서브메뉴
검색
Modern Quantum Chemistry with GPU Acceleration
Modern Quantum Chemistry with GPU Acceleration
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20260202104743
- ISBN
- 9798290649412
- DDC
- 500
- 저자명
- Wang, Yuanheng.
- 서명/저자
- Modern Quantum Chemistry with GPU Acceleration
- 발행사항
- [Sl] : Stanford University, 2024
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2024
- 형태사항
- 241 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 87-01, Section: B.
- 주기사항
- Advisor: Martinez, Todd.
- 학위논문주기
- Thesis (Ph.D.)--Stanford University, 2024.
- 초록/해제
- 요약Electronic structure is one of the key areas in the field of computational chemistry. It describes the arrangement of electrons around the nuclei, and can be used to model how the electrons and nuclei move over time. This typically requires a large amount of computation, thus necessitating the adoption of high-performance computing (HPC) systems. Graphics processing units (GPUs) have become increasingly popular in HPC clusters due to their ability to perform parallel computation efficiently. In this work, I will present our GPU-accelerated implementation for electronic structure calculations in the TeraChem software package, as well as how to leverage TeraChem's efficiency for interactive applications.One computational bottleneck for mean-field electronic structure calculations is the evaluation of the Hamiltonian integrals. We show our equation-to-code converter, data structure and optimal parallel algorithm for GPU acceleration of the Hamiltonian integrals, which enables TeraChem to be one of the fastest electronic structure packages. We also extend our package to support f type orbitals and Γ-point crystalline orbitals, thus permitting efficient simulation of large transition metal complexes and extended systems. Furthermore, we have developed an efficient and easy-to-use interface to TeraChem based on a server-client model, that permits other packages to use TeraChem as an electronic structure engine for molecular dynamics simulation.The advantages of GPU acceleration are most evident when applied to interactive applications. We developed a real-time molecular dynamics visualizer, that utilizes virtual reality (VR) headsets to promote human perception of and interaction with molecules. The users can exert forces on atoms to trigger reactions and visualize how the system responds to their applied impulse, thus exploring chemical reactions based on their chemical intuition (and refining their intuition in the process). Additionally, the platform enables simulation of excited-state dynamics, allowing investigation of photochemical processes.
- 일반주제명
- Decomposition
- 일반주제명
- Software
- 일반주제명
- Energy
- 일반주제명
- Electrons
- 기타저자
- Stanford University.
- 기본자료저록
- Dissertations Abstracts International. 87-01B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008260126s2024 us c eng d■001000017358729
■00520260202104743
■006m o d
■007cr#unu||||||||
■020 ▼a9798290649412
■035 ▼a(MiAaPQ)AAI32149728
■035 ▼a(MiAaPQ)Stanfordtc490bq3169
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a500
■1001 ▼aWang, Yuanheng.
■24510▼aModern Quantum Chemistry with GPU Acceleration
■260 ▼a[Sl]▼bStanford University▼c2024
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2024
■300 ▼a241 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 87-01, Section: B.
■500 ▼aAdvisor: Martinez, Todd.
■5021 ▼aThesis (Ph.D.)--Stanford University, 2024.
■520 ▼aElectronic structure is one of the key areas in the field of computational chemistry. It describes the arrangement of electrons around the nuclei, and can be used to model how the electrons and nuclei move over time. This typically requires a large amount of computation, thus necessitating the adoption of high-performance computing (HPC) systems. Graphics processing units (GPUs) have become increasingly popular in HPC clusters due to their ability to perform parallel computation efficiently. In this work, I will present our GPU-accelerated implementation for electronic structure calculations in the TeraChem software package, as well as how to leverage TeraChem's efficiency for interactive applications.One computational bottleneck for mean-field electronic structure calculations is the evaluation of the Hamiltonian integrals. We show our equation-to-code converter, data structure and optimal parallel algorithm for GPU acceleration of the Hamiltonian integrals, which enables TeraChem to be one of the fastest electronic structure packages. We also extend our package to support f type orbitals and Γ-point crystalline orbitals, thus permitting efficient simulation of large transition metal complexes and extended systems. Furthermore, we have developed an efficient and easy-to-use interface to TeraChem based on a server-client model, that permits other packages to use TeraChem as an electronic structure engine for molecular dynamics simulation.The advantages of GPU acceleration are most evident when applied to interactive applications. We developed a real-time molecular dynamics visualizer, that utilizes virtual reality (VR) headsets to promote human perception of and interaction with molecules. The users can exert forces on atoms to trigger reactions and visualize how the system responds to their applied impulse, thus exploring chemical reactions based on their chemical intuition (and refining their intuition in the process). Additionally, the platform enables simulation of excited-state dynamics, allowing investigation of photochemical processes.
■590 ▼aSchool code: 0212.
■650 4▼aDecomposition
■650 4▼aSoftware
■650 4▼aEnergy
■650 4▼aElectrons
■690 ▼a0791
■71020▼aStanford University.
■7730 ▼tDissertations Abstracts International▼g87-01B.
■790 ▼a0212
■791 ▼aPh.D.
■792 ▼a2024
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17358729▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
