# US 8,279,865 B2 · Efficient pipeline parallelism using frame shared memory

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

- Number: US 8,279,865 B2
- Document ID: US8279865B2
- Status: Granted
- Kind: B2
- Application number: 12/106,684
- Filed: 2008-04-21
- Published: 2009-02-26
- Granted: 2012-10-02
- Earliest claimed: 2007-04-20
- Inventors: John Giacomoni; Manish Vachharajani
- Web page: https://patentgrep.com/patent/US8279865B2

## CPC

- G06F 12/0846: Cache with multiple tag or data arrays being simultaneously accessible
- G06F 9/544: {Buffers; Shared memory; Pipes}
- G06F 12/128: adapted to multidimensional cache systems, e.g. set-associative, multicache, multiset or multilevel
- G06F 12/0806: Multiuser, multiprocessor or multiprocessing cache systems
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A systems and methods are disclosed that provide an efficient parallel pipeline for data processing using a multi-core processor. Embodiments allocate a shared memory portion of the memory that is accessible from more than one context of execution and/or process a frame in a plurality of processing stages processed by a context of execution. In some embodiments, each of the plurality of processing stages may be bound to a processing core of the multi-core processor. In other embodiments include one or more processing stages with a point-to-point communication mechanism that operates in shared memory.

## Claims

### Claim 1

A computer system comprising:
a general purpose multi-core processor including a plurality of processing cores; and
memory communicably coupled with the general purpose multi-core processor, wherein the memory includes a plurality of instructions executed by the general purpose multi-core processor, the plurality of instructions comprising:
instructions executed by the general purpose multi-core processor to allocate a shared memory portion of the memory, wherein the shared memory is accessible from more than one context of execution;
instructions executed by the general purpose multi-core processor to process a frame in a plurality of processing stages comprising at least a first processing stage executed in kernel space and a second processing stage executed in user space, wherein the first processing stage and the second processing stage are processed by separate contexts of execution;
instructions executed by the general purpose multi-core processor to bind each of the plurality of processing stages to a processing core of the multi-core processor; and
instructions executed by the general purpose multi-core processor to connect one or more processing stages with a point-to-point communication mechanism that allows the first processing stage and the second processing stage to exchange data related to the frame through mutual access to the shared memory portion of the memory.

### Claim 2 (depends on 1)

The computer system according to claim 1, wherein a context of execution includes at least one of user space and kernel space.

### Claim 3 (depends on 1)

The computer system according to claim 1, wherein the memory further includes instructions to extend the processing time of stages with a short processing interval.

### Claim 4 (depends on 1)

The computer system according to claim 1, wherein the memory further includes:
instructions to receive a frame through a network interface device communicatively coupled with the general purpose multi-core processor; and
instructions to process the frame in each processing stage, wherein the processing time for each stage is about the frame arrival period of the frame.

### Claim 5 (depends on 1)

The computer system according to claim 1, wherein the point-to-point communication mechanism includes instructions to use a head pointer without a tail pointer to enqueue.

### Claim 6 (depends on 1)

The computer system according to claim 1, wherein the point-to-point communication mechanism includes instructions to use a tail pointer without a head pointer to dequeue.

### Claim 7

A method for establishing parallel frame processing, the method comprising:
receiving a plurality of frames at a general purpose multi-core processor that operates contexts of execution;
segmenting frame processing into a plurality of stages comprising at least a first stage in kernel space and a second stage in user space, wherein the processing of at least one stage occurs at the general purpose multi-core processor in at least one of the contexts of execution;
allocating a shared portion of memory for access by both the first stage and the second stage;
associating each of the plurality of stages to a processing core corresponding to a processing core of the general purpose multi-core processor; and
associating stages with a point-to-point communication mechanism, wherein the point-to-point communication mechanism uses the shared portion of the memory to communicate data related to the plurality of frames between at least the first stage and the second stage.

### Claim 8 (depends on 7)

The method according to claim 7, wherein a context of execution operates in at least one of user space and kernel space.

### Claim 9 (depends on 7)

The method according to claim 7, further comprising processing at least two stages of the plurality of stages in parallel on different frames wherein the at least two stages provide the same processing.

### Claim 10 (depends on 7)

The method according to claim 7, wherein at least one stage of the plurality of stages comprises a stage selected from the group consisting of: an input stage; a decryption stage, an application stage, an encryption stage, and an output stage.

### Claim 11 (depends on 7)

The method according to claim 7, further comprising processing each stage within the inter-frame arrival period of the frames.

### Claim 12 (depends on 7)

The method according to claim 7, wherein the point-to-point communication mechanism is a low latency mechanism.

### Claim 13 (depends on 7)

The method according to claim 7, further comprising maintaining substantially consistent frame arrival rates.

### Claim 14 (depends on 13)

The method according to claim 13, further comprising extending the processing time of stages with a short processing interval.

### Claim 15

A method for sequentially processing a plurality of frames comprising:
allocating shared memory in a general purpose multi-core processor that operates contexts of execution, wherein the shared memory is accessible by multiple contexts of execution, and wherein the multi-core processor includes a first processing core and a second processing core;
receiving a first frame at the general purpose multi-core processor;
placing the first frame in the shared memory;
processing by the general purpose multi-core processor the first frame with the first processing core during a first time period, the processing of the first frame with the first processing core comprising a first stage in kernel space accessing data related to the first frame in the shared memory;
receiving a second frame at the general purpose multi-core processor after the first frame is received;
placing the second frame in the shared memory;
processing by the general purpose multi-core processor the first frame with the second processing core during a second time period, wherein the second time period occurs after the first time period, the processing of the first frame with the second processing core comprising a second stage in user space accessing the data related to the first frame in the shared memory; and
processing by the general purpose multi-core processor the second frame with the first processing core during the second time period, the processing of the second frame with the second processing core comprising a second stage in user space accessing data related to the second frame in the shared memory.

### Claim 16 (depends on 15)

The method according to claim 15, wherein the shared memory is accessible from a first context of execution using a first set of pointers and is accessible from a second context of execution using a second set of pointers.

### Claim 17 (depends on 15)

The method according to claim 15, further comprising reading frame data from the shared memory.
