# US 2022/0276803 A1 · Near-Memory Computing Systems And Methods

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2022/0276803 A1
- Document ID: US20220276803A1
- Status: Published application
- Kind: A1
- Application number: 17/187,483
- Filed: 2021-02-26
- Published: 2022-09-01
- Earliest claimed: 2021-02-26
- Assignees: Petaio Inc.
- Applicants: PetaIO Inc.
- Inventors: Fan Yang; Peirong Ji; Changyou Xu
- Web page: https://patentgrep.com/patent/US20220276803A1

## CPC

- G06F 3/0656: {Data buffering arrangements}
- G06F 9/54: Interprogram communication
- G06F 3/0613: {in relation to throughput}
- G06F 3/0659: {Command handling arrangements, e.g. command buffers, queues, command scheduling}
- G06F 3/0683: {Plurality of storage devices}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Example near-memory computing systems and methods are described. In one implementation, a system includes a host command processing system and a computational engine associated with a solid-state drive. In some situations, the computational engine includes multiple versatile processing unit slices coupled to one another. The multiple versatile processing unit slices are configured to perform different tasks in parallel with one another. The system also includes a host direct memory access module configured to access memory devices independently of a central processing unit.

## Claims

### Claim 1

A near-memory computing system comprising:
a host command processing system;
a computational engine associated with a solid-state drive, wherein the computational engine includes a plurality of versatile processing unit slices coupled to one another, wherein the plurality of versatile processing unit slices are configured to perform different tasks in parallel with one another; and
a host direct memory access module configured to access memory devices independently of a central processing unit.

### Claim 2 (depends on 1)

The near-memory computing system of claim 1, further comprising a peripheral component interconnect interface coupled to the host command processing system and configured to provide high-speed communication with the host command processing system and the host direct memory access module.

### Claim 3 (depends on 1)

The near-memory computing system of claim 1, wherein the computational engine further includes an inter process communication hub coupled to the plurality of versatile processing unit slices.

### Claim 4 (depends on 3)

The near-memory computing system of claim 3, wherein the inter process communication hub is configured to deliver messages between multiple central processing units.

### Claim 5 (depends on 1)

The near-memory computing system of claim 1, wherein the plurality of versatile processing unit slices are coupled to an interconnect that allows the plurality of versatile processing unit slices to communicate with one another.

### Claim 6 (depends on 1)

The near-memory computing system of claim 1, further comprising a NAND flash controller configured to manage a plurality of NAND flash devices.

### Claim 7 (depends on 1)

The near-memory computing system of claim 1, wherein each of the plurality of versatile processing unit slices includes:
a local static random access memory; and
a plurality of field-programmable gate array/digital signal processing arrays.

### Claim 8 (depends on 7)

The near-memory computing system of claim 7, wherein each of the plurality of versatile processing unit slices further includes:
a central processing unit;
a memory device;
a direct memory access interface; and
a buffer management interface.

### Claim 9 (depends on 1)

The near-memory computing system of claim 1, wherein each of the plurality of versatile processing unit slices is configured to receive data from a host memory and store the data in a local static random access memory.

### Claim 10 (depends on 1)

The near-memory computing system of claim 1, wherein each of the plurality of versatile processing unit slices is configured to read data from a buffer management unit.

### Claim 11 (depends on 1)

The near-memory computing system of claim 1, wherein the plurality of versatile processing unit slices share a common data path, and wherein a quality of service arbiter optimizes access to the data path by the plurality of versatile processing unit slices.

### Claim 12 (depends on 1)

The near-memory computing system of claim 1, wherein at least two of the versatile processing unit slices cooperate with one another to complete a particular task.

### Claim 13 (depends on 1)

The near-memory computing system of claim 1, wherein each of the plurality of versatile processing unit slices includes a central processing unit configured to manage job scheduling by the particular versatile processing unit slice.

### Claim 14 (depends on 1)

The near-memory computing system of claim 1, wherein each of the plurality of versatile processing unit slices includes a static random access memory that includes a plurality of memory banks, and wherein each of the plurality of memory banks are coupled to at least one arbiter.

### Claim 15 (depends on 1)

The near-memory computing system of claim 1, wherein each of the plurality of versatile processing unit slices includes a static random access memory that includes a plurality of memory banks, and wherein the plurality of static random access memories are configured in at least one of a parallel mode, a serial mode, or a mixed mode.

### Claim 16

A near-memory computing system comprising:
a host command processing system; and
a computational engine associated with a solid-state drive, wherein the computational engine includes a plurality of versatile processing unit slices coupled to one another, wherein the plurality of versatile processing unit slices are configured to perform different tasks in parallel with one another, and wherein each of the plurality of versatile processing unit slices includes:
a local static random access memory; and
a plurality of field-programmable gate array/digital signal processing arrays.

### Claim 17 (depends on 16)

The near-memory computing system of claim 16, wherein the computational engine further includes an inter process communication hub coupled to the plurality of versatile processing unit slices, and wherein the inter process communication hub is configured to deliver messages between multiple central processing units.

### Claim 18 (depends on 16)

The near-memory computing system of claim 16, wherein the plurality of versatile processing unit slices are coupled to an interconnect that allows the plurality of versatile processing unit slices to communicate with one another.

### Claim 19 (depends on 16)

The near-memory computing system of claim 16, wherein each of the plurality of versatile processing unit slices further includes:
a central processing unit;
a memory device;
a direct memory access interface; and
a buffer management interface.

### Claim 20 (depends on 16)

The near-memory computing system of claim 16, wherein each of the plurality of versatile processing unit slices includes a central processing unit configured to manage job scheduling by the particular versatile processing unit slice.
