Data processing apparatus for pipeline execution acceleration and method thereof

• US 9,804,903 B2
• Filed: 02/09/2015
• Issued: 10/31/2017
• Est. Priority Date: 03/04/2014
• Status: Active Grant

• Current Assignee: Electronics and Telecommunications Research Institute
• Original Assignee: Electronics and Telecommunications Research Institute
• Inventors: Kim, Kang-Ho, Koh, Kwang-Won, Jeon, Seung-Hyub
• Primary Examiner(s): Ho, Andy
• Assistant Examiner(s): SEYE, ABDOU K
• Application Number:
  US14/616,894
• Publication Number:
  US 20150254116A1
• Time in Patent Office: 995 Days
• Field of Search: None
• CPC Class Codes: G06F 9/54 Interprogram communication
• CPC Class Codes: G06F 9/544 Buffers; Shared memory; Pipes
• CPC Class Codes: G06F 9/546 Message passing systems or structures, e.g. queues

Abstract

Disclosed herein are a data processing apparatus for pipeline execution acceleration and a method thereof. According to an exemplary embodiment of the present invention, the data processing apparatus for pipeline execution acceleration includes: a processor configured to sequentially execute a first application program and a second application program reading or writing a specific file; and a file system configured to complete a write for a file data for the specific file to a data block previously allocated from the first application program and provide the file data for the specific file to the second application program prior to executing a close call for the specific file from the first application program, when executing a read call for the specific file from the second application program.

Claims

1. A data processing apparatus for pipeline execution acceleration, comprising:

a processor configured to;

execute a first application program writing file data to a specific file, and

execute a second application program reading the file data from the specific file; and

a file system communicationally coupled to the processor and configured as a virtual ring type storage space in form of a circular queue of data blocks to store at least some of the file data for the specific file to cause the at least some of the file data be readable from a data block among the data blocks by the second application program prior to executing a close call for the writing of the file data to for the specific file from the first application program,

wherein the file system is further configured to,

set for the circular queue a queue_tail that represents a position for reading the file data and set a queue_head that represents a position for writing a file data,

perform a setting of a queue_index for a read that represents a start position of the circular queue at a time of executing the read from the specific file by the second application program to read file data having a size from the set queue_index and in response to the read, change the queue_tail of the circular queue,

the setting of the queue_index for the read includes to,

check, in a read request message, for an offset, which represents a value up to a position spaced apart from a previous start position of the specific file previously stored for the circular queue, and for a size, which represents a size of the file data to be read, and

set the queue_index using the offset when a size of the circular queue is larger than the size included in the read request message and the specific file is closed.

2. The data processing apparatus of claim 1, wherein the queue_tail is obtained by the following equation (queue_tail=(queue_tail+size) modulo queue_size), in which the queue_tail represents the queue_tail within the circular queue, the size represents the size of the file data to be read, and the queue_size represents a size of the circular queue.

3. The data processing apparatus of claim 1, wherein the file system is further configured to,

set a queue_index for the writing representing the start position of the circular queue for the writing at a time of executing the writing from the second application program, and

write the file data having a size from the set queue_index and in response to the write, change the queue_head of the circular queue.

4. The data processing apparatus of claim 3, wherein to set the queue_index for the writing includes,

checking, in a write request message, for an offset, which represents a value up to a position spaced apart from a previous start position of the specific file previously stored within the circular queue and for a size, which represents a size of the file data to be written, and

setting the queue_index using the offset when the offset is included in an empty section of the circular queue and an available space of the circular queue is larger than the size.

5. The data processing apparatus of claim 3, wherein the queue_head is obtained by the following equation (queue_head=(queue_head+size) modulo queue_size), wherein the queue_head represents the queue_head within the circular queue, the size represents the size of the file data to be written, and the queue_size represents a size of the circular queue.

6. The data processing apparatus of claim 1, the file system is further configured to:

present an application programming interface (API) configured to perform write and read operations of the specific file according to the circular queue at the time of executing the read and the write for the specific file;

update previously generated call information for the read and the write according to the circular queue at the time of executing the read and the write for the specific file and provide the call information to allow the API to perform the read and write operations for the specific file; and

manage the file data for the specific file and read or write the file data according to the circular queue at the time of executing the read and the write from the first and second application programs.

7. A data processing method for pipeline execution acceleration, comprising:

executing, by a processor, a first application program writing file data to a specific file;

executing, by the processor, a second application program reading the file data from the specific file; and

providing, by a file system, a virtual ring type storage space configured in form of a circular queue of data blocks that stores at least some of the file data for the specific file to cause the at least some of the file data for the specific file be readable from a data block among the data blocks by the second application program prior to executing a close call for the writing of the file data to the data for the specific file from the first application program,

wherein by the file system,

set for the circular queue a queue_tail that represents a position for reading the file data and set a queue_head that represents a position for writing the file data,

perform a setting of a queue_index for a read that represents a start position of the circular queue at a time of executing the read from the specific file by the second application program to read file data having a size from the set queue_index and in response to the read, change the queue_tail of the circular queue,

the setting of the queue_index for the read includes to,

check, in a read request message, for an offset, which represents a value up to a position spaced apart from a previous start position of the specific file previously stored for the circular queue, and for a size, which represents a size of the file data to be read, and

set the queue_index using the offset when a size of the circular queue is larger than the size included in the read request message and the specific file is closed.

8. The data processing method of claim 7, wherein the queue_tail is obtained by the following equation (queue_tail=(queue_tail+size) modulo queue_size), in which the queue_tail represents the queue_tail within the circular queue, the size represents the size of the file data to be read, and the queue_size represents a size of the circular queue.

9. The data processing method of claim 7, wherein by the file system,

setting a queue_index for the writing representing the start position of the circular queue for the writing at a time of executing the writing from the second application program, and

writing the file data having a size from the set queue_index and in response to the writing, changing the queue_head of the circular queue.

10. The data processing method of claim 9, wherein to set the queue_index for the writing includes,

checking, in a write request, for an offset, which represents a value up to a position spaced apart from a previous start position of the specific file previously stored within the circular queue and for a size, which represents a size of the file data to be written, and

setting the queue_index is set using the offset when the offset is included in an empty section of the circular queue and an available space of the circular queue is larger than the size.

11. The data processing method of claim 9, wherein the queue_head is obtained by the following equation (queue_head=(queue_head+size) modulo queue_size), in which the queue_head represents the queue_head within the circular queue, the size represents the size of the file data to be written, and the queue_size represents a size of the circular queue.