# US 2026/0299948 A1 · Descriptor-based hardware prefetcher

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

- Number: US 2026/0299948 A1
- Document ID: US20260299948A1
- Status: Published application
- Kind: A1
- Application number: 19/093,937
- Filed: 2025-03-28
- Published: 2026-10-01
- Earliest claimed: 2025-03-28
- Assignees: Advanced Micro Devices, Inc.
- Applicants: Advanced Micro Devices, Inc.
- Inventors: Onur Kayiran; Gabriel H. Loh; Rishabh Jain; Masab Ahmad
- Web page: https://patentgrep.com/patent/US20260299948A1

## CPC

- G06F 9/30047: {Prefetch instructions; cache control instructions}
- G06F 9/30065: {Loop control instructions; iterative instructions, e.g. LOOP, REPEAT}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A processor prefetches data based on access-pattern descriptors that delineate the expected storage patterns for data stored at a memory. An executing program, such as a machine learning model, generates the access-pattern descriptors to indicate the expected storage patterns. A hardware prefetcher of the processor employs the access-pattern descriptors to generate prefetch requests for the data, and cache of the processor executes the prefetch requests to transfer data to the cache for subsequent retrieval and use by the executing program. By employing the hardware prefetcher to generate the prefetch requests based on the access-pattern descriptors, the processing system reduces memory latency associated with more complex data structures, thereby improving overall processing efficiency

## Claims

### Claim 1

A method comprising:
generating, at a hardware prefetcher of a processing unit, a plurality of prefetch requests based on a descriptor indicating a memory access pattern; and
prefetching data to a cache based on the plurality of prefetch requests.

### Claim 2 (depends on 1)

The method of claim 1, wherein the descriptor indicates a plurality of loop sizes for N different dimensions of memory access, where N is an integer greater than one.

### Claim 3 (depends on 1)

The method of claim 1, wherein the descriptor indicates a plurality of stride lengths for N different dimensions of memory access, where N is an integer greater than one.

### Claim 4 (depends on 1)

The method of claim 1, wherein the descriptor indicates a last address to be prefetched to the cache for the memory access pattern.

### Claim 5 (depends on 1)

The method of claim 1, wherein the descriptor indicates a pooling factor for a machine learning model to be executed at the processing unit.

### Claim 6 (depends on 1)

The method of claim 1, wherein the descriptor indicates an embedding dimension for a machine learning model to be executed at the processing unit.

### Claim 7 (depends on 1)

The method of claim 1, further comprising:
retrieving the descriptor from memory based on a prefetch instruction including a pointer to the descriptor, wherein the prefetch instruction indicates a size of the descriptor based on a number of dimensions associated with the memory access pattern.

### Claim 8 (depends on 1)

The method of claim 1, wherein the descriptor indicates N dimensions of the memory access pattern, wherein N is an integer greater than two.

### Claim 9 (depends on 1)

The method of claim 1, wherein the descriptor indicates a plurality of loop sizes for N different dimensions of memory access, and wherein at least one of the plurality of loop sizes is based on a corresponding stride length indicated by the descriptor.

### Claim 10 (depends on 1)

The method of claim 1, further comprising:
adjusting a rate of issuance of subsequent prefetch requests at the hardware prefetcher based on measuring a timeliness of the plurality of prefetch requests relative to a corresponding plurality of demand requests at the processing unit.

### Claim 11

A processing unit, comprising:
a cache;
a hardware prefetcher configured to generate a plurality of prefetch requests based on a descriptor indicating a memory access pattern; and
wherein the cache is configured to prefetch data based on the plurality of prefetch requests.

### Claim 12 (depends on 11)

The processing unit of claim 11, wherein the descriptor indicates a plurality of loop sizes for N different dimensions of memory access, where N is an integer greater than one.

### Claim 13 (depends on 11)

The processing unit of claim 11, wherein the descriptor indicates a plurality of stride lengths for N different dimensions of memory access, where N is an integer greater than one.

### Claim 14 (depends on 11)

The processing unit of claim 11, wherein the descriptor indicates a last address to be prefetched to the cache for the memory access pattern.

### Claim 15 (depends on 11)

The processing unit of claim 11, wherein the descriptor indicates a pooling factor for a machine learning model to be executed at the processing unit.

### Claim 16 (depends on 11)

The processing unit of claim 11, wherein the descriptor indicates an embedding dimension for a machine learning model to be executed at the processing unit.

### Claim 17 (depends on 11)

The processing unit of claim 11, wherein the hardware prefetcher is configured to:
retrieve the descriptor from memory based on a prefetch instruction including a pointer to the descriptor, wherein the prefetch instruction indicates a size of the descriptor based on a number of dimensions associated with the memory access pattern.

### Claim 18 (depends on 11)

The processing unit of claim 11, wherein the descriptor indicates N dimensions of the memory access pattern, wherein N is an integer greater than two.

### Claim 19 (depends on 18)

The processing unit of claim 18, wherein the descriptor indicates a first loop size for a first dimension of memory access and indicates a first stride length for a second dimension of memory access, and wherein the first loop size is based on the first stride length.

### Claim 20

A processing system comprising:
a memory; and
a processing unit comprising:
a cache;
a hardware prefetcher configured to generate a plurality of prefetch requests based on a descriptor indicating a memory access pattern; and
wherein the cache is configured to prefetch data based on the plurality of prefetch requests.
