EXTREME-SCALE QUANTUM SIMULATIONS (Spring 2018)

Course Number: CSCI 699
Class Number: 29950D
Instructor: Aiichiro Nakano; office: VHE 610; phone: (213) 821-2657; email: anakano@usc.edu
Lecture: 5:00-6:50pm M W, GFS 223
Office Hour: 5:00-5:50pm F, VHE 610
HPC Office Hour: 2:30-5:00 pm T, LVL 3M
Prerequisites: (1) CSCI596 or basic experience in parallel computing; and (2) PHYS 516 or basic knowledge of numerical methods in computational sciences.
Textbooks: None; all course materials are provided online on the course Web page. See also:
Electronic structure, R. M. Martin (Cambridge Univ. Press, '04),
Ab initio molecular dynamics, D. Marx & J. Hutter (Cambridge Univ. Press, '09),
Time-dependent density-functional theory, C. A. Ullrich (Oxford Univ. Press, '12),
Effective computation in physics, A. Scopatz & K. D. Huff (O'Reilly, '15).

Course Description
Computer simulation of quantum-mechanical dynamics has become an essential enabling technology for physical, chemical and biological sciences and engineering. Quantum-dynamics simulations on extreme-scale parallel supercomputers would provide unprecedented predictive power, but pose enormous challenges as well. This course surveys and projects algorithmic and computing technologies that will make quantum-dynamics simulations metascalable, i.e., "design once, continue to scale on future computer architectures".

The course first covers how the exponential time complexity for solving the quantum N-body problem is reduced to (1) O(N³) within the density functional theory (DFT), for which Walter Kohn received a Nobel chemistry prize in 1998, and (2) O(N) based on physical data-locality principles (e.g., Kohn's quantum nearsightedness principle). The course then introduces key abstractions (e.g., pseudopotentials and exchange-correlation functionals) and representation issues (e.g., plane-wave basis vs. real-space multigrids), which are necessary for efficient implementation of quantum molecular dynamics (QMD) simulations. This is followed by the design of QMD simulation algorithms on massively parallel supercomputers using message passing and multithreading, including our metascalable divide-conquer-recombine (DCR) algorithmic framework, as well as performance optimization on modern many-core processors and accelerators through memory hierarchies and vectorization. Advanced topics to be covered include (1) DCR approaches to excitation dynamics, (2) intersection of machine learning and quantum N-body problem, and (3) merger of quantum Monte Carlo (QMC) and QMD methods. The course ends with best software practices for co-developing extreme-scale QMD software for million-way parallelism.

Students will learn fundamental knowledge and gain hands-on experience in order to: (1) reduce the intractable quantum many-body problem to lower-complexity problems, while retaining the essential physics; (2) design scalable parallel algorithms for linear-scaling quantum-dynamics simulations; (3) develop metascalable quantum-dynamics software on current and future computer architectures.

Nonadiabatic quantum molecular dynamics simulation to study photoexcitation dynamics in MoSe₂ bilayer [M.-F. Lin et al., Nature Commun. 8, 1745 (2017)].

Announcements

• 1/8 (M): Class begins.
• 1/8 (M): This class will have guest participants: Prof. Kohei Shimamura, Prof. Shigenori Tanaka, and several graduate students from Kobe University in Japan.
• 1/8 (M): Postdoctoral scientist position for computational materials science on Japanese post-K supercomputer; information; see also Prof. Takeo Hoshi's 100 million-atom quantum simulation paper.
• 1/11 (Th): Seminar on machine learning for materials modeling by Prof. Efthimios Kaxiras (Harvard) at 12:45 pm in ZHS 159.
• 1/15 (M): No class (Martin Luther King Jr. Day).
• 1/17 (W): Getting started with QXMD software -- please bring your laptop to the class.
• 1/22 (M): HPC accounts for USC students are ready. Please work in the class director, /home/rcf-proj/an1.
• 1/22 (M): See Aurora Early Science Program (ESP) on simulation and data & learning.
• 1/22 (M): Lecture on basic QMD equations.
• 1/24 (W): Lecture & discussion on DFT.
• 1/29 (M): Lecture on exchange-correlation functional.
• 1/31 (W): Lecture on pseudopotentials.
• 2/5 (M): Programs on SCAN metaGGA functional (metagga.F), DFTD3 (dftd3/) & PAW construction (atm7.f) are found in Google Drive.
• 2/5 (M): Discussion on pseudopotentials.
• 2/7 (W): Our Aurora/A21 Early Science Program (ESP) project mentioned in Science magazine news; article.
• 2/7 (W): First Japanese USC blog by Fatemeh.
• 2/7 (W): Lecture on representation and solution.
• 2/12 (M): Hands on assignment 1: Born-Oppenheimer molecular dynamics (BOMD).
• 2/14 (W): Linear-scaling QMD basics.
• 2/16 (F): Discussion session for assignment 1 at 5:00 pm in VHE 610
• 2/19 (M): No class (President's Day).
• 2/21 (W): See DOE exascale reports, in particular BES report.
• 2/21 (W): Advanced linear-scaling QMD methods.
• 2/26 (M): Hands on assignment 2: Construct your own projector augmented waves (PAW).
• 2/28 (W): Lecture on nonadiabatic quantum molecular dynamics (NAQMD).
• 2/28 (W): Kobe-USC photoshoot: photo1, photo2
• 3/2 (F): No office hour; MAGICS workshop.
• 3/5 (M): Hands on assignment 3: NAQMD.
• 3/7 (W): Lecture on linear-response time-dependent density functional theory (LR-TDDFT).
• 3/7 (W): Information session by QuantLab.
• 3/12 (M), 3/14 (W): No class (spring recess).
• 3/19 (M): Lecture on hybrid exact-exchange and excited-state force calculations.
• 3/21 (W): Lecture on parallel QMD.
• 3/26 (M): Lecture on advanced parallel QMD.
• 3/28 (W): Discussion of real-space DFT on K computer: Hasegawa et al., Proc. Supercomputing, SC11 (2011) led by Xiangyu.
• 4/2 (M): Hands on assignment 4: Linear-response time-dependent density functional theory (LR-TDDFT).
• 4/4 (W): Discussion of density matrix renormalization group (DMRG) papers led by Yongqian.
• 4/9 (M): Lecture on neural network force field (NNFF) by Prof. Kohei Shimamura (Kobe Univ.); jointly with MASC 576 in RTH 109 (please noe the different class room, only for this day).
• 4/11 (W): Discussion of NNFF papers led by Size, Yuan and Shichun.
• 4/16 (M): Discussion of Harris-functional papers led by Ben.
• 4/18 (W): Discussion of (1) information-theory/renormalization-group/neural-network papers led by Kyle, and (2) "quantum chemistry on quantum computer" paper by Guoqing and Yuzi.
• 4/23 (M), 4/25 (W): Final project presentations.
• 5/9 (W): Final project report due.