Pushing Back the Limit of Ab-initio Quantum Transport Simulations on Hybrid Supercomputers

被引：23

作者：

Calderara, Mauro ^{[1
]}

Brueck, Sascha ^{[1
]}

Pedersen, Andreas ^{[1
]}

Bani-Hashemian, Mohammad H. ^{[2
]}

VandeVondele, Joost ^{[2
]}

Luisier, Mathieu ^{[1
]}

机构：

[1] ETH, Integrated Syst Lab, CH-8092 Zurich, Switzerland

[2] ETH, Nanoscale Simulat, CH-8093 Zurich, Switzerland

来源：

PROCEEDINGS OF SC15: THE INTERNATIONAL CONFERENCE FOR HIGH PERFORMANCE COMPUTING, NETWORKING, STORAGE AND ANALYSIS | 2015年

关键词：

HIGH-PERFORMANCE; 1ST-PRINCIPLES; SCHEMES; DEVICES; SOLVER;

D O I：

10.1145/2807591.2807673

中图分类号：

TP301 [理论、方法];

学科分类号：

081202 ;

摘要：

The capabilities of CP2K, a density-functional theory package and OMEN, a nano-device simulator, are combined to study transport phenomena from first-principles in unprecedentedly large nanostructures. Based on the Hamiltonian and overlap matrices generated by CP2K for a given system, OMEN solves the Schrodinger equation with open boundary conditions (OBCs) for all possible electron momenta and energies. To accelerate this core operation a robust algorithm called SplitSolve has been developed. It allows to simultaneously treat the OBCs on CPUs and the Schrodinger equation on GPUs, taking advantage of hybrid nodes. Our key achievements on the Cray-XK7 Titan are (i) a reduction in time-to-solution by more than one order of magnitude as compared to standard methods, enabling the simulation of structures with more than 50000 atoms, (ii) a parallel efficiency of 97% when scaling from 756 up to 18564 nodes, and (iii) a sustained performance of 15 DP-PFlop/s.

引用

页数：12

共 56 条

[1] Multifrontal parallel distributed symmetric and unsymmetric solvers [J].

Amestoy, PR ;

Duff, IS ;

L'Excellent, JY .

COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING, 2000, 184 (2-4) :501-520

[2]

Anderson E., 1999, LAPACK Users' Guide, V3rd

[3] Quantum Dot Light-Emitting Devices with Electroluminescence Tunable over the Entire Visible Spectrum [J].

Anikeeva, Polina O. ;

Halpert, Jonathan E. ;

Bawendi, Moungi G. ;

Bulovic, Vladimir .

NANO LETTERS, 2009, 9 (07) :2532-2536

[4]

Baboulin M, 2015, P INT C PAR PROC APP

[5] A NEW APPROACH TO HIGH-EFFICIENCY MULTI-BAND-GAP SOLAR-CELLS [J].

BARNHAM, KWJ ;

DUGGAN, G .

JOURNAL OF APPLIED PHYSICS, 1990, 67 (07) :3490-3493

[6] An integral method for solving nonlinear eigenvalue problems [J].

Beyn, Wolf-Juergen .

LINEAR ALGEBRA AND ITS APPLICATIONS, 2012, 436 (10) :3839-3863

[7] Density-functional method for nonequilibrium electron transport -: art. no. 165401 [J].

Brandbyge, M ;

Mozos, JL ;

Ordejón, P ;

Taylor, J ;

Stokbro, K .

PHYSICAL REVIEW B, 2002, 65 (16) :1654011-16540117

[8] Atomically precise bottom-up fabrication of graphene nanoribbons [J].

Cai, Jinming ;

Ruffieux, Pascal ;

Jaafar, Rached ;

Bieri, Marco ;

Braun, Thomas ;

Blankenburg, Stephan ;

Muoth, Matthias ;

Seitsonen, Ari P. ;

Saleh, Moussa ;

Feng, Xinliang ;

Muellen, Klaus ;

Fasel, Roman .

NATURE, 2010, 466 (7305) :470-473

[9] A two-dimensional domain decomposition technique for the simulation of quantum-scale devices [J].

Cauley, Stephen ;

Balakrishnan, Venkataramanan ;

Klimeck, Gerhard ;

Koh, Cheng-Kok .

JOURNAL OF COMPUTATIONAL PHYSICS, 2012, 231 (04) :1293-1313

[10] Electronically conductive phospho-olivines as lithium storage electrodes [J].

Chung, SY ;

Bloking, JT ;

Chiang, YM .

NATURE MATERIALS, 2002, 1 (02) :123-128

← 1 2 3 4 5 6 →