# US 2012/0254542 A1 · Gather cache architecture

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

- Number: US 2012/0254542 A1
- Document ID: US20120254542A1
- Status: Published application
- Kind: A1
- Application number: 13/078,380
- Filed: 2011-04-01
- Published: 2012-10-04
- Earliest claimed: 2011-04-01
- Assignees: Intel Corporation
- Inventors: Shlomo Raikin; Robert Valentine
- Web page: https://patentgrep.com/patent/US20120254542A1

## CPC

- G06F 12/0815: Cache consistency protocols
- G06F 12/0804: with main memory updating (G06F12/0806 takes precedence)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Apparatuses and methods to perform gather instructions are presented. In one embodiment, an apparatus comprises a gather logic module which includes a gather logic unit to identify locality of data elements in response to a gather instruction. The apparatus includes memory comprising a plurality of memory rows including a memory row associated with the gather instruction. The apparatus further includes memory structure to store data element addresses accessed in response to the gather instruction.

## Claims

### Claim 1

An apparatus comprising a processor operable to perform one or more vector operations, wherein the processor comprises
a gather logic module comprising a first gather logic unit to identify locality of data elements in response to a first gather instruction;
a first memory comprising a plurality of memory rows including a first memory row associated with the first gather instruction; and
a second memory to store one or more data element addresses accessed in response to the first gather instruction.

### Claim 2 (depends on 1)

The processor of claim 1, wherein the processor further comprises a memory execution unit, coupled to the first memory, to manage gather instructions including the first gather instruction and to write back only once a result of the first gather instruction.

### Claim 3 (depends on 1)

The processor of claim 1, wherein the first gather logic unit is operable to determine data locality based at least on cache line locality.

### Claim 4 (depends on 1)

The processor of claim 1, wherein the first gather logic unit is operable to access the first memory row in a manner of a content-addressable memory for identifying data elements that reside in a same cache line.

### Claim 5 (depends on 1)

The processor of claim 1, wherein the first gather logic unit is operable to
iteratively identify data elements that reside in a same cache line;
identify offsets of the data elements;
generate control information to a data cache rotator based on the offsets; and
send one or more load operations until all non-masked data elements are gathered.

### Claim 6 (depends on 1)

The processor of claim 1, wherein the first gather logic unit is operable to schedule one write-back cycle to an execution unit of the processor in response to the first gather instruction.

### Claim 7 (depends on 1)

The processor of claim 1, wherein the first memory row is capable of storing architectural information associated with the first gather instruction including status of hardware resources.

### Claim 8 (depends on 1)

The processor of claim 1, wherein the first memory row is capable of storing information associated with the first gather instruction including blocked hardware resources required by the first gather instruction.

### Claim 9 (depends on 1)

The processor of claim 1, wherein the first gather logic unit is operable to set or to clear a snoop indication associated with the first gather instruction and to determine whether to retire the first gather instruction based on the snoop indication.

### Claim 10

A method comprising:
identifying locality of data elements in response to a first gather instruction, wherein information about the first gather instruction is stored in a first memory row of a first memory; and
storing one or more data element addresses accessed in response to the first gather instruction.

### Claim 11 (depends on 10)

The method of claim 10, further comprising accessing the first memory row in a manner of a content-addressable memory for identifying data elements that reside in a same cache line.

### Claim 12 (depends on 10)

The method of claim 10, further comprising:
identifying data elements that reside in a same cache line;
identifying offsets of the data elements;
generating control information to a data cache rotator based on the offsets; and
sending one or more load operations until all non-masked data elements are gathered.

### Claim 13 (depends on 10)

The method of claim 10, further comprising scheduling one write-back cycle to an execution unit of the processor in response to the first gather instruction.

### Claim 14 (depends on 10)

The method of claim 10, further comprising:
setting or clearing a snoop indication associated with the first gather instruction; and
determining whether to retire the first gather instruction based on the snoop indication.

### Claim 15

A system comprising:
a data cache; and
a processor comprising
a first gather logic unit to identify locality of data elements in response to a first gather instruction;
a first memory comprising a plurality of memory rows including a first memory row associated with the first gather instruction; and
a second memory to store one or more data element addresses accessed in response to the first gather instruction; and
a memory execution unit, coupled to the first memory, to manage gather instructions including the first gather instruction and to write back only once a result of the first gather instruction.

### Claim 16 (depends on 15)

The system of claim 15, wherein the first gather logic unit is operable to
iteratively identify data elements that reside in a same cache line;
identify offsets of the data elements;
generate control information to a data cache rotator based on the offsets; and
send one or more load operations until all non-masked data elements are gathered.

### Claim 17 (depends on 15)

The system of claim 15, wherein the first gather logic unit is operable to schedule one write-back cycle to an execution unit of the processor in response to the first gather instruction.

### Claim 18 (depends on 15)

The system of claim 15, wherein the first gather logic unit is operable to set or to clear a snoop indication associated with the first gather instruction and to determine whether to retire the first gather instruction based on the snoop indication.
