DOCSLIB.ORG

High-Performance Computing at CSCS: Co-Design and New Separations of Concerns for HPC

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
High-Performance Computing at CSCS: Co-Design and New Separations of Concerns for HPC High-Performance Computing at CSCS: Co-Design and New Separations of Concerns for HPC Thomas C. Schulthess Computer performance and application performance increase ~103 every decade ~100 Kilowatts ~5 Megawatts 20-30 MW ~1 Exaflop/s 100 million or billion 1.35 Petaflop/s processing cores (!) Cray XT5 150’000 processors This system was built with commodity 1.02 Teraflop/s processors Cray T3E 1’500 processors 1 Gigaflop/s Cray YMP 8 processors 1988 1998 2008 2018 First sustained GFlop/s First sustained TFlop/s First sustained PFlop/s Another 1,000x increase in Gordon Bell Prize 1988 Gordon Bell Prize 1998 Gordon Bell Prize 2008 sustained performance Wednesday, March 13, 2013 HPC Advisory Council Switzerland Conference, Lugano HOW WELL CAN APPLICATIONS DEAL WITH CONCURRENCY? HOW EFFICIENT ARE THEY? Wednesday, March 13, 2013 HPC Advisory Council Switzerland Conference, Lugano Applications running at scale on Jaguar @ ORNL (Spring 2011) Domain area Code name Institution # of cores Performance Notes 2008 Gordon Bell Materials DCA++ ORNL 213,120 1.9 PF Prize Winner 2009 Gordon Bell Materials WL-LSMS ORNL/ETH 223,232 1.8 PF Prize Winner 2008 Gordon Bell Chemistry NWChem PNNL/ORNL 224,196 1.4 PF Prize Finalist 2010 Gordon Bell Materials DRC ETH/UTK 186,624 1.3 PF Prize Hon. Mention 2010 Gordon Bell Nanoscience OMEN Duke 222,720 > 1 PF Prize Finalist 2010 Gordon Bell Biomedical MoBo GaTech 196,608 780 TF Prize Winner Chemistry MADNESS UT/ORNL 140,000 550 TF 2008 Gordon Bell Materials LS3DF LBL 147,456 442 TF Prize Winner 2008 Gordon Bell Seismology SPECFEM3D USA (multiple) 149,784 165 TF Prize Finalist Combustion S3D SNL 147,456 83 TF Weather WRF USA (multiple) 150,000 50 TF Wednesday, March 13, 2013 HPC Advisory Council Switzerland Conference, Lugano Hirsch-Fey quantum Monte Carlo with Delayed updates (or Ed updates) Ed D’Azevedo, ORNL t Gc({si,l}k+1)=Gc({si,l}k)+ak × bk t Gc({si,l}k+1)=Gc({si,l}0)+[a0|a1|...|ak] × [b0|b1|...|bk] 2 6000 Complexity for k updates remains O(kNt ) mixed precision double precision But we can replace k rank-1 updates 4000 with one matrix-matrix multiply plus some additional bookkeeping. 2000 time to solution [sec] G. Alvarez, M. S. Summers, D. E. Maxwell, M. Eisenbach, J. S. Meredith, J. M. Larkin, J. Levesque, T. A. Maier, P. R. C. Kent, E. F. D'Azevedo, T. C. Schulthess; New algorithm to enable 400+ TFlop/ 0 0 20 40 60 80 100 s sustained performance in simulations of disorder effects in high-Tc superconductors, Proceedings of the 2008 ACM/IEEE conference on Supercomputing 61, 2008 delay Wednesday, March 13, 2013 HPC Advisory Council Switzerland Conference, Lugano Arithmetic intensity of a computation floating point operations ↵ = data transferred t Gc({si,l}k+1)=Gc({si,l}k)+ak × bk α o(1) ⇡ t Gc({si,l}k+1)=Gc({si,l}0)+[a0|a1|...|ak] × [b0|b1|...|bk] α o(k) ⇡ Wednesday, March 13, 2013 HPC Advisory Council Switzerland Conference, Lugano Algorithmic motifs and their arithmetic intensity Arithmetic intensity: number of operations per word of memory transferred Finite difference / stencil in S3D and WRF (& COSMO) Rank-1 update in HF-QMC Rank-N update in DCA++ QMR in WL-LSMS Sparse linear algebra Linpack (Top500) Matrix-Vector Vector-Vector Fast Fourier Transforms Dense Matrix-Matrix BLAS1&2 FFTW & SPIRAL BLAS3 O(1) O(log N) O(N) Supercomputers are designed for certain algorithmic motifs – which ones? Wednesday, March 13, 2013 HPC Advisory Council Switzerland Conference, Lugano Petaflop/s = 1015 64-bit floating point operations / sec. which takes more energy? 64-bit floating-point fused multiply add or moving three 64-bit operands 20 mm across the die 934,569.299814557 x 52.827419489135904 ---------------------------- = 49,370,884.442971624253823 + 4.20349729193958 ---------------------------- = 49,370,888.64646892 20 mm this takes over 3x the energy! loading the data from off chip takes > 10x more yet source: Steve Scott, Cray Inc. moving data is expensive – exploiting data locality is critical to energy efficiency If we care about energy consumption, we have to worry about these and other physical considerations of the computation – but where is the separation of concerns? Wednesday, March 13, 2013 HPC Advisory Council Switzerland Conference, Lugano COSMO in production for Swiss weather prediction ECMWF COSMO-7 2x per day 3x per day 72h forecast 16 km lateral grid, 91 6.6 km lateral grid, 60 layers layers COSMO-2 8x per day 24h forecast 2.2 km lateral grid, 60 layers § Some of the products generated from these simulations § Daily weather forecast § Forecasting for air traffic control (Sky Gide) § Safety management in event of nuclear incidents Wednesday, March 13, 2013 HPC Advisory Council Switzerland Conference, Lugano Insight into model/methods/algorithms used in COSMO § PDE on structured grid (variables: velocity, temperature, pressure, humidity, etc.) § Explicit solve horizontally (I, J) using finite difference stencils § Implicit solve in vertical direction (K) with tri-diagonal solve in every column (applying Thomas algorithm in parallel – can be expressed as stencil) ~2km Due to implicit solves in the vertical we can work with 60m longer time steps (2km and not 60m grid size K J is relevant) Tri-diagonal solves I Wednesday, March 13, 2013 HPC Advisory Council Switzerland Conference, Lugano Hence the algorithmic motif are § Tri-diagonal solve § vertical K-diretion § with loop carried dependencies in K J K I § Finite difference stencil computations J § focus on horizontal IJ-plane access K § no loop carried dependencies Wednesday, March 13, 2013 HPC Advisory Council Switzerland Conference, Lugano Performance profile of (original) COSMO-CCLM Runtime based 2 km production model of MeteoSwiss % Code Lines (F90) % Runtime Wednesday, March 13, 2013 HPC Advisory Council Switzerland Conference, Lugano Analyzing two examples/motifs – how are they different? Physics 3 memory accesses 136 FLOPs è compute bound Dynamics 3 memory accesses 5 FLOPs è memory bound Wednesday, March 13, 2013 HPC Advisory Council Switzerland Conference, Lugano Running the simple examples on the Cray XK6 Compute bound (physics) problem Machine Interlagos Fermi (2090) GPU+transfer Time 1.31 s 0.17 s 1.9 s Speedup 1.0 (REF) 7.6 0.7 Memory bound (dynamics) problem Machine Interlagos Fermi (2090) GPU+transfer Time 0.16 s 0.038 s 1.7 s Speedup 1.0 (REF) 4.2 0.1 The simple lesson: leave data on the GPU! Wednesday, March 13, 2013 HPC Advisory Council Switzerland Conference, Lugano Performance profile of (original) COSMO-CCLM Runtime based 2 km production model of MeteoSwiss % Code Lines (F90) % Runtime Original code (with OpenACC for GPU) Rewrite in C++ (with CUDA backend for GPU) Wednesday, March 13, 2013 HPC Advisory Council Switzerland Conference, Lugano Stencil Library Ideas § Implement a stencil library using C++ and template meta programming § 3D structured grid § Parallelization in horizontal IJ-plane (sequential loop in K for tridiagonal solves) § Multi-node support using explicit halo exchange (Generic Communication Library – not covered by presentation) § Abstract the hardware platform (CPU/GPU) § Adapt loop order and storage layout to the platform § Leverage software caching § Hide complex and “ugly” optimizations § Blocking § Single source code compiles to multiple platforms § Currently, efficient back-ends are implemented for CPU and GPU CPU GPU Storage Order (Fortran notation) KIJ IJK Parallelization OpenMP CUDA Wednesday, March 13, 2013 HPC Advisory Council Switzerland Conference, Lugano “Apples to Oranges” comparison of new vs. current COSMO Speedup of new code on K20x GPU vs. current code on AMD Interlagos (@2.1GHz) 7.0# new C++ code with CUDA (*) 6.0# 5.0# 4.0# Physics# Dynamics# 3.0# Dyn+Physics# Speedup& 2.0# current F90 code with OpenACC 1.0# 0.0# 64x128# 128x128# 256*256# (*) new C++ code with OpenMP is about 2x faster than current F90 code Wednesday, March 13, 2013 HPC Advisory Council Switzerland Conference, Lugano Solving Kohn-Sham equation is the bottleneck of most DFT based materials science codes 2 ~ 2 Kohn-Sham Eqn. + vLDA(⇤r) ⇥ (⇤r)= ⇥ (⇤r) −2mr i i i ✓ ◆ Ansatz ⇥i(⇤r)= ciµφµ(⇤r) µ X 2 ~ 2 Hermitian matrix H = φ⇤ (⇥r) + v (⇥r) φ (⇥r)d⇥r µν µ −2mr LDA ν Z ✓ ◆ Basis may not be orthogonal Sµν = φµ⇤ (⇥r)φν (⇥r)d⇥r Z Solve generalized eigenvalue problem (H " S)=0 − i Wednesday, March 13, 2013 HPC Advisory Council Switzerland Conference, Lugano Solving Kohn-Sham equation is the bottleneck of most DFT based materials science codes 2 ~ 2 Kohn-Sham Eqn. + vLDA(⇤r) ⇥ (⇤r)= ⇥ (⇤r) −2mr i i i ✓ ◆ Remarks:Ansatz ⇥i(⇤r)= ciµφµ(⇤r) Typical matrix size is up to 10’000 µ X Only a few methods require high accuracy and matrix size2 up to 100’000 Hermitian matrix ~ 2 Hµν = φµ⇤ (⇥r) + vLDA(⇥r) φν (⇥r)d⇥r Some methods try to avoid eigensolvers (orthogonalization−2m willr be important motif instead) Z ✓ ◆ When lower accuracy is an option, order(N) methods are used instead Basis may not be orthogonal (this is coveredSµ byν =one of φourµ⇤ ( HP2C⇥r)φν projects(⇥r)d⇥r see CP2K on hp2c.ch) Z Dominant motif: Solve generalized eigenvalue problem (H " S)=0 − i We will need between 10% and 50% of the eigenvectors Wednesday, March 13, 2013 HPC Advisory Council Switzerland Conference, Lugano Solving the generalized eigenvalue problem Ax = λBx Standard 1 stage solver H A0y = λy xPOTRF B = LL 1 H H xHETRD xHEGST A0 = L− AL− T = Q A0Q Most time consuming step, dominated by level 2 BLAS xHEEVx A0y = λy (memory bound) xSTExx Ty0 = λy0 xTRSM H x = L− y y = Qy0 xUNMTR Wednesday, March 13, 2013 HPC Advisory Council Switzerland Conference, Lugano Solving
Load more

Recommended publications
  • Free and Open Source Software for Computational Chemistry Education
    Free and Open Source Software for Computational Chemistry Education Susi Lehtola∗,y and Antti J. Karttunenz yMolecular Sciences Software Institute, Blacksburg, Virginia 24061, United States zDepartment of Chemistry and Materials Science, Aalto University, Espoo, Finland E-mail: [email protected].fi Abstract Long in the making, computational chemistry for the masses [J. Chem. Educ. 1996, 73, 104] is finally here. We point out the existence of a variety of free and open source software (FOSS) packages for computational chemistry that offer a wide range of functionality all the way from approximate semiempirical calculations with tight- binding density functional theory to sophisticated ab initio wave function methods such as coupled-cluster theory, both for molecular and for solid-state systems. By their very definition, FOSS packages allow usage for whatever purpose by anyone, meaning they can also be used in industrial applications without limitation. Also, FOSS software has no limitations to redistribution in source or binary form, allowing their easy distribution and installation by third parties. Many FOSS scientific software packages are available as part of popular Linux distributions, and other package managers such as pip and conda. Combined with the remarkable increase in the power of personal devices—which rival that of the fastest supercomputers in the world of the 1990s—a decentralized model for teaching computational chemistry is now possible, enabling students to perform reasonable modeling on their own computing devices, in the bring your own device 1 (BYOD) scheme. In addition to the programs’ use for various applications, open access to the programs’ source code also enables comprehensive teaching strategies, as actual algorithms’ implementations can be used in teaching.
    [Show full text]
  • Living at the Top of the Top500: Myopia from Being at the Bleeding Edge
    Living at the Top of the Top500: Myopia from Being at the Bleeding Edge Bronson Messer Oak Ridge Leadership Computing Facility & Theoretical Astrophysics Group Oak Ridge National Laboratory Department of Physics & Astronomy University of Tennessee Friday, July 1, 2011 Outline • Statements made without proof • OLCF’s Center for Accelerated Application Readiness • Speculations on task-based approaches for multiphysics applications in astrophysics (e.g. blowing up stars) 2 Friday, July 1, 2011 Riffing on Hank’s fable... 3 Friday, July 1, 2011 The Effects of Moore’s Law and Slacking 1 on Large Computations Chris Gottbrath, Jeremy Bailin, Casey Meakin, Todd Thompson, J.J. Charfman Steward Observatory, University of Arizona Abstract We show that, in the context of Moore’s Law, overall productivity can be increased for large enough computations by ‘slacking’orwaiting for some period of time before purchasing a computer and beginning the calculation. According to Moore’s Law, the computational power availableataparticular price doubles every 18 months. Therefore it is conceivable that for sufficiently large numerical calculations and fixed budgets, computing power will improve quickly enough that the calculation will finish faster if we wait until the available computing power is sufficiently better and start the calculation then. The Effects of Moore’s Law and Slacking 1Figureon Large 1: Computations Chris Gottbrath, Jeremy Bailin, Casey Meakin, Todd Thompson, J.J. Charfman 1 The Effects of Moore’sSteward Observatory, Law and University Slacking of Arizona on Large astro-ph/9912202 Computations Abstract Chris Gottbrath, Jeremy Bailin, Casey Meakin, Todd Thompson, We show that, in the context of Moore’s Law, overall productivity can be increased forJ.J.
    [Show full text]
  • A Deep Dive Into ASUS Solutions
    From DataCenters to Supercomputers A Deep Dive Into ASUS Solutions Christopher Liang / Server/WS Product manager ASUS is a global technology leader in the Who is ASUS? digital era. We focus on the mastery of technological innovation and design perfection. We’re very critical of our own work when it comes to only delivering consumers our very best. ASUS Worldwide ASUS has a strong presence in over 50 countries, with offices in Europe, Asia, Australia and New Zealand, the Americas, and South Africa. • > 11,000 employees worldwide (source : HR dept ) • > 3,100 R&D employees (source : HR dept ) • 900+ support centers worldwide (source : TSD dept ) Business Update 11 (estimated) ASUS closed 2011 on a high, with revenues around US$11.8 billion. As of September 2010, the brand is estimated to be worth US$1.285 billion*. 10.1 *2010 Top Taiwan Global Brand (Interbrand) ** Due to Q1-Q2 worldwide economy crisis 8.21 7.66** 6.99 5.087 3.783 3.010 2.081 Revenue US$ (billions) Leader in Performance and Reliability #1 Motherboard Since 1989, ASUS has shipped over 420 million motherboards. Placed end to end, they can form a chain long enough to circumnavigate the globe more than three times. #1 Windows-based Desktop PC Reliability Ranked most reliable Window’s based PC brand 2 years in a row by PCWorld. The 2011 PCWorld Reliability and Service survey was conducted with 63,000 PCWorld readers. 1. Though design thinking to provide cutting Why ASUS ? edge SPEC 2. BIOS – superior performance through increased functionality and upgradeability 3.
    [Show full text]
  • An Overview on the Libxc Library of Density Functional Approximations
    An overview on the Libxc library of density functional approximations Susi Lehtola Molecular Sciences Software Institute at Virginia Tech 2 June 2021 Outline Why Libxc? Recap on DFT What is Libxc? Using Libxc A look under the hood Wrapup GPAW 2021: Users' and Developers' Meeting Susi Lehtola Why Libxc? 2/28 Why Libxc? There are many approximations for the exchange-correlation functional. But, most programs I ... only implement a handful (sometimes 5, typically 10-15) I ... and the implementations may be buggy / non-standard GPAW 2021: Users' and Developers' Meeting Susi Lehtola Why Libxc? 3/28 Why Libxc, cont'd This leads to issues with reproducibility I chemists and physicists do not traditionally use the same functionals! Outdated(?) stereotype: B3LYP vs PBE I how to reproduce a calculation performed with another code? GPAW 2021: Users' and Developers' Meeting Susi Lehtola Why Libxc? 4/28 Why Libxc, cont'd The issue is compounded by the need for backwards and forwards compatibility: how can one I reproduce old calculations from the literature done with a now-obsolete functional (possibly with a program that is proprietary / no longer available)? I use a newly developed functional in an old program? GPAW 2021: Users' and Developers' Meeting Susi Lehtola Why Libxc? 5/28 Why Libxc, cont'd A standard implementation is beneficial! I no need to keep reinventing (and rebuilding) the wheel I use same collection of density functionals in all programs I new functionals only need to be implemented in one place I broken/buggy functionals only need to be fixed in one place I same implementation can be used across numerical approaches, e.g.
    [Show full text]
  • Density-Functional Theory of Atoms and Molecules • W
    3.320: Lecture 7 (Feb 24 2005) DENSITYDENSITY--FUNCTIONALFUNCTIONAL THEORY,THEORY, ANDAND DENSITYDENSITY--FUNCTIONALFUNCTIONAL PRACTICEPRACTICE Feb 24 2005 3.320 Atomistic Modeling of Materials -- Gerbrand Ceder and Nicola Marzari Hartree-Fock Equations r rr ϕϕαβ()rr11() L ϕν()r1 ϕϕ()rrr ()rrϕ()r rr r 1 αβ22L ν2 ψ (,rr12,...,rn )= n! MMOM r rr ϕϕαβ()rrnn()L ϕν()rn ⎡⎤1 2 r r r ⎢⎥−∇iI+∑VR()−riϕλ ()ri + ⎣⎦2 I ⎡⎤1 ϕϕ* ()rrrr()drrϕ ()rr − ⎢⎥∑ ∫ µµjjr r jλ i ⎣⎦⎢⎥µ ||rrji− ⎡⎤1 ϕϕ* ()rdrrϕϕ()rrr r(rr)= ε(r) ∑ ⎢⎥∫ µµj rrλ j j i λ i µ ⎣⎦⎢⎥||rrj − i Feb 24 2005 3.320 Atomistic Modeling of Materials -- Gerbrand Ceder and Nicola Marzari Image removed for copyright reasons. Screenshot of online article. “Nobel Focus: Chemistry by Computer.” Physical Review Focus, 21 October 1998. http://focus.aps.org/story/v2/st19 Feb 24 2005 3.320 Atomistic Modeling of Materials -- Gerbrand Ceder and Nicola Marzari The Thomas-Fermi approach • Let’s try to find out an expression for the energy as a function of the charge density • E=kinetic+external+el.-el. • Kinetic is the tricky term: how do we get the curvature of a wavefunction from the charge density ? • Answer: local density approximation Feb 24 2005 3.320 Atomistic Modeling of Materials -- Gerbrand Ceder and Nicola Marzari Local Density Approximation • We take the kinetic energy density at every point to correspond to the kinetic energy density of the homogenous electron gas 5 T(rr) = Aρ 3 (rr) 5 1 ρ(rr)ρ(rr ) E [ρ] = A ρ 3 (rr)drr + ρ(rr)v (rr)drr + 1 2 drrdrr Th−Fe ∫ ∫ ext ∫∫ r r 1 2 2 | r1 − r2 | Feb 24 2005
    [Show full text]
  • Trends in Atomistic Simulation Software Usage [1.3]
    A LiveCoMS Perpetual Review Trends in atomistic simulation software usage [1.3] Leopold Talirz1,2,3*, Luca M. Ghiringhelli4, Berend Smit1,3 1Laboratory of Molecular Simulation (LSMO), Institut des Sciences et Ingenierie Chimiques, Valais, École Polytechnique Fédérale de Lausanne, CH-1951 Sion, Switzerland; 2Theory and Simulation of Materials (THEOS), Faculté des Sciences et Techniques de l’Ingénieur, École Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland; 3National Centre for Computational Design and Discovery of Novel Materials (MARVEL), École Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland; 4The NOMAD Laboratory at the Fritz Haber Institute of the Max Planck Society and Humboldt University, Berlin, Germany This LiveCoMS document is Abstract Driven by the unprecedented computational power available to scientific research, the maintained online on GitHub at https: use of computers in solid-state physics, chemistry and materials science has been on a continuous //github.com/ltalirz/ rise. This review focuses on the software used for the simulation of matter at the atomic scale. We livecoms-atomistic-software; provide a comprehensive overview of major codes in the field, and analyze how citations to these to provide feedback, suggestions, or help codes in the academic literature have evolved since 2010. An interactive version of the underlying improve it, please visit the data set is available at https://atomistic.software. GitHub repository and participate via the issue tracker. This version dated August *For correspondence: 30, 2021 [email protected] (LT) 1 Introduction Gaussian [2], were already released in the 1970s, followed Scientists today have unprecedented access to computa- by force-field codes, such as GROMOS [3], and periodic tional power.
    [Show full text]
  • (DFT) and Its Application to Defects in Semiconductors
    Introduction to DFT and its Application to Defects in Semiconductors Noa Marom Physics and Engineering Physics Tulane University New Orleans The Future: Computer-Aided Materials Design • Can access the space of materials not experimentally known • Can scan through structures and compositions faster than is possible experimentally • Unbiased search can yield unintuitive solutions • Can accelerate the discovery and deployment of new materials Accurate electronic structure methods Efficient search algorithms Dirac’s Challenge “The underlying physical laws necessary for the mathematical theory of a large part of physics and the whole of chemistry are thus completely known, and the difficulty is only that the exact application of these laws leads to equations much too complicated to be soluble. It therefore becomes desirable that approximate practical methods of applying quantum mechanics should be developed, which can lead to an P. A. M. Dirac explanation of the main features of Physics Nobel complex atomic systems ... ” Prize, 1933 -P. A. M. Dirac, 1929 The Many (Many, Many) Body Problem Schrödinger’s Equation: The physical laws are completely known But… There are as many electrons in a penny as stars in the known universe! Electronic Structure Methods for Materials Properties Structure Ionization potential (IP) Absorption Mechanical Electron Affinity (EA) spectrum properties Fundamental gap Optical gap Vibrational Defect/dopant charge Exciton binding spectrum transition levels energy Ground State Charged Excitation Neutral Excitation DFT
    [Show full text]
  • Lawrence Berkeley National Laboratory Recent Work
    Lawrence Berkeley National Laboratory Recent Work Title From NWChem to NWChemEx: Evolving with the Computational Chemistry Landscape. Permalink https://escholarship.org/uc/item/4sm897jh Journal Chemical reviews, 121(8) ISSN 0009-2665 Authors Kowalski, Karol Bair, Raymond Bauman, Nicholas P et al. Publication Date 2021-04-01 DOI 10.1021/acs.chemrev.0c00998 Peer reviewed eScholarship.org Powered by the California Digital Library University of California From NWChem to NWChemEx: Evolving with the computational chemistry landscape Karol Kowalski,y Raymond Bair,z Nicholas P. Bauman,y Jeffery S. Boschen,{ Eric J. Bylaska,y Jeff Daily,y Wibe A. de Jong,x Thom Dunning, Jr,y Niranjan Govind,y Robert J. Harrison,k Murat Keçeli,z Kristopher Keipert,? Sriram Krishnamoorthy,y Suraj Kumar,y Erdal Mutlu,y Bruce Palmer,y Ajay Panyala,y Bo Peng,y Ryan M. Richard,{ T. P. Straatsma,# Peter Sushko,y Edward F. Valeev,@ Marat Valiev,y Hubertus J. J. van Dam,4 Jonathan M. Waldrop,{ David B. Williams-Young,x Chao Yang,x Marcin Zalewski,y and Theresa L. Windus*,r yPacific Northwest National Laboratory, Richland, WA 99352 zArgonne National Laboratory, Lemont, IL 60439 {Ames Laboratory, Ames, IA 50011 xLawrence Berkeley National Laboratory, Berkeley, 94720 kInstitute for Advanced Computational Science, Stony Brook University, Stony Brook, NY 11794 ?NVIDIA Inc, previously Argonne National Laboratory, Lemont, IL 60439 #National Center for Computational Sciences, Oak Ridge National Laboratory, Oak Ridge, TN 37831-6373 @Department of Chemistry, Virginia Tech, Blacksburg, VA 24061 4Brookhaven National Laboratory, Upton, NY 11973 rDepartment of Chemistry, Iowa State University and Ames Laboratory, Ames, IA 50011 E-mail: [email protected] 1 Abstract Since the advent of the first computers, chemists have been at the forefront of using computers to understand and solve complex chemical problems.
    [Show full text]
  • MADNESS-HFB Nuclear DFT in 3-D Toward a Multiresolution Nuclear
    MADNESS-HFB Nuclear DFT in 3-D Toward a Multiresolution Nuclear DFT and Hartree-Fock Bogoliubov Solver Towards DFT solver based on MADNESS George Fann, Yvonne Ou and Judith Hill Oak Ridge National Laboratory/UT-JICS Junchen Pei and Witold Nazarewicz University of Tennessee/ORNL Robert Harrison Joint Faculty, ORNL/UTK UNEDF 2010 Meeting, Michigan State East Lansing, MI, June 22, 2010 The funding The development of MADNESS is funded by the U.S. Department of Energy, Office of Advanced Scientific Computing Research (OASCR) and the division of Basic Energy Science, Office of Science, under contract DE-AC05-00OR22725 with Oak Ridge National Laboratory, by the SciDAC Base Math Program and was by SAP in computational chemistry (PI: Fann) and SciDAC BES (PI:Harrison). The application of MADNESS-NDFT to nuclear physics is funded through SciDAC UNEDF (PIs: R. Lusk and W. Nazarewicz, co-PI: Fann) by the DOE- OASCR. This research was performed in part using – resources of the National Energy Scientific Computing Center which is supported by the Office of Energy Research of the U.S. Department of Energy under contract DE-AC03-76SF0098, – the Center for Computational Sciences at Oak Ridge National Laboratory under contract DE-AC05-00OR22725 . http://code.google.com/p/m-a-d-n-e-s-s MADNESS release 1, Jan. 2010 MADNESS v. 1 released in Jan. 2010. – Guaranteed accuracy with controlled precision – With examples in nuclear DFT with two-cosh potential, 3-D – Non-linear Schrodinger equation (1-D version of slda) – Time-dependent Hartree-Fock and Density
    [Show full text]
  • Computational Chemistry at the Petascale: Tools in the Tool Box
    Computational chemistry at the petascale: tools in the tool box Robert J. Harrison [email protected] [email protected] The DOE funding • This work is funded by the U.S. Department of Energy, the division of Basic Energy Science, Office of Science, under contract DE-AC05-00OR22725 with Oak Ridge National Laboratory. This research was performed in part using – resources of the National Energy Scientific Computing Center which is supported by the Office of Energy Research of the U.S. Department of Energy under contract DE-AC03-76SF0098, – and the Center for Computational Sciences at Oak Ridge National Laboratory under contract DE-AC05-00OR22725 . Chemical Computations on Future High-end Computers Award #CHE 0626354, NSF Cyber Chemistry • NCSA (T. Dunning, PI) • U. Tennessee (Chemistry, EE/CS) • U. Illinois UC (Chemistry) • U. Pennsylvania (Chemistry) • Conventional single processor performance no longer increasing exponentially – Multi-core • Many “corners” of chemistry will increasingly be best served by non-conventional technologies – GP-GPU, FPGA, MD Grape/Wine Award No: NSF CNS-0509410 Project Title: Collaborative research: CAS-AES: An integrated framework for compile-time/run-time support for multi-scale applications on high-end systems Investigators: P. Sadayappan (PI) G. Baumgartner, D. Bernholdt, R.J. Harrison, J. Nieplocha, J. Ramanujam and A. Rountev Institution: The Ohio State University Louisiana State University University of Tennessee, Knoxville Website: Description of Graphic Image: http://www.cse.ohio-state.edu/~saday/ A tree-based algorithm (here in 1D, but in actuality in 3-12D) is naturally implemented via recursion. The data in subtrees is managed in chunks and the computation expressed very compactly.
    [Show full text]
  • Wavelet-Based Linear-Response Time-Dependent Density
    Wavelet-Based Linear-Response Time-Dependent Density-Functional Theory Bhaarathi Natarajana,b,∗ Luigi Genoveseb,† Mark E. Casidaa,‡ Thierry Deutschb,§ Olga N. Burchaka,¶ Christian Philouzea,∗∗ and Maxim Y. Balakirevc†† a Laboratoire de Chimie Th´eorique, D´epartement de Chimie Mol´ecularie (DCM, UMR CNRS/UJF 5250), Institut de Chimie Mol´eculaire de Grenoble (ICMG, FR2607), Universit´eJoseph Fourier (Grenoble I), 301 rue de la Chimie, BP 53, F-38041 Grenoble Cedex 9, France b UMR-E CEA/UJF-Grenoble 1, INAC, Grenoble, F-38054, France c iRTSV/Biopuces, 17 rue des Martyrs, 38 054 Grenoble Cedex 9, France Linear-response time-dependent (TD) density-functional theory (DFT) has been implemented in the pseudopotential wavelet-based electronic structure program BigDFT and results are compared against those obtained with the all-electron Gaussian-type orbital program deMon2k for the cal- culation of electronic absorption spectra of N2 using the TD local density approximation (LDA). The two programs give comparable excitation energies and absorption spectra once suitably exten- sive basis sets are used. Convergence of LDA density orbitals and orbital energies to the basis-set limit is significantly faster for BigDFT than for deMon2k. However the number of virtual orbitals used in TD-DFT calculations is a parameter in BigDFT, while all virtual orbitals are included in TD-DFT calculations in deMon2k. As a reality check, we report the x-ray crystal structure and the measured and calculated absorption spectrum (excitation energies and oscillator strengths) of the small organic molecule N-cyclohexyl-2-(4-methoxyphenyl)imidazo[1,2-a]pyridin-3-amine. I. INTRODUCTION The last century witnessed the birth, growth, and increasing awareness of the importance of quantum mechanics for describing the behavior of electrons in atoms, molecules, and solids.
    [Show full text]
  • Ab Initio Organic Chemistry. a Survey of Ground- and Excited States and Aromaticity
    Ab Initio Organic Chemistry. A survey of ground- and excited states and aromaticity Ab initio organische chemie. Een onderzoek naar grond- en aangeslagen toestanden en aromaticiteit (met een samenvatting in het Nederlands) Proefschrift Ter verkrijging van de graad van doctor aan de Universiteit Utrecht, op gezag van de Rector Magnificus, Prof. Dr. H.O. Voorma, ingevolge het besluit van het College voor Promoties in het openbaar te verdedigen op dinsdag 28 november 2000 des ochtends te 10.30 uur door Remco Walter Alexander Havenith geboren op 24 juni 1971, te Apeldoorn. promotor: Prof. Dr. L.W. Jenneskens co-promotor: Dr. J.H. van Lenthe beiden verbonden aan het Debye Instituut van de Universiteit Utrecht. ISBN 90-393-2499-9 This work was partially sponsored by NCF, Cray Research Inc (NCF grant CRG 96.18) and the Stichting Nationale Computerfaciliteiten (NCF) (use of supercomputer facilities), with financial support from the Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO). 2 I want to believe in the madness that calls now I want to believe that a light is shining through somehow (David Bowie, The Cygnet Committee, 1969) Aan mijn ouders 3 Contents CHAPTER 1 9 GENERAL INTRODUCTION 9 1. General introduction 10 2. The application of quantum chemical calculations in organic chemistry 11 3. Subjects covered in this thesis 26 CHAPTER 2 33 A GLANCE AT ELECTRONIC STRUCTURE THEORY 33 1. The Schrödinger equation 34 2. SCF procedure 35 3. Electron correlation 36 4. The MCSCF and CASSCF methods 39 5. Magnetic properties 39 6. Valence Bond theory 40 7. Density functional theory 41 8.
    [Show full text]