[보고서]극한규모 슈퍼컴퓨팅 기반 과학·공학 분야 거대문제 해결·지원

염민선

[국가R&D연구보고서] 극한규모 슈퍼컴퓨팅 기반 과학·공학 분야 거대문제 해결·지원
Solving grand-challenge problems in science and engineering by using extreme scaling of supercomputing 원문보기

보고서 정보
주관연구기관	한국과학기술정보연구원 Korea Institute of Science and Technology Information
연구책임자	염민선
참여연구자	강궁원 , 강지훈 , 권오경 , 김지민 , 김한슬 , 류훈 , 문명환 , 박기홍 , 박정우 , 안설아 , 안영선 , 여인성 , 윤태호 , 이홍석 , 정민중 , 정영균 , 정희진 , 조기현 , 조덕연 , 조지호 , 최성환 , 한혜림 , 현영환 , Bui Khac Hoai Nam , 고영일 , 최해천
보고서유형	3단계보고서
발행국가	대한민국
언어	한국어
발행년월	2020-12
과제시작연도	2020
주관부처	과학기술정보통신부 Ministry of Science and ICT
연구관리전문기관	한국과학기술정보연구원 Korea Institute of Science and Technology Information
등록번호	TRKO202100005926
과제고유번호	1711121658
사업명	한국과학기술정보연구원연구운영비지원(R&D)(주요사업비)
DB 구축일자	2021-07-03
키워드	거대문제.슈퍼컴퓨팅.병렬처리.계산과학공학.매니코어 연산.Grand-challenge Problems.Supercomputing.Parallel Processing.Computational Science and Engineering.Manycore Computing.

초록 ▼

ㅇ 대규모 병렬처리 기반 초고성능컴퓨팅 기술 확보
- 12PF 대규모 양자수송 현상 해석을 위한 병렬처리 기술 개발
- 7.5PF FSI 해석을 위한 2기술 개발
- 슈퍼컴퓨터 5호기 누리온 버스트버퍼 성능사례 확보

ㅇ 초고성능컴퓨팅 기술 기반 과학·공학 초거대규모 문제 해결
- 12.5PF 대규모·고정밀 거대 계산 실현 및 열유동해석 문제 발굴
- 7.5PF 페로브스카이트 소재 기반의 LED 설계 연구
- 7.5PF 근우주의 은하 생성 및 진화 연구

ㅇ 과학·공학 분야 난제

Abstract ▼

IV. Results of the study
□ Securing ultra-high performance computing technology based on large-scale parallel processing
- Development of parallel processing technology for analyzing 12PF large-scale quantum transport phenomena
- Development of 2 technologies for analyzing 7.5PF FSI
- Acquisition of performance improvement example for the burst buffer of the Nurion supercomputer

□ Solving ultra grand-challenge problems in science and engineering based on ultra-high performance computing technology
- Realization of 12.5PF large scale calculation with high precision and descovery of heat flow analysis problem
- Research on LED design based on perovskite material
- Research on Galactic Generation and Evolution of 7.5PF Near Space

□ Discovering, solving and supporting group research for solving outstanding issue in science and engineering
- Research on enhancing light-emission rate of organic Perovskite
- Research on designing quantum logic gate device in silicon

(출처 : SUMMARY 7p)

목차 Contents

표지 ... 1
제 출 문 ... 2
보고서 초록 ... 3
요 약 문 ... 4
Summary ... 6
Contents ... 9
목차 ... 10
표목차 ... 11
그림목차 ... 12
제 1 장 연구개발 과제의 개요 ... 17
제 1 절 연구목적 및 필요성 ... 17
제 2 절 연구내용 및 범위 ... 19
제 3 절 추진전략 ... 33
제 2 장 국내외 기술개발 현황 분석 ... 35
제 1 절 국내 동향 ... 35
제 2 절 해외 동향 ... 38
제 3 장 연구개발 수행 내용 및 결과 ... 43
제 1 절 대규모 병렬처리 기반 초고성능컴퓨팅 기술 확보 ... 43
제 2 절 초고성능컴퓨팅 기술 기반 과학ž공학 초거대규모 문제 해결 ... 81
제 3 절 과학·공학 분야 난제 해결을 위한 집단 연구 문제 발굴 및 수행·지원 ... 88
제 4 절 지적 재산권 현황 ... 108
제 4 장 목표달성도 및 관련분야에의 기여 ... 118
제 1 절 목표달성도 ... 118
제 2 절 관련 분야에의 기여도 ... 120
제 5 장 연구개발결과의 활용계획 ... 122
제 6 장 참고문헌 ... 123
끝페이지 ... 126

표/그림 (90)

표 Comparison of parallel multigrid methods with different coarse-level processing. (a) Coarse-grid solver (b) Coarse-grid aggregation Coarse-grid partial semi-aggregation
표 Scheme of the Recursive Green’s Function method for a 3D system. The new Green’s function G(N+1) can be calculated from the old Hamiltonian HN (light grey), the new slice Hamiltonian HN+1 (dark grey) and the coupling tN (solid arrows)
표 Schematic of 2D domain decomposition of FSI solver
표 (a) Isosurfaces of the streamwise velocity fluctuations in the channel flow simulation code (b) Isosurfaces of the filtered streamwise velocity fluctuation
표 List of 2020 calls for collaboration of performance optimization
표 Recent simulation status and target of this simulation
표 1940 Tacoma Narrows Bridge collapses
표 Large-scale simulation for electric device design
표 Visualization of real-time sensing data for machine learning system
표 Convergence research on high energy physics using experiment-theory-simulation
표 The chart of evolving universe
표 Bacterial chromosome and polyampholytes
표 Relationships among superconductivity, physical property, and new materials/devices from high-Tc superconductors
표 Phase Diagram of High-Tc Cuprates
표 Research Strategies and Methods
표 Relationship between sub-projects (R&D program
표 Annual status of performance optimization and parallelization
표 Best papers by parallel optimization technical support
표 parallel performance test of turbulence simulation by using maximum 6,144cores
표 Major scientific areas supported by NERSC program
표 Load balancing efficiency of E3SM meteorological model: load balancing is maintained when an application is executed with up to 33792 KNL processors
표 Subprojects funded by FET-HPC in Europe
표 list of 5 major areas, 9 major tasks of computational science research in Japan and their implementation institutions
표 . Gordon Bell Award List
표 (Left) Computing steps of Recursive Green’s Function algorithm (Right) Performance gain driven by each optimization technique
표 (a) Strong scalability of the quantum transport simulation that is conducted for a Silicon nanowire consisting of 102,400 atoms. (b) Transmission coefficient and (c) local density states that are given as a function of conduction band energy
표 Scalability of turbulent flow solver (VDMPM-STD) (a) execution time (b) speedup
표 Coarse-grid aggregation (CGA) method (a) Grid aggregation process (b) Relative performance profiling
표 CGPSA process. Grid aggregations are conducted at Level2 and Level3
표 Parallel performance of CGPSA (a) Scalability until 6.3PFlops (131k core) (b) Comparison with other MG methods
표 Scalability: Strong scale of optimized FSI poisson solver
표 Performance comparison of resource allocation sequence
표 Scalability: Strong scale of FSI solver up to 2500 nodes (160,000 cores)
표 Write performance comparison of with an without hybrid I/O method
표 I/O performance comparison of global file and burst buffer system: (a) read, (b) write
표 Scalability: Strong scale of optimized channel code up to 256 nodes(16,384 cores)
표 Scalability: Weak scale of optimized channel code up to 512 nodes(32,768 cores)
표 Peridynamics model 1
표 Performance comparision of peridynamics code – model 1 (4,000 iterations)
표 Performance comparision of peridynamics code – model 2 (300 iterations)
표 Performance result of Nek500 code
표 Schematic of halo-cell of 2D domain (Selvan et al, AIAA SciTech, 2015)
표 Performance result of 3D N-S code (a) Performance speedup (b) Communication cost
표 Performance result of AlphaMet code at Jeju area(X-axis:cores, Y-axis: elapsed time)
표 Performance of AlphaMet code at Busan area(X-axis:cores, Y-axis: elapsed time)
표 Fluid analysis of turbine nozzle using OpenFOAM
표 Fluid analysis of first stage(stator+rotor) of turbine using OpenFOAM
표 FSI multiphysics analysis of bar(OpenFOAM application)
표 Nurion system test of LES turbulent fluid analysis for the simple model(OpenFOAM application)
표 Exascale analysis application using OpenFOAM about wind power generator and wind farm
표 Nurion linkage improvement of KISTIFoam
표 Summary of supercomputer resource share
표 1st / 2nd / 3rd list of support tasks of Innovation Support Program(giant research)
표 Systematic cooperation among industry, academia, institute, and government for finding COVID-19 drug candidate
표 News article from HPC Wire
표 Plum structure variation depending on Ra (a) Ra=107 (b) Ra=5x108 (c) Ra=2x1010
표 Barycentric map of velocity gradient from maximul JPDF location: RBC(color) and turbulent (black)
표 Cross section of bridge deck and Flow vorticity at Re=10,000
표 Field vaule over time at Re=10,000 (a)Vertical vibration and (b)Twist angle
표 Reproduction of distortion through simulation of the collapse of the Tacoma bridge in Nurion supercomputer
표 Calculation of Bertch parameters using NLEFT
표 (a) Stark-shift in uniformly mixed and segregated (induced by phase separation) 10nm-scale MAPbBrCl quantum dot structures. (b) Inter-band light transition rate and (c) characteristic wavelength that are simulated under external fields
표 (a) Device structure and simulation domain (b) Multi-scale modeling method that employs the Thomas-Fermi calculation and effective-mass electronic structure calculations to describe “Classical” and “Quantum” region, respectively. (c) 2-qubit time responses of the DQD system that show a timing point where CNOT operation is completed, and the corresponding fidelity plotted as a function of inaccuracy in the exchange interaction energy (J)
표 Visualization of multi-dimensional sensor data based on ML
표 Performance evaluation of the proposed machine learning to recognize gunshot sound
표 Bacterial chromosome and self-avoidance
표 Polyampholyte and its translocation
표 The predicted U(1) phase diagrams for 5 different lattice sizes, i.e., 10ⅹ10, 20ⅹ20, 30ⅹ30, 40ⅹ40, 50ⅹ50 lattice with the Heisenberg coupling constant, J/t = 0.2, 0.3, 0.4
표 A conceptual diagram of an artificial intelligence module that predicts the density matrix proposed in the study
표 Scattering angles of black hole encounters for various impact parameters obtained by numerical relativity simulations (NR) and three approximation methods, e.g., Newtonian gravity (Newtonian), 3rd order post-Minkowskian (3PM), and effective one-body Hamiltonian (EOB), respectively
표 The purpose of high energy physics
표 Masses and cross-sections of dark matter depending on theory
표 The scheme of dark matter search based on machine learning and big data with theory
표 The production and transfer of big data from particle experiments and astronomical observatory
표 Results of 7 layer deep learning using 4096 node of KIST-5 supercomputer
표 Big data analysis using KISTI-5 supercomputer with deep learning
표 Newspaper reports on Belle II Data Handling (2020.02.25)
표 The Feynman diagram of signal event
표 Cross-section depending on center of mass energy
표 CPU time and wall clock time depending on one or more cores
표 Wall clock time of machines depending on the number of jobs
표 The role of the KISTI at the DUNE
표 The location of DUNE experiment at Fermilab
표 The far detector and near detector at the DUNE
표 The result of production cross-section of neutron rich No isotopes (I)
표 The result of production cross-section of neutron rich No isotopes (II)
표 The comparisons of CPU time at Nurion KNL system depending Geant4 on versions
표 The comparison of memory of first events at Nurion KNL system depending on Geant4 versions
표 The comparison of CPU time of KISTI-4 and KISTI-5 supercomputers
표 The comparison of memory of KISTI-4 and KISTI-5 supercomputers

연구자의 다른 보고서 :

참고문헌 (25)

활용도 분석정보

상세보기

다운로드

내보내기

활용도 Top5 보고서

해당 보고서가 속한 카테고리에서 활용도가 높은 상위 5개 콘텐츠를 보여줍니다.
더보기 버튼을 클릭하시면 더 많은 관련자료를 살펴볼 수 있습니다.

과제명(ProjectTitle) :	-
연구책임자(Manager) :	-
과제기간(DetailSeriesProject) :	-
총연구비 (DetailSeriesProject) :	-
키워드(keyword) :	-
과제수행기간(LeadAgency) :	-
연구목표(Goal) :	-
연구내용(Abstract) :	-
기대효과(Effect) :	-

내보내기 구분	파일저장 인쇄 메일전송
구성항목	기본정보 상세정보 관리번호, 제목(한글), 저자명(한글), 발행일자, 전자원문, 초록(한글), 초록(영문) 관리번호, 제목(한글), 제목(영문), 저자명(한글), 저자명(영문), 주관연구기관(한글), 주관연구기관(영문), 발행일자, 총페이지수, 주관부처명, 과제시작일, 보고서번호, 과제종료일, 주제분류, 키워드(한글), 전자원문, 키워드(영문), 입수제어번호, 초록(한글), 초록(영문), 목차
저장형식	Text(ASCII format) Excel format
메일정보	받는사람 (필수) @ 보내는사람 (선택) @ 제목 내용 KISTI 검색결과 이메일 서비스
안내	총 건의 자료가 검색되었습니다. 다운받으실 자료의 인덱스를 입력하세요. (1-10,000) 검색결과의 순서대로 최대 10,000건 까지 다운로드가 가능합니다. 데이타가 많을 경우 속도가 느려질 수 있습니다.(최대 2~3분 소요) 다운로드 파일은 UTF-8 형태로 저장됩니다. 파일의 내용이 제대로 보이지 않을실 때는 웹브라우저 상단의 보기 -> 인코딩 -> 자동선택 여부를 확인하십시오. ~ Text(ASCII format) Excel format

연합인증

[국가R&D연구보고서] 극한규모 슈퍼컴퓨팅 기반 과학·공학 분야 거대문제 해결·지원
Solving grand-challenge problems in science and engineering by using extreme scaling of supercomputing 원문보기