Patent References 3287649 Data processing apparatus for highly parallel execution of data structure operations Method and apparatus for building knowledge-based systems Method of integrating software application programs using an attributive data model database Method for integrating a knowledge-based system with an arbitrary database system Computer based inference engine device and method thereof for integrating backward chaining and forward chaining reasoning Expert system for communications network System for modifying relocatable object code files to monitor accesses to dynamically allocated memory Method and apparatus for executing tasks by following a linked list of memory packets Task flow computer architecture InventorsAssigneeApplicationNo. 580921 filed on 12/29/1995US Classes:709/216, Accessing another computer's memory709/212, COMPUTER-TO-COMPUTER DIRECT MEMORY ACCESSING709/237, Computer-to-computer handshaking709/248, MULTICOMPUTER SYNCHRONIZING719/310MISCELLANEOUSExaminersPrimary: Rinehart, Mark H.Attorney, Agent or FirmInternational ClassesG06F 013/38G06F 015/17 AbstractA cell-based system for computation and communication. A cell is a well-defined structure that associates executable code and data into a basic computational module. A processor can traverse a cell and execute the instructions therein. Cells are arranged into a network with each cell linked to other cells by forward branches or other paths. A processor branches to a cell by loading its program counter with the address of the cell or by sending a packet of data across a communications network to activate remote processing in a cell. Cells have a path selection section for evaluating conditional branches and causing branches to be followed. Multiple branches may be followed in parallel from any cell. A tree-like organization is superimposed on the network by distinguishing one branch to each cell as a superbranch incorporating a return path (called "recession path"). A cell also has a convergence section that handles processors and threads of execution that return to that cell by this path (called "recession"). The convergence section implements rules that control which, if any, processors or threads continue to execute upon recession as well as rules for handling any resulting values associated with the processors or threads. The cellular computational module controls highly parallel and distributed processing and supports dynamic self-modification of the system. Information is acquired within the system not only through the acquisition of data and the modification of executable code within the cells, but also through the encoding of information and behavior into the structure of links that connect the cells.Other References
| |