Search International and National Patent Collections
Some content of this application is unavailable at the moment.
If this situation persists, please contact us atFeedback&Contact
1. (KR1020150088045) METHOD FOR PERFORMING PARALLEL PROGRAMMING IN MANY-CORE CLUSTER SYSTEM AND MANY-CORE CLUSTER SYSTEM

Office : Republic of Korea
Application Number: 1020140008392 Application Date: 23.01.2014
Publication Number: 1020150088045 Publication Date: 31.07.2015
Publication Kind : A
IPC:
G06F 9/38
G06F 9/46
G PHYSICS
06
COMPUTING; CALCULATING; COUNTING
F
ELECTRIC DIGITAL DATA PROCESSING
9
Arrangements for programme control, e.g. control unit
06
using stored programme, i.e. using internal store of processing equipment to receive and retain programme
30
Arrangements for executing machine- instructions, e.g. instruction decode
38
Concurrent instruction execution, e.g. pipeline, look ahead
G PHYSICS
06
COMPUTING; CALCULATING; COUNTING
F
ELECTRIC DIGITAL DATA PROCESSING
9
Arrangements for programme control, e.g. control unit
06
using stored programme, i.e. using internal store of processing equipment to receive and retain programme
46
Multiprogramming arrangements
CPC:
G06F 9/505
G06N 99/005
H04L 67/10
H04L 67/1004
Applicants: 서울대학교산학협력단
SNU RDB FOUNDATIONSNU RDB FOUNDATION
Inventors: LEE, JAE JINLEE, JAE JIN
KIM, JUNG WONKIM, JUNG WON
이재진
김정원
Agents: 특허법인 아이스퀘어
Priority Data:
Title: (EN) METHOD FOR PERFORMING PARALLEL PROGRAMMING IN MANY-CORE CLUSTER SYSTEM AND MANY-CORE CLUSTER SYSTEM
(KO) 매니코어 클러스터 시스템 상에서 병렬 프로그래밍을 수행하는 방법 및 매니코어 클러스터 시스템
Abstract:
(EN) The present invention relates to a method for performing a parallel programming in a many-core cluster system and a many-core cluster system, which can operate an OpenCL application written for one calculation device in a heterogeneous many-core cluster system composed of multiple calculation devices without any alteration of the source. The method for performing a parallel programming in a many-core cluster system including multiple nodes, each of which includes at least one calculation device, comprises the steps of: executing the parallel programming application at at least one node; estimating an amount of processed data of at least one calculation device included in the multiple nodes from the at least one node; and distributing the workload involved in the execution of the parallel programming application to at least one calculation device included in the multiple nodes according to the estimated amount of the processed data of the calculation device. Conventionally, a programmer should manually calculate and properly distribute the workload according to the processed amount of each calculation device. However, according to an embodiment of the present invention, such barrier to programming can be eliminated, thereby providing a heterogeneous many-core cluster environment which can be readily used by a programmer. COPYRIGHT KIPO 2015
(KO) 본 발명의 일 실시예에 따르는 매니코어 클러스터 시스템 상에서 병렬 프로그래밍을 수행하는 방법 및 매니코어 클러스터 시스템은 한 개의 계산 디바이스를 위해 작성된 OpenCL 어플리케이션을, 소스 수정 없이 다수 개의 계산 디바이스로 구성된 이종 매니코어 클러스터 시스템 상에서 동작시킬 수 있다. 종래에는 프로그래머가 각 계산 디바이스의 처리량에 따라 수동적으로 워크로드의 양을 계산 디바이스에 적절히 배분해야 했으나, 본 발명의 일 실시예는 이에 대한 프로그래밍 장벽을 제거할 수 있으므로, 프로그래머에게 용이한 이종 매니코어 클러스터 환경을 제공할 수 있다.