Virtual supercomputer

• US 20040205751A1
• Filed: 04/09/2004
• Published: 10/14/2004
• Est. Priority Date: 04/09/2003
• Status: Active Grant

• Current Assignee: Veriscape, Inc.
• Original Assignee: Veriscape, Inc.
• Inventors: Berkowitz, Gary Charles, Wurtz, Charles Christopher
• Granted Patent: US 7,774,191 B2
• Time in Patent Office: Days
• US Class Current: 718/1
• CPC Class Codes: G06F 16/22 Indexing; Data structures therefor; Storage structures
• CPC Class Codes: G06F 9/3004 to perform operations on memory
• CPC Class Codes: G06F 9/5077 Logical partitioning of resources; Management or configuration of virtualized resources specific details on emulation or internal functioning of virtual machines G06F9/455
• CPC Class Codes: G06Q 40/06 Asset management; Financial planning or analysis

Abstract

The virtual supercomputer is an apparatus, system and method for generating information processing solutions to complex and/or high-demand/high-performance computing problems, without the need for costly, dedicated hardware supercomputers, and in a manner far more efficient than simple grid or multiprocessor network approaches. The virtual supercomputer consists of a reconfigurable virtual hardware processor, an associated operating system, and a set of operations and procedures that allow the architecture of the system to be easily tailored and adapted to specific problems or classes of problems in a way that such tailored solutions will perform on a variety of hardware architectures, while retaining the benefits of a tailored solution that is designed to exploit the specific and often changing information processing features and demands of the problem at hand.

Claims

1. A virtual supercomputer apparatus, comprising:

a reconfigurable virtual machine (processor), having its own instruction set, and operating on an underlying physical hardware processor;

or a local or distributed network of such virtual machines;

where the virtual machine architecture is designed to handle the class of problems having a solution describable in terms of nodes,

where a node comprises an index word and a data word, and

where each node represents one or more data structures such as;

numeric tags, character tags, boolean flags, numeric values, character values, objects IDs, database-record IDs, simple arrays, variable-density multidimensional arrays, symbolic functions, mathematical functions, connection pointers to other nodes, function pointers, lookup-table list pointers, linked-lists, or even pointers to other solution spaces or data representations; and

where nodes are interconnected in any topology, such as;

independent point-clouds, ordered sets of points, acyclic graphs, cyclic graphs, balanced trees, recombining graphs, meshes, lattices, and various hybrids or combinations of such representations;

and where a virtual machine includes virtual hardware units to;

configure nodes and virtual processor architecture,

including register structures, node data structures, arithmetic modes, and memory schemes;

and units to

create (that is, instantiate) nodes, compute results for (that is, populate) nodes, move among (that is, navigate) nodes, and alter (that is, evolve) nodes and their interconnections;

and a unit to

provide highly-optimized function evaluation and fixed-point integer arithmetic, with application-selectable precision;

and a unit to

manage the distribution of data and processes to networked machines.

2. A virtual supercomputing system, comprising:

the virtual machine of claim 1, and

a multi-tasking operating system for the virtual machine, that creates a new virtual CPU for each task thread, and contains;

software engines for configuring, instantiating, populating, navigating, and modifying (evolving) nodes;

autonomous daemons for background processing of nodes; and

a toolbox containing frequently-used engine programs;

and allows for multiprocessing via multiple virtual machines implemented on a network of underlying hardware processors, in a local or distributed cluster;

an assembler for translating operating-system calls into virtual machine operation codes; and

platform drivers for implementing virtual-machine operations on the underlying physical platform processor, via

a platform assembler for translating virtual machine operations into instruction-codes for platform operations.

3. A method, or procedure, using the apparatus of claim 1 and the system of claim 2, that allows for faster, and therefore lower cost, software application creation, and that produces computer programs that rapidly generate ‘

good enough’

solutions to computationally complex and/or high-demand problems that are describable in terms of nodes, via a set of non-sequential processes, comprising;

a. matching the virtual machine architecture and solution manifold to the problem architecture;

b. adapting the solution manifold in response to changing demands in problem architecture or data;

c. adapting the virtual-machine architecture in response to changing demands in problem architecture or data;

d. using application-selectable arithmetic precision to rapidly compute ‘

accurate enough’

calculations when evaluating nodes;

e. using software-emulation of supercomputing techniques, such as small instruction set, simple and efficient data representation and handling, inherent vector representation, limited data/calculation modes, interleaved memory, table lookup, induced pointers, and distributed &

parallelized computation;

f. separating the populating and navigating of nodes, to allow for pre-computation of manifolds, so that navigation of possible solutions occurs in near real-time;

g. using autonomous, second-order dedicated processes that operate in background, as concurrent tasks, to collect garbage, prune trees, condense redundancies, process edit-queues, interpolate with finer granularity (mesh enhancement) around selected nodes in state-space, or to extrapolate and elaborate the data structures, during both population and navigation phases;

h. generating virtual CPUs for each operating-system task thread.