naga_patent/drafts/claims.md

Preliminary Claims

Independent Claim 1: Computing Substrate

1. A computing system comprising: a memory array configured to store cell states of a cellular automaton; update circuitry configured to repeatedly apply a fixed local evolution function to the cell states stored in the memory array; and an image loader configured to load into the memory array a cellular automaton state image encoding an executable computation; wherein the fixed local evolution function computes a next state for a cell from a current state of the cell and states of a bounded neighborhood of the cell; wherein the cellular automaton state image comprises spatially arranged computational structures that evolve under the fixed local evolution function; and wherein computational output is obtained by reading one or more designated output regions of the memory array after one or more applications of the fixed local evolution function.

Dependent Claims

2. The computing system of claim 1, wherein the cell states comprise multi-bit cell states.

3. The computing system of claim 2, wherein each cell state comprises a 6-bit value.

4. The computing system of claim 1, wherein a memory word stores cell states for a plurality of independent cellular automaton cores.

5. The computing system of claim 4, wherein a 64-bit memory word stores ten 6-bit cell states and one or more additional bits for metadata, parity, cyclic redundancy checking, or error detection.

6. The computing system of claim 1, wherein the fixed local evolution function is cell-type-count preserving.

7. The computing system of claim 1, wherein the fixed local evolution function is information-conserving.

8. The computing system of claim 1, wherein the fixed local evolution function is reversible.

9. The computing system of claim 1, wherein the spatially arranged computational structures comprise persistent propagating structures.

10. The computing system of claim 9, wherein the persistent propagating structures comprise gliders.

11. The computing system of claim 9, wherein the cellular automaton state image comprises interaction regions configured to cause persistent propagating structures to collide or interact to perform logic, routing, detection, emission, or transformation operations.

12. The computing system of claim 1, wherein the cellular automaton state image defines an effective processor architecture.

13. The computing system of claim 12, wherein the effective processor architecture comprises at least one of register regions, instruction-processing regions, memory-interface regions, routing regions, or output regions.

14. The computing system of claim 1, wherein the cellular automaton state image defines a nonlinear simulation engine.

15. The computing system of claim 14, wherein the nonlinear simulation engine is configured for at least one of particle simulation, radiation transport, plasma wakefield simulation, field propagation, collision cascade simulation, nonlinear wave computation, or spatial constraint solving.

16. The computing system of claim 1, wherein loading a different cellular automaton state image changes an effective hardware architecture executed by the computing system without modifying the update circuitry implementing the fixed local evolution function.

17. The computing system of claim 1, wherein the fixed local evolution function is implemented by an application-specific integrated circuit.

18. The computing system of claim 1, wherein the fixed local evolution function is implemented by near-memory processing circuitry or processing-in-memory circuitry.

19. The computing system of claim 1, wherein the update circuitry applies a pull-based update in which each cell computes and writes only its own next state.

20. The computing system of claim 19, wherein simultaneous neighbor influences are collected into a local pattern and mapped to the next state without selecting a winning influence according to a fixed global direction priority.

21. The computing system of claim 19, wherein the fixed local evolution function is equivariant under one or more lattice rotations, reflections, coordinate inversions, or coordinate permutations.

22. The computing system of claim 19, wherein a deterministic priority orientation for resolving simultaneous neighbor influences varies across cells, rows, planes, tiles, or cores according to a balanced spatial pattern.

23. The computing system of claim 22, wherein the balanced spatial pattern assigns each of a plurality of lattice directions an equal number of priority assignments within a repeating supercell.

24. The computing system of claim 22, wherein the deterministic priority orientation is derived from address bits of a cell location.

25. The computing system of claim 1, further comprising validation circuitry configured to compare outputs from two or more independent cellular automaton cores.

26. The computing system of claim 25, wherein the two or more independent cellular automaton cores execute mirrored, rotated, coordinate-transformed, or otherwise transformed versions of the cellular automaton state image.

27. The computing system of claim 1, wherein the cellular automaton state image comprises one or more emitter structures configured to generate persistent propagating structures under the fixed local evolution function.

28. The computing system of claim 1, wherein the cellular automaton state image comprises one or more detector structures configured to alter one or more output regions in response to arrival of persistent propagating structures.

29. The computing system of claim 1, wherein the executable computation is encoded primarily as a spatial configuration rather than as a sequential instruction stream.

30. The computing system of claim 1, wherein update cost for a fixed lattice region is determined primarily by a number of cells updated rather than by a number of pairwise interactions represented within the cellular automaton state image.

Independent Claim 31: Method

31. A method of executing a computation, comprising: storing cell states of a cellular automaton in a memory array; loading into the memory array a cellular automaton state image that encodes an executable computation; repeatedly applying, by update circuitry, a fixed local evolution function to the cell states, wherein the fixed local evolution function computes next states from bounded neighborhoods of cells; evolving the cellular automaton state image into one or more output states; and reading one or more designated output regions of the memory array to obtain a result of the executable computation.

32. The method of claim 31, further comprising loading a second cellular automaton state image that defines a different effective processor, accelerator, or simulation engine without modifying the fixed local evolution function.

33. The method of claim 31, wherein the cellular automaton state image directly implements a nonlinear simulation without emulating a conventional central processing unit.

34. The method of claim 31, wherein the cellular automaton state image implements a virtual processor using persistent propagating structures as information carriers.

35. The method of claim 31, further comprising validating the result by comparing outputs from redundant cellular automaton cores.

Independent Claim 36: Non-Transitory Medium

36. A non-transitory computer-readable medium storing instructions that, when executed by one or more processors or update circuits, cause performance of operations comprising: receiving a cellular automaton state image encoding an executable computation; storing the cellular automaton state image in a memory array; repeatedly applying a fixed local evolution function to cell states of the cellular automaton state image; and reading one or more designated output regions after one or more applications of the fixed local evolution function to obtain computational output.