• Proceedings of the Korea Information Processing Society Conference
• /
• 2010.11a
• /
• pp.1732-1734
• /
• 2010

멀티 코어 또는 매니 코어 기반의 HPC 시스템 보급이 늘어나면서 HPC 어플리케이션이 사용하는 프로세스의 수 또한 증가하고 있다. 이런 경우, 기존의 IO 방식이 아닌 병렬 IO 의 사용을 고려하여야 하는데 그 중 특히 집합 IO 는 중요한 역할을 수행한다. 본 연구에서는 IBM Power 595 기반 슈퍼 컴퓨터에서 집합 IO 특성을 알아 본다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2020.11a
• /
• pp.98-101
• /
• 2020

누리온 시스템은 한국과학기술정보연구원(KISTI) 국가슈퍼컴퓨팅센터에서 2018년 5번째로 구축하여 운영하고 있는 슈퍼컴퓨터이다. 본 논문에서는 전산구조공학 분야에서 많이 활용되고 있는 ABAQUS, NASTRAN 등과 같은 응용소프트웨어를 누리온 시스템에서 활용하기 위한 방법을 소개하고 간단한 전산구조공학 모델을 누리온 시스템과 기존에 운영되었던 슈퍼컴퓨터 4호기 신바람 시스템에서 ABAQUS를 이용하여 분석하고 성능을 비교한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.11a
• /
• pp.118-121
• /
• 2021

슈퍼컴퓨터는 수백~수천 노드의 컴퓨팅 자원이 연결되어 복잡한 계산이나 대규모 데이터를 병렬 처리하며 일부 노드에서 발생하는 예상치 못한 문제는 전체적인 시스템 성능 저하로 이어질 수 있기 때문에 슈퍼컴퓨터 모니터링과 성능 분석은 슈퍼컴퓨터를 구축하고 운영하는데 필수적인 요소로 볼 수 있다. 본 논문에서는 오픈소스 소프트웨어인 Performance Co-Pilot, Bpftrace, Grafana를 활용한 슈퍼컴퓨터 모니터링 및 성능분석 시스템 구축 방안을 제안하였으며 이를 통해서 확장가능하면서도 유연한 구조의 슈퍼컴퓨터 모니터링 및 성능 분석이 가능함을 보였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.11a
• /
• pp.6-8
• /
• 2021

슈퍼컴퓨터 5호기 누리온은 8,437대의 계산노드와 33.88PB 규모의 병렬스토리지가 100Gbps의 Omni-Path(OPA) 인터커넥트로 연결되어 있는 초대형 클러스터 시스템이다. 누리온의 계산자원은 PBS 작업스케줄러를 통해 관리되고 있고 병렬 스토리지는 DDN사의 Exascaler Monitoring System(ESMON)을 통해 influxDB에 read/write IO 통계 데이터를 수집하고 있다. 본 논문에서는 PBS의 과금 데이터와 ESMON influxDB의 IO 통계 데이터를 활용하여 사용자의 작업별 IO 통계 정보를 생성하는 방안에 대하여 소개한다.

• Journal of Technology Innovation
• /
• v.23 no.2
• /
• pp.33-54
• /
• 2015

Supercomputing performance demand forecasting at the national level is an important information to the researchers in fields of the computational science field, the specialized agencies which establish and operate R&D infrastructure, and the government agencies which establish science and technology infrastructure. This study derived the factors affecting the scientific and technological capability through the analysis of supercomputing performance prediction research, and it proposed a hybrid forecasting model of applying the super-computer technology trends. In the cross-sectional analysis, multiple regression analysis was performed using factors with GDP, GERD, the number of researchers, and the number of SCI papers that could affect the supercomputing performance. In addition, the supercomputing performance was predicted by multiplying in the cross-section analysis with technical progress rate of time period which was calculated by time series analysis using performance(Rmax) of Top500 data. Korea's performance scale of supercomputing in 2016 was predicted using the proposed forecasting model based on data of the top500 supercomputer and supercomputing performance demand in Korea was predicted using a cross-sectional analysis and technical progress rate. The results of this study showed that the supercomputing performance is expected to require 15~30PF when it uses the current trend, and is expected to require 20~40PF when it uses the trend of the targeting national-level. These two results showed significant differences between the forecasting value(9.6PF) of regression analysis and the forecasting value(2.5PF) of cross-sectional analysis.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2016.05a
• /
• pp.554-557
• /
• 2016

Recently, supercomputer has been increasingly adopted as a computing environment for scientific simulation as well as education, healthcare and national defence. Especially, supercomputing system with heterogeneous computing resources is gaining resurgence of interest as a next-generation problem solving environment, allowing theoretical and/or experimental research in various fields to be free of time and spatial limits. However, traditional supercomputing services have only been handled through a simple form of command-line based console, which leads to the critical limit of accessibility and usability of heterogeneous computing resources. To address this problem, in this paper, we provide the design and implementation of web-based HPC (High Performance Computing) job management framework for computational scientific simulation. The proposed framework has highly extensible design principles, providing the abstraction interfaces of job scheduler (as well as bundle scheduler plug-ins for LoadLeveler, Sun Grid Engine, OpenPBS scheduler) in order to easily incorporate the broad spectrum of heterogeneous computing resources such as cluster, computing cloud and grid. We also present the detailed specification of HTTP standard based RESTful endpoints, which manage simulation job's life-cycles such as job creation, submission, control and status monitoring, etc., enabling various 3rd-party applications to be newly created on top of the proposed framework.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• v.10 no.2
• /
• pp.81-88
• /
• 2006

The advent of supercomputers contribute greatly in overcoming scientific and academic problems that were previously difficult to solve. However, the supercomputer itself suffers from the problem of being considerable cost. In response, the concept of grid computing, to use the resources of distribute computers connected with each other, was created. This system uses connection oriented protocols to integrate and manage the resources of different types of distributed systems, yet it has the problem of compatibility between protocols of other types. In this paper, a system to manage grid network performance through XML-based SOAP web service which is platform-independent, is proposed.

• Proceedings of the Korean Society For Composite Materials Conference
• /
• 2005.04a
• /
• pp.5-8
• /
• 2005

In this paper, the electromechanical displacements of curved piezoelectric actuators composed of PZT ceramic and laminated composite materials are calculated based on high performance computing technology and the optimal configuration of composite curved actuator is examined. To accurately predict the local pre-stress in the device due to the mismatch in coefficients of thermal expansion, carbon-epoxy and glass-epoxy as well as PZT ceramic are numerically modeled by using hexahedral solid elements. Because the modeling of these thin layers increases the number of degrees of freedom, large-scale structural analyses are performed through the PEGASUS supercomputer, which is installed in our laboratory. In the first stage, the curved shape of the actuator and the internal stress in each layer are obtained by the cured curvature analysis. Subsequently, the displacement due to the piezoelectric force (which is resulted from applied voltage) is also calculated. The performance of composite curved actuator is investigated by comparing the displacements obtained by the variation of thickness and elastic modulus of laminated composite layers. In order to consider the finite deformation in the first analysis stage and include the pre-stress due to curing process in the second stage, nonlinear finite element analyses are carried out.

• Journal of Information Processing Systems
• /
• v.2 no.2
• /
• pp.101-106
• /
• 2006

Grid computing is an emerging technology that enables global resource sharing. In Korea, the $K^*$Grid provides an extremely powerful research environment to both industries and academia. As part of the $K^*$Grid project, we have constructed, together with the Korea Institute of Science and Technology Information and a number of domestic universities, a supercomputer Grid test bed which connects several types of supercomputers based on the globus toolkit. To achieve efficient networking in this Grid testbed, we propose a novel method of available bandwidth measurement, called Decoupled Capacity measurement with Initial Gap (DCIG), using packet trains. DCIG can improve the network efficiency by selecting the best path among several candidates. Simulation results show that DCIG outperforms previous work in terms of accuracy and the required measurement time. We also define a new XML schema for DCIG request/response based on the schema defined by the Global Grid Forum (GGF) Network Measurement Working Group (NM-WG).

• Proceedings of the Korean Information Science Society Conference
• /
• 2010.06b
• /
• pp.294-298
• /
• 2010

슈퍼컴퓨터는 스케줄러를 이용하여 사용자의 작업(task) 실행 및 자원할당 요청을 관리하고, 사용자는 배치 프로세스 형태로 스케줄러에 작업을 제출한다. 스케줄러에 배치 프로세스 형태로 작업을 제출하는 방식은 사용자의 작업이 비동기적으로 수행되어 작업의 상태가 변경될 경우 확인할 수 있는 서비스를 필요로 한다. 이를 위해 대부분의 슈퍼컴퓨터는 작업이 완료되거나 오류가 발생하여 중단되었을 때 전자메일을 발송하여 사용자에게 정보를 제공하는 이벤트 알림 서비스를 제공하지만, 장기간 수행되는 작업의 경우 주기적으로 전자메일을 확인해야 하는 불편함이 있고, 다수의 작업을 동시에 수행하는 경우 각각의 이벤트 정보를 포함한 전자메일을 관리하는데 어려움이 따른다. 본 논문에서는 슈퍼컴퓨터에서 제공하는 전자메일 기반 이벤트 알림 서비스를 이용하여 이벤트를 통지받을 때 지정된 방식으로 후처리 작업이 실행될 수 있는 자동화된 후처리 시스템의 설계 및 구현 방법을 제시한다.