# US 8,279,886 B2 · Dataport and methods thereof

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

- Number: US 8,279,886 B2
- Document ID: US8279886B2
- Status: Granted
- Kind: B2
- Application number: 11/024,909
- Filed: 2004-12-30
- Published: 2006-07-06
- Granted: 2012-10-02
- Earliest claimed: 2004-12-30
- Assignees: Intel Corporation
- Inventors: Dinakar Munagala; Hong Jiang; Bishara Shomar; Val Cook; Michael K. Dwyer; Thomas Piazza
- Web page: https://patentgrep.com/patent/US8279886B2

## CPC

- G06F 13/1668: {Details of memory controller}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A context-free (stateless) dataport may allow multiple processors to perform read and write operations on a shared memory. The operations may include, for example, structured data operations such as image and video operations. The dataport may perform addressing computations associated with block memory operations. Therefore, the dataport may be able, for example, to relieve the processors that it serves from this duty. The dataport may be accessed using a message interface that may be implemented in a standard and generalized manner and that may therefore be easily transportable between different types of processors.

## Claims

### Claim 1

A system comprising:
a first processor;
a second processor;
a dataport, said processors coupled in parallel to the dataport; and
a shared memory shared by said processors, said dataport coupled to said shared memory, said dataport to perform memory access operations on said shared memory for each of said processors, each processor to send a message to the dataport and to specify a location where a response from said dataport is to be stored and to enable the thread to lock said location when the thread sends the message to the dataport, wherein said dataport further comprises:
an input bus to receive a message from at least one of said processors for executing one or more threads requesting an operation on said shared memory; a message decoder to determine solely from said message which steps are to be performed in order to execute said operation and to identify requester of said operation; and a data processor to perform said steps.

### Claim 2 (depends on 1)

The system of claim 1, wherein said operation includes writing data from said message into said shared memory.

### Claim 3 (depends on 1)

The system of claim 1, further comprising a write-back bus to transmit a write-back message, wherein said data processor generates said write-back message.

### Claim 4 (depends on 1)

The system of claim 1, wherein said data processor performs data-scattering, data-gathering and format conversion operations.

### Claim 5

A method comprising:
coupling each of two processors in parallel to a dataport, in turn coupled to a memory shared by said processors;
enabling said dataport to execute shared memory access operations for each of said processors;
enabling a thread to send a message to the dataport and to specify a location where a response from said dataport is to be stored;
enabling the thread to lock said location when the thread sends the message to the dataport; and
receiving a message in said dataport from two or more threads running on said processors requesting an operation on the shared memory and decoding said message to determine, solely from said message, which steps are to be performed in order to execute said operation and to identify a requester of said operation, and performing said steps by a data processor.

### Claim 6 (depends on 5)

The method of claim 5, wherein said operation includes writing data from said message into said shared memory.

### Claim 7 (depends on 5)

The method of claim 5, further comprising the step of transmitting a write-back message.

### Claim 8 (depends on 7)

The method of claim 7, wherein said operation includes reading data from said shared memory and inserting said data into said write-back message.

### Claim 9 (depends on 7)

The method of claim 7, wherein said write-back message contains a target identifier.

### Claim 10 (depends on 5)

The method of claim 5, further comprising the steps of: organizing a queue of messages; and suspending arrival of said message when said queue is full.

### Claim 11 (depends on 5)

The method of claim 5, wherein said operation refers to a data structure and said message includes attributes of said data structure.

### Claim 12 (depends on 5)

The method of claim 5, further comprising the steps of: processing said steps to be performed; computing addresses for accessing said shared memory; checking said steps and said addresses for errors; and accessing said shared memory using any one of a set of out-of-bounds addressing modes.

### Claim 13 (depends on 5)

The method of claim 5, further comprising performing data- scattering and data-gathering operations.
