Granted Patent B2

US 12,670,034 · App. 19/030,706 · Granted Jun 30, 2026

Inventors: Zvi Tannenbaum (Newport Coast, CA); Dean E. Dauger (Huntington Beach, CA)

Assignee: Advanced Cluster Systems, Inc.

G06F9/5072G06F9/54G06F15/76H04L41/04

Abstract

Claims (34)

1 . A computer cluster comprising:

a first processor that accesses a memory that stores instructions associated with a first kernel and instructions associated with a first cluster node;

wherein the first processor:

executes the instructions associated with the first cluster node to communicate with the first kernel and receive user commands from a user interface, wherein the first cluster node communicates with the first kernel by accessing the memory within which the first kernel resides;

executes the instructions associated with the first kernel to translate the user commands into calls for execution of matrix operations on an array of data elements;

transmits the calls for execution of the matrix operations on the array of data elements to a second cluster node, wherein the second cluster node executes the matrix operations on the array of data elements; and

receives a result of the execution of the matrix operations on the array of data elements.

2 . The computer cluster of claim 1 , wherein the second cluster node accesses a second kernel residing in the memory.

3 . The computer cluster of claim 2 , wherein the second kernel is executed by a second processor.

4 . The computer cluster of claim 2 , wherein the memory comprises a first memory coupled to at least the first processor and a second memory coupled to at least a second processor, wherein the first kernel and the first cluster node reside in the first memory, and wherein the second kernel and the second cluster node reside in the second memory.

5 . The computer cluster of claim 2 , wherein the first cluster node communicates with the second cluster node using commands that implement at least a portion of a Message-Passing Interface configured to distribute tasks among at least one of the first kernel or the second kernel without requiring user commands.

6 . The computer cluster of claim 1 , wherein a first result of executing the matrix operations on the array of data elements are sent to a third cluster node to execute second matrix operations on the first result using a third kernel node in communication with the third cluster node.

7 . The computer cluster of claim 6 , wherein the third kernel is executed by a third processor, and wherein the third processor receives the first result from a second processor that executes the second kernel.

8 . The computer cluster of claim 6 , wherein the result of the execution of the matrix operations comprises a result of the second matrix operations.

9 . The computer cluster of claim 6 , wherein the result of the execution of the matrix operations is received from the third cluster node.

10 . The computer cluster of claim 6 , further comprising the first cluster node, the second cluster node, and the third cluster node.

11 . The computer cluster of claim 6 , wherein expressions are evaluated across the first cluster node, the second cluster node, and the third cluster node in parallel.

12 . The computer cluster of claim 1 , further comprising a cluster configuration module that initializes at least one of the first cluster node or the second cluster node.

13 . The computer cluster of claim 1 , wherein the first cluster node communicates with the second cluster node via a communications interface that implements a peer-to-peer architecture.

14 . The computer cluster of claim 13 , wherein the communications interface comprises at least one of: a local bus, a local area network (LAN), a wide area network (WAN), a storage area network (SAN), a TCP/IP interface over Ethernet, or a wireless network interface.

15 . The computer cluster of claim 1 , wherein the first cluster node comprises an advanced functions module that evaluates an expression across cluster nodes of the computer cluster in parallel.

16 . The computer cluster of claim 15 , wherein the expression comprises at least one of a Fourier transform or a matrix operation.

17 . The computer cluster of claim 1 , wherein the first processor is further to provide the result to the user interface.

18 . The computer cluster of claim 1 , wherein at least one of the first kernel or the second kernel comprises a single-node kernel.

19 . The computer cluster of claim 1 , wherein at least the first cluster node and the second cluster node communicate with each other to act as a cluster.

20 . The computer cluster of claim 1 , wherein the first cluster node is configured to receive a command to start a cluster initialization process for the computer cluster.

21 . A computer cluster comprising:

a plurality of nodes, wherein each node of the plurality of nodes comprises a single-node kernel and wherein each of the plurality of nodes is configured to access a non-transitory computer-readable medium comprising program code that, when executed, causes a hardware processor to evaluate mathematical expressions; and

a first hardware processor that:

accesses the program code from the non-transitory computer-readable medium;

provides a user interface responsive to execution of the program code by a first node of the plurality of nodes;

interprets user instructions received via the user interface;

evaluates a mathematical expression responsive to interpretation of the user instructions by the first node; and

distributes calls to at least one other node of the plurality of nodes for execution of a second mathematical expression responsive to communication from the first node.

Assignments (1)

Continuity (14)

References Cited (400)