements of different four-dimensional velocities of different energy states, different rates of energy being transferred in one direction then in an inverse direction.

Claim 7

A system as claimed in claim 1 further comprising a masking input to the interaction processing means to prevent performance of interaction rules.

Claim 8

A data processor as claimed in claim 1 wherein the move processing means provides movement of elements on a nonuniform lattice grid, portions of the grid having different unit dimensions.

Claim 9

A data processor as claimed in claim 1 wherein the state vector identifies at least 54 elements at each site of the lattice.

Claim 10

A data processor as claimed in claim 1 wherein the interaction processing means processes only a subset of bits of each state vector in each cycle.

Claim 11

A system as claimed in claim 1 wherein the interaction processing means selects like bits of a plurality of sites of the lattice together and applies a common interaction rule to all selected bits simultaneously in a common processor.

Claim 12

A system as claimed in claim 10 further comprising a sequencer for selecting the bits to be processed in an interaction rule from all bits of the plurality of sites to be processed in the common processor.

Claim 13

A data processor as claimed in claim 1 wherein the elements are particles and the rule performed by the logic is a collision rule, the logic causing a collision when the state vector indicates that all of selected colliding particles are present and that there are holes where all of selected collided particles are to be present after the collision.

Claim 14

A data processor as claimed in claim 12 wherein the collision rule is performed in combinational logic which is specific to the collision rule.

Claim 17

A method as claimed in claim 14 wherein the elements represented by the state vector represent particles and the rule applied by the logic is a collision rule, the logic allowing a collision when the state vector indicates that all colliding particles are present and that there are holes where all collided particles are to be present after the collision.

Claim 18

A method as claimed in claim 14 wherein the state vector is processed to cause a second state vector which reflects a transfer of energy between particles, different rates of energy being transferred in one direction then in an inverse direction.

Claim 19

A method as claimed in claim 14 further comprising a masking input to the logic to prevent application of a collision rule.

Claim 20

A method as claimed in claim 14 comprising selecting like bits of a plurality of sites of the lattice together and applying a common rule to all selected bits in a common processor.

Claim 21

A method as claimed in claim 14 herein full state vectors of plural adjacent sites of a volume of the lattice are retrieved from memory and stored local to the logic, and like subsets of bits from all state vectors of the volume are applied in parallel to the logic for application of a common rule in parallel to plural sites, different subsets of bits being applied to the logic in sequence to apply the common rule to different permutations of bits.

Claim 22

A method as claimed in claim 14 wherein the new state vector reflects a transfer of energy between moving elements of different four-dimensional velocities of different energy states, different rates of energy being transferred in one direction than in an inverse direction.

Claim 23

A method as claimed in claim 14 wherein the state vectors are processed to provide movement of elements on a nonuniform lattice grid, portions of the grid having different unit dimensions.

Claim 24

A method as claimed in claim 14 wherein the state vector identifies at least 54 elements at each site of the lattice.

Claim 25

A method as claimed in claim 14 wherein only a subset of bits of each state vector is selected in each cycle.

Claim 26

A flow processing system comprising: storage storing particle representations of plural particles for each of plural sites of a lattice, the lattice comprising a first grid of a first unit dimension spanning a first volume and a second grid of a second unit dimension spanning a second volume; collision processing means for processing, for each lattice site, particle representations to generate second particle representations according to collision rules; move processing means for processing the representations across the lattice to generate new representations which reflect movement of particles in the lattice to new sites; and grid interface processing means for coalescing particles of the first grid into larger particles of the second grid and for exploding particles of the second grid into multiple smaller particles in the first grid.

Claim 27

A flow processing system comprising: storage storing particle representations of plural particles for each of plural sites of a lattice, the lattice comprising a first grid of a first unit dimension spanning a first volume and a second grid of a second unit dimension spanning a second volume; collision processing means for processing, for each lattice site, particle representations to generate second particle representations according to collision rules; move processing means for processing the representations across the lattice to generate new representations which reflect movement of particles in the lattice to new sites; and grid interface processing means for coalescing particles of the first grid into larger particles of the second grid and for exploding particles of the second grid into multiple smaller particles in the first grid.

Claim 28

A flow processing system as claimed in claim 24 wherein the collision processing means processes particle representations of the respective grids at rates which are inversely related to the unit dimensions of the grids.

Claim 29

A flow processing system as claimed in claim 25 wherein the first and second grids overlap in a volume in which the particles coalesce and explode.

Claim 30

A flow processing system as claimed in claim 26 further comprising means for coalescing particles of the first grid into a particle of the second grid having a larger mass than the total mass of the coalesced particles and for subsequently eliminating particles of the first grid without coalescing them into particles of the second grid in order to maintain a conservation of mass at the grid interface over time.

Claim 31

A flow processing system as claimed in claim 27 further comprising means for storing a count for a block of sites in the first grid which represents an imbalance in the mass of particles coalesced from the first grid into particles of the second grid.

Claim 32

A flow processing system as claimed in claim 27 further comprising means for creating particles in the first grid other than by explosion of particles from the second grid and subsequently annihilating particles in the first grid to maintain conservation of mass at the interface over time.

Claim 33

A flow processing system as claimed in claim 29 wherein the particles in the first grid are created and annihilated as a function of the presence of other particles at the interface moving from the first grid toward the second grid.

Claim 34

A flow processing system as claimed in claim 26 wherein the particle representations identify, for each particle, one of plural possible particle speeds.

Claim 35

A flow processing system as claimed in claim 31 wherein the particle representations comprise sets of bits for each lattice site, each bit of a set of bits representing a particle of particular energy and a particular direction of movement at the site.

Claim 36

A flow processing system as claimed in claim 31 wherein the first and second grids overlap by different amounts for particles of different speeds.

Claim 37

A flow processing system as claimed in claim 31 wherein the collisions processing means processes a reduced set of collision rules at the grid interface to prevent transfers of energy between particles which increase the speed of particles.

Claim 40

A method as recited in claim 35 further comprising processing particle representations of the respective grids at rates which are inversely related to the unit dimensions of the grids.

Claim 41

A method as claimed in claim 36 wherein the first and second grids overlap in a volume in which the particles coalesce and explode.

Claim 42

A method as claimed in claim 37 wherein particles of the first grid are coalesced into a particle of the second grid having a larger mass than the total mass of the coalesced particles and wherein particles of the first grid are eliminated without coalescing them into particles of the second grid in order to maintain a conservation of mass at the grid interface over time.

Claim 43

A method as claimed in claim 38 further comprising storing a count for a block of sites in the first grid which represents an imbalance in the mass of particles coalesced from the first grid into particles of the second grid.

Claim 44

A method as claimed in claim 38 further comprising creating particles in the first grid other than by explosion of particles from the second grid and subsequently annihilating particles in the first grid to maintain conservation of mass at the interface over time.

Claim 45

A method as claimed in claim 40 wherein the particles in the first grid are created and annihilated as a function of the presence of other particles at the interface moving from the first grid toward the second grid.

Claim 46

A method as claimed in claim 37 wherein the particle representations identify, for each particle, one of plural possible particle speeds.

Claim 47

A method as claimed in claim 42 wherein the particle representations comprise sets of bits for each lattice site, each bit of a set of bits representing a particle of particular energy and a particular direction of movement at the site.

Claim 48

A method as claimed in claim 42 wherein the first and second grids overlap by different amounts for particles of different speeds.

Claim 49

A method as claimed in claim 42 wherein a reduced set of collision rules are processed at the grid interface to prevent transfers of energy between particles which increase the speed of particles.

Claim 50

A data processor for processing data of a plurality of sites of a lattice to simulate an interactive process across the lattice, the data processor comprising: storage storing a state vector for each site of the lattice, each state vector including a like set of bits representing elements at the site of the lattice, each element having specific amplitude and specific direction of movement within the lattice; interaction processing means for processing, for each lattice site, the state vector to generate a second state vector in a time step, the interaction processing means selecting, in each of successive cycles within a time step, like subsets of bits of the state vectors of a plurality of sites of the lattice together and applying a common interaction rule to all selected bits in a common processor; and move processing means for processing the state vectors across the lattice to generate with each time step new state vectors which reflect movement of elements in the lattice to new sites.

Claim 51

A data processor for processing data of a plurality of sites of a lattice to simulate an interactive process across the lattice, the data processor comprising: storage storing a state vector for each site of the lattice, each state vector including a like set of bits representing elements at the site of the lattice, each element having specific amplitude and specific direction of movement within the lattice; interaction processing means for processing, for each lattice site, the state vector to generate a second state vector in a time step, the interaction processing means selecting, in each of successive cycles within a time step, like subsets of bits of the state vectors of a plurality of sites of the lattice together and applying a common interaction rule to all selected bits in a common processor; and move processing means for processing the state vectors across the lattice to generate with each time step new state vectors which reflect movement of elements in the lattice to new sites.

Claim 52

A data processor as claimed in claim 46 further comprising a sequencer for selecting the subset of bits to which an interaction rule is applied from all bits of the plurality of sites to be processed in the common processor.

Claim 55

A method as claimed in claim 48 wherein the subset of bits is selected by a sequencer from all bits of the plurality of sites to be processed in the common processor.

Claim 56

A data processor for processing data of a plurality of sites of a lattice to simulate an interactive process across the lattice, the data processor comprising: storage storing state vectors of plural elements, including moving elements of at least three possible energy states, for each of plural sites of a lattice, the sum over the directions of the four-fold tensor product of unit vectors in each energy state forming a purely isotropic rank four tensor in each energy state; interaction processing means for processing, for each lattice site, the state vectors to generate a second state vector which reflects a transfer of energy between elements of different energy states; and move processing means for processing the state vectors across the lattice to generate new state vectors which reflect movement of elements in the lattice to new sites.

Claim 57

A data processor for processing data of a plurality of sites of a lattice to simulate an interactive process across the lattice, the data processor comprising: storage storing state vectors of plural elements, including moving elements of at least three possible energy states, for each of plural sites of a lattice, the sum over the directions of the four-fold tensor product of unit vectors in each energy state forming a purely isotropic rank four tensor in each energy state; interaction processing means for processing, for each lattice site, the state vectors to generate a second state vector which reflects a transfer of energy between elements of different energy states; and move processing means for processing the state vectors across the lattice to generate new state vectors which reflect movement of elements in the lattice to new sites.

Claim 60

A data processor for processing data of a plurality of sites of a lattice to simulate an interactive process across the lattice, the data processor comprising: storage storing a state vector for each site of the lattice, each state vector including a like set of bits representing at least fifty-four particles at the site of the lattice, each particle having specific amplitude and specific direction of movement within a four dimensional lattice; collision processing means for processing, for each lattice site, the state vector to generate a second state vector in a time step, the collision processing means comprising a sequencer for selecting like bits of a plurality of sites of the lattice together and combinational logic for applying a common collision rule on all selected bits to modify this state vector, the combinational logic applying the collision rule by causing a collision when the state vector indicates that all of selected colliding particles are present and that there are holes where all of selected collided particles are to be present after the collision, the combinational logic causing a second state vector which reflects a transfer of energy between moving elements of different four dimensional velocities and different energy states, different rates of energy being transferred in one direction than in an inverse direction, and the combinational logic applying the collision rule in successive cycles within each time step to different permutations of bits of the modified state vector; and move processing means for processing the state vectors across the lattice to generate with each time step new state vectors which reflect movement of elements in the lattice to new sites.

Claim 61

A data processor for processing data of a plurality of sites of a lattice to simulate an interactive process across the lattice, the data processor comprising: storage storing a state vector for each site of the lattice, each state vector including a like set of bits representing at least fifty-four particles at the site of the lattice, each particle having specific amplitude and specific direction of movement within a four dimensional lattice; collision processing means for processing, for each lattice site, the state vector to generate a second state vector in a time step, the collision processing means comprising a sequencer for selecting like bits of a plurality of sites of the lattice together and combinational logic for applying a common collision rule on all selected bits to modify this state vector, the combinational logic applying the collision rule by causing a collision when the state vector indicates that all of selected colliding particles are present and that there are holes where all of selected collided particles are to be present after the collision, the combinational logic causing a second state vector which reflects a transfer of energy between moving elements of different four dimensional velocities and different energy states, different rates of energy being transferred in one direction than in an inverse direction, and the combinational logic applying the collision rule in successive cycles within each time step to different permutations of bits of the modified state vector; and move processing means for processing the state vectors across the lattice to generate with each time step new state vectors which reflect movement of elements in the lattice to new sites.

Claim 62

A data processor as claimed in claim 52 wherein the move processing means provides movement of elements on a nonuniform lattice grid, portions of the grid having different unit dimensions.

Claim 65

A method as claimed in claim 54 wherein the state vectors are processed across the lattice to provide movement of elements on a nonuniform lattice grid, portions of the grid having different unit dimensions.