A/Prof. Anthony G. Williams

1
Probing the Heart of Matter
A/Prof. Anthony G. Williams CSSMAdelaide
University 22 November 2001, Melbourne
2
Outline of Talk

• Introduction and Context
• Special Research Centre for the Subatomic
  Structure of Matter (CSSM)
• The Standard Model of Particle Physics
• Quantum Chromodynamics (QCD)
• Quarks and Gluons and the Origin of 98 of the
  Mass of Tangible Matter
• Lattice Gauge Theory and Lattice QCD
• Orion Supercomputer
• Centre for High-Performance Computing and
  Applications (CHPCA)
• Conclusions and Outlook

3
Introduction and Context

• Strongly interacting matter makes up almost the
  entire mass of the tangible universe, from the
  protons and neutrons in the nuclei of atoms and
  molecules to neutron stars.
• The strong interaction fuels the sun and stars
  and determines which nuclei are stable and hence
  which elements can exist.
• The underlying theory of the strong interaction
  is called quantum chromodynamics (QCD) and has
  quarks (like electrons) and gluons (like photons
  but self-interacting) as its fundamental
  constituents.
• Quarks and gluons are bound so strongly together
  that they can never appear as free particles.
  This is called confinement.
• When probed at increasingly higher energies the
  interaction between them becomes progressively
  weaker. This is called asymptotic freedom.
• The quarks in your body represent only about 2
  of your mass with the rest of your mass being
  generated by the strong interaction itself.
• The world's fastest supercomputers are being used
  to improve our understanding of the strong
  interaction and the unusual properties of quarks
  and gluons.

4
Concepts

• Strongly interacting matter is referred to as
  hadronic matter and strongly interacting
  particles are called hadrons, e.g., protons,
  neutrons, and pions are all hadrons
• Hadrons with 1/2 integer spin (e.g., 1/2, 3/2,
  5/2,) are fermions and are called baryons.
  Protons and neutrons are baryons.
• Hadrons with integer spin (e.g., 0, 1, 2, ) are
  bosons and are called mesons. The pion is a
  meson.
• Nuclei consist of protons and neutrons bound
  together by the strong interaction.
• The elements and hence all of chemistry is
  determined by which nuclei are stable.

Boson Bose-Einstein statistics Fermion
Fermi-Dirac statistics
5
Scales

• Typical atomic size is 10-10 m 1 Angstrom
• Typical nuclear sizes are 10-14 m 10 fermi
• 1 fermi 1 fm 10-15 m
• Proton radius is 0.8 fm,i.e., approximately 1 fm
• No substructure of electrons or quarks has ever
  been observed at resolutions down to
  approximately 1/100,000,000 Angstroms 10-18 m.
  
• At the present time we assume that electrons and
  quarks are elementary particles.

6
Atomic Structure

• Warning Sketch not to scale!
• If the protons and neutrons in this nucleus are
  10cm across then
• the nucleus is about 100cm across,
• the electrons and quarks are less than 0.1mm
  across,
• and the atom is 10km across!

7
Fundamental Forces

• There are four fundamental forces which are
  believed to give rise to all observed physical
  phenomena.
• Gravity holds us to the earth, binds stars,
  solar systems, galaxies, etc.
• Electromagnetic e.m. radiation, chemistry hence
  biology, touch, electronics, etc.
• Weak radioactivity, neutrino physics of
  supernovae, etc.
• Strong all familiar matter, nuclear energy,
  powers sun and stars, etc.

8
Special Research Centre for the Subatomic
Structure of Matter (CSSM)

• The CSSM is a Special Research Centre of the
  Australian Research Council (ARC)
• funded for nine years (since 1997) to carry out
  theoretical research in subatomic physics.
• Mission
• - Make major advances in the understanding of
  the structure of hadronic matter
• - Cross fertilization enhances opportunities for
  breakthroughs in understanding
• - Pursue lattice, models, phenomenology, and
  strong links to experimental results
• - Develop strong international links, exchanges
• Personnel
• - High level postgraduate and postdoctoral
  training
• - Interact with best researchers in a
  stimulating atmosphere
• Service
• - LANL Archive for Australia- Workshop program
  (support external students and affiliate staff)
  - Stimulate school students to science
  (brochures, school visits) - Physics Guru,
  Public Lectures, Newspapers, radio, Television

9
CSSM Current Academic Staff

• Professor Anthony Thomas (Director)
• A/Prof. Anthony Williams (Deputy Director)
• Dr. P. Coddington
• Dr. Alex Kalloniatis (Australian Research Fellow)
• Dr. Derek Leinweber
• Dr. Andreas Schreiber (Australian Research
  Fellow)
• Dr. Ingo Bojak
• Dr. Xin-Heng Guo
• Dr. Ayse Kizilersü
• Dr. Vadim Guzey
• Dr. Martin Oettel (joint appointment Alexander
  von Humbolt Stiftung/Foundation CSSM)
• Dr. Jianbo Zhang

10
Postgraduate Students

• Sundance Bilson-Thompson (Ph.D.). Supervisors
  D.B. Leinweber A.G. Williams
• Francois Bissey (Ph.D.). Cotutelle (U. Blaise
  Pascal) A.W. Thomas.
• Frederic Bonnet (Ph.D.). Supervisors D.B.
  Leinweber A.G. Williams.
• Patrick Bowman (Ph.D.). Supervisors D.B.
  Leinweber A.G. Williams position at Florida
  State Univ.
• Shane Braendler (Ph.D.). Supervisor A.W.
  Thomas
• Will Detmold (Ph.D.). Supervisors A. Bender
  A.W. Thomas
• Emily Hackett-Jones (M.Sc.). Supervisors D.B.
  Leinweber A.W. Thomas
  

• Waseem Kamleh (Ph.D.). Supervisor A.G.
  Williams
• Daniel Kusterer (M.Sc. Baden-Wuerttemberg/S.A.
  Exchange Program) - Supervisor D.B. Leinweber
• Olivier Leitner (Ph.D.). Cotutelle (U. Blaise
  Pascal) A.W. Thomas
• Tom Sizer (Ph.D.). Supervisor A.G. Williams
• Stewart Wright (Ph.D.). Supervisors D.B.
  Leinweber A.W. Thomasposition at Liverpool,
  UK.
• Ross Young (Ph.D.). Supervisors D.B. Leinweber
  A.W. Thomas
• James Zanotti (Ph.D.). Supervisors D.B.
  Leinweber A.G. Williams

11
International Collaborative Agreements

• Abdus Salam International Centre for Theoretical
  Physics - Italy
• Argonne National Laboratory - USA
• Bonn University - Germany
• Chinese Academy of Sciences (Beijing) - China
• Commissariat à l'Energie Atomique - France
• European Centre for Theoretical Studies in
  Nuclear Physics and Related Areas (Trento) -
  Europe
• Indiana University (Bloomington) - USA
• Institute for Nuclear Theory, University of
  Washington (Seattle) - USA
• Instituto De Fisica Teòrica (IFT-UNESP) - Brazil
• Joint Institute for Nuclear Research (JINR -
  Dubna) - Russia
• Jülich (FZ) - Germany

12
International Collaborative Agreements (cont.)

• MESON (Medium Energy Science Open Network)
  involving IUCF Indiana Yonsei, Korea RCNP
  Osaka KVI Groningen IMP Lanzhou TSL Uppsala
  NAC Cape Town SAHA Calcutta FZ Jülich CIAE
  Beijing
• Osaka University - Japan
• Computational Science and Information Technology
  (CSIT, Florida) - USA
• Svedberg Laboratory - Sweden
• Thomas Jefferson National Accelerator Facility
  (Newport News) - USA
• Technical University of Munich - Germany
• TRIUMF (Vancouver) - Canada
• Université Blaise Pascal - France
• University of Tübingen - Germany

13
CSSM Workshops/Conferences 2000

• 3rd International Symposium on Symmetries in
  Subatomic Physics - MarchTotal Number of
  Participants 85Overseas Participants
  45Interstate Participants 12Local
  Participants 28
• International Conference on Quark Nuclear Physics
  - FebruaryTotal Number of Participants
  109Overseas Participants 75Interstate
  Participants 1Local Participants 33
• HallD Workshop - FebruaryTotal Number of
  Participants 45Overseas Participants
  24Interstate Participants NilLocal
  Participants 21

14
CSSM Workshops/Conferences 2001

• Lattice Hadron Physics Workshop - JulyTotal
  Number of Participants 42Overseas Participants
  26Interstate Participants 1Local Participants
  15
• Hamiltonian Lattice Gauge Theories Workshop -
  AprilTotal Number of Participants 19Overseas
  Participants NilInterstate Participants
  4Local Participants 15
• Leptonic Scattering Workshop - MarchTotal Number
  of Participants 64Overseas Participants
  35Local Participants 29

15
Visitor Program 2001

• Dr. P. Bowman, Florida, USA
• Dr. A. Chian, Sao Paolo, Brazil
• Dr. M. Chaichian, Helsinki
• Dr. G. Dunne, Connecticut
• Dr. Y. Hoshino, Kushiro, Japan
• Dr. C.-S. Huang, Beijing, China
• Prof. A. Ioannides, RIKEN, Japan
• Dr. S. Krewald, Jueilich, Germany
• Dr. R. Landau, Oregon, USA
• Dr. D. Lu, Zhejiang, China
• Prof. W.-X. Ma, Beijing, China

• Dr. K. Maltman, York, Canada
• Dr. S. Sharpe, Seattle, USA
• Dr. A. Signal, Massey, NZ
• Dr. D. Sinclair, Argonne, USA
• Prof. J. Speth, Juelich, Germany
• Dr. P. Tandy, Ohio, USA
• Dr. G. Valencia, Iowa, USA
• Prof. M. Veltman, Utrecht
• Dr. J. Vergados, Ionnina, Greece
• Dr. L. von Smekal, Erlangen
• Dr. M. Weyrauch, Bundesanstalt, Germany

16
Future Workshops

• Joint Workshop with JHF, March 2002
• The Structure of the Nucleon (Joint with ECT in
  Trento), September 2002
• NUPP Summer School, February 2003
• 2nd Lattice Workshop, Cairns, June/July(?) 2003

17
The Standard Model
Let us review some aspects of the standard model
briefly before beginning to focus on QCD itself.
18
The Standard Model Fermions

• In addition to the 6 known flavors of quarks they
  come in 3 colours red, blue, and green
• Lepton comes from the Greek for small mass
• Leptons do not carry color charge, i.e., they do
  not feel the strong force

19
The Standard Model Bosons

• The very massive W-, W, and Z0 bosons mediate
  the weak interaction, which as a result is very
  short range
• The massless photon mediates the long-range e.m.
  interactions
• Gluons carry color and mediate the strong
  interaction

20
The Standard Model Forces

• Gravitons are thought to mediate the
  gravitational force but have not yet been seen
• Gravitational waves are to gravitons what e.m.
  radiation is to photons
• Above we see the relative strengths and relative
  ranges of the four fundamental forces

21
Standard Model Sample Processes

• Standard model processes and interactions
• neutron beta decay (neutrons are only stable in
  nuclei which is just as well!) - imagine the
  universe if this was not so ...
• electron-positron annihilation to meson-antimeson
  pair
• proton-proton collision producing two Z0 bosons
  and other hadrons

22
Early Universe

• Free quarks and gluons existed until about 10-5
  seconds
• atoms formed at about 300,000 years
• stars formed at about 1 billion years
• solar systems and life at about 12 billion years

23
Neutron Stars

• Different phases of hadronic matter can co-exist
  within a neutron star.
• For this sample neutron star, it is expected that
  quark matter becomes a stable phase 1km beneath
  the surface.
• In the central core quark matter is dominant.

24
Neutron Star Phase vs Density

• Protons and neutrons are collectively referred to
  as nucleons
• Ordinary nuclear matter density is approximately
  0.17 nucleons/fm-3

?
25
Quantum Chromodynamics (QCD)

• All hadrons are color-singlet (white)
• Baryons contain three quarks (red blue green)
  - Different baryons arise from the three quarks
  having different flavor combinations.
• Mesons contain colour anticolour combinations
  of quark and antiquark pairs (red anti-red,
  green anti-green, blue anti-blue) - Different
  mesons arise from different flavor combinations
  of the pair.
• Each flavor of quark cycles through the three
  colors by exchanging gluons with the other quarks
  or anti-quark.

26
QCD Nuclear Forces
All hadrons are color-singlet (neutral or
colorless or white) combinations of quarks (3
quarks for baryons or quark-antiquark pair for
mesons). But just as electrically neutral atoms
interact by van der Waals forces, so can color
neutral particles. This has very important
consequences the strong nuclear force.

• and hence to atoms, molecules, chemistry, and
  all of biology.

27
QCD Baryons and Antibaryons

• Baryons are color-singlet combinations of three
  quarks.
• Anti-baryons are color-singlet combinations of
  three anti-quarks.

28
QCD Mesons

• Mesons are made up of a quark - anti-quark pair
  with equal and opposite color charges giving a
  color-singlet.

29
Confinement

• The strong force between two quarks arises from
  the exchange of gluons.
• There is also a strong force between two gluons -
  since they carry color themselves they can
  interact with each other. This is not the case
  for photons in QED since they carry no electric
  charge.
• The force between two quarks is constant, i.e.,
  independent of their separation. This
  corresponds to a linearly rising potential
  between them, which is referred to as a string
  tension.
• This force between colored objects is equivalent
  to a wieight of approximately 10 tons! - This is
  why no free quarks or gluons are ever seen. QCD
  has the property of CONFINEMENT.

30
String Breaking Quark - Anti-quark Pairs

• What happens when we try to pull apart two
  colored objects by pumping more and more energy
  into the system, e.g., through energetic
  collisions at a particle accelerator?
• When the energy put in is large enough to make a
  quark - anti-quark pair, then a meson can be
  created and the string is broken.

• Confinement survives however as no free color
  particles are produced.

31
Asymptotic Freedom

• In quantum electrodynamics (QED) the fine
  structure constant is ? ? 1/137.
• In QCD the coupling runs by decreasing with
  increasing energy scale, i.e., at short
  distances.
• At the energy scale of the Z0 mass the strong
  coupling constant has decreased to a value ?s
  ?0.12.
• Quarks and gluons appear almost free at high
  energies.
• At low energies ?s ? 1, i.e., the coupling is
  very STRONG and perturbation theory fails..

ASYMPTOTIC FREEDOM
32
Strong Coupling Constant at Z Mass

• At the energy scale of the Z0 mass the strong
  coupling constant has decreased to ?s ?0.12.
• This has been confirmed in a variety of different
  experiments with a remarkable degree of
  consistency.

33
QCD Produces 98 of Your Mass

• Proton mass is 938 MeV 0.938 GeV.
• Up and down quark masses are approximately 3 and
  6 MeV respectively.

Where does the rest of the hadron mass come from?

• The interactions in the low-energy
  (long-distance) are so strong (i.e.,
  non-perturbative) that they induce a mass in the
  quarks of approximately 300 MeV. This is
  referred to as dynamical mass generation.
  This is how three quarks in a bound state can
  have a mass exceeding their naïve sum.
• The fact that the up and down quarks are so light
  and because of this mass-generation, Goldstones
  theorem states that there will be nearly massless
  Goldstone bosons associated with the spontaneous
  breaking of this symmetry - these Goldstone
  bosons are the pions which are anomalously light
  mesons. This is the basis of what is called
  chiral symmetry and dynamical chiral symmetry
  breaking.

34
Lattice Gauge Theory

• Physicists at the CSSM and elsewhere use the
  techniques of lattice gauge theory to put all of
  four-dimensional space-time on a grid or lattice.
• We formulate the gauge field theory (e.g., QED,
  QCD, etc.) on this discrete lattice with
  finite-difference techniques. This is done in
  Euclidean space-time for numerical reasons. We
  use methods adapted from statistical mechanics to
  study the physical properties of the theory from
  the confining to the asymptotically free regimes.
  
• We use this to study and model subatomic
  particles and their interactions. It is VERY
  computationally expensive to do this -
  Teraflop-years of computer time are needed.
• More accurate results need a finer lattice with
  larger space-time volumes. This is what makes it
  computationally costly.

35
Lattice QCD
Depiction of electron scattering from quark
through a virtual photon exchange in a
background gluon field. Lattice QCD does
weighted averages over gauge field configurations
to obtain physical quantities of interest.
36
Lattice QCD Gluon Configurations

• Typical gluon field configuration used in lattice
  calculations, (3-dim slice of 4-dim lattice
  showing action density, where red depicts
  highest).
• The estimate of the integral over gluon fields
  does a weighted average over hundreds of these.

37
High-Performance Computing

• Traditionally supercomputers were very expensive,
  contained purpose-built hardware, and were
  obsolete within 5 years.
• Most modern high-performance computing uses
  cost-effective clusters of mass-produced
  commodity off-the-shelf (COTS) hardware, rather
  than expensive proprietary hardware.
• These clusters of workstations or PCs are
  connected either by commodity Fast Ethernet or
  Gigabit switched networking or by (more
  expensive) special ultrafast, low-latency
  networks for clustering (e.g., Myrinet,
  ServerNet, GigaNet, SCI, etc).
• Clusters are flexible in their design and easy to
  build and upgrade - e.g., add more nodes upgrade
  some or all of the nodes upgrade some or all of
  the networking.
• Can get an order of magnitude better
  price/performance ratio!

38
Top 500 Supercomputers

• Clusters of PCs or Unix workstations have become
  incredibly popular in the last few years --
  ranging from a few networked PCs to around 1000
  Compaq Alpha workstations.
• The unit used to measure the performance of
  computers is the flop, i.e., 1 flop 1
  floating-point operation per second.
• In other words 1 flop 1 calculation per
  second.
• The 12 fastest supercomputers in the world all
  exceed 1 Teraflop, i.e., 1 Teraflop 1,000
  Gigaflops 1,000 billion calculations per
  second.
• The current fastest is 7.2 Teraflops, (cluster
  with 8,192 CPUs called ASCI White at Lawrence
  Livermore National Laboratory in the USA- models
  nuclear explosions in place of nuclear tests).
• Three of the top 4 machines serve this same
  purpose.

39
Top 500 Supercomputers (cont.)

• The fastest supercomputer in Australia at present
  is the NEC SX-5/32M2 at the CSIRO Bureau of
  Meteorology in Melbourne. It is number 99 in the
  world and is 241.4 /(256 peak) Gigaflops.
• The current second fastest is the APAC facility
  at ANU in Canberra, which is a Compaq alpha
  workstation cluster and is number 134 in the
  world and 167.5/(245 peak) Gigaflops. It will
  very soon be upgraded to 980 peak Gigaflops
  (almost a Teraflop peak) and at that time will be
  Australias fastest.
• The third is a similar Compaq cluster run by VPAC
  in Melbourne and is number 151 at 149.1/(213
  peak) Gigaflops.
• Fourth is our own Sun cluster in Adelaide, called
  Orion, which is number 246 at 110/(144 peak)
  Gigaflops.

40
Cluster Computing Architectures

• Many possible design choices in building a
  compute cluster -- depends on type of application
  (or applications).
• Unix workstations or Beowulf PC cluster?
• What processor, clock speed, cache, bus speed,
  etc?
• Single processor or Shared-Memory Processor (SMP)
  nodes?
• Linux or commercial operating system (OS)?
• How much memory and disk per node?
• Buy from vendor or systems integrator, or build
  it yourself?
• What networking technology to use?
• Commodity Fast or Gigabit switched ethernet is
  relatively cheap, but has high latency.
• Low-latency, ultra-fast networks are
  significantly more expensive, but far superior.

41
Orion Supercomputer

• Orion is a Sun Technical Compute Farm with 40 Sun
  E420R 4-way SMP nodes (160 processors). Fast
  switched ethernet AND high-speed Myrinet network.
  
• 110/(144 peak) Gflops, 160 Gigabytes RAM, 640
  Megabytes cache memory.
• Fastest computer in Australia when installed in
  June 2000 and now number 4.
• The CSSM together with lattice theorists from
  UNSW and the University of Melbourne obtained
  RIEF funds to seed the National Computing
  Facility for Lattice Gauge Theory (NCFLGT) which
  houses the Orion supercomputer.

42
Orion Software

• Nodes run Solaris and standard Sun software and
  compilers.
• Sun HPC ClusterTools includes
• Sun Cluster Runtime Environment (CRE)
• Sun MPI, optimized for SMP cluster (and recently
  for Myrinet)
• Sun Scientific Software Library (S3L)
• Prism debugger and performance analysis tool
• S3L and Prism are developed from CM software.
• Sun Fortran 90/95 compiler, supports automated
  parallelisation and OpenMP directives for shared
  memory parallelism.
• Portland Group HPF compiler, converts code to Sun
  F77/90 plus MPI.
• Sun Grid Engine (formerly CODINE) cluster
  management system is more advanced than CRE,
  supports batch queueing, detailed logging of
  system usage, access levels, etc.

43
Centre for High-Performance Computing and
Applications (CHPCA)

• The CHPCA is a newly created cross-disciplinary
  Research Centre at Adelaide University. Its goal
  is to bring together all researchers with an
  interest and commitment to High-Performance
  Computing (HPC) in order to share expertise and
  to develop a very large shared computing
  platform. Director AGW Deputy Directors Paul
  Coddington (DHPC), Derek Lienweber (CSSM), and
  Francis Vaughan (SAPAC).
• Partner researchers include Physics, Chemistry,
  Engineering, Biology and Bioinformatics, Plant
  Science, Geology and Geophysics, Water Resource
  Management, etc.
• Next year (2002) the CHPCA already has enough
  funds from its members and partners to construct
  a 250 Gigaflop cluster consisting of 40 dual
  processor 2 GHz Pentium 4 nodes with Myrinet 2000
  cluster networking.
• The goal is to raise funds to turn this into a
  Teraflop supercomputer.

44
CHPCA Engineering Applications

• Computational fluid dynamics.
• Drag and noise reduction in planes, ships,
  submarines, etc.
• Petroleum geology, oil and gas reservoir
  modeling.
• Flow through porous media.
• Water quality management and salinity.
• Optimization of distribution systems for water
  piplines, power lines, and telecommunication
  networks.
• Optimization of engineering design over a large
  parameter space
• search over multiple parameters using task farm
  approach, many sequential jobs each with
  different parameters

45
CHPCA Biological Applications

• Bluegene - IBM plans to build a 1 Petaflop
  computer to study fundamental problems in
  computational biology and protein science - (1
  Petaflop 1,000 Teraflops!) - 1 million
  processors and each processor with multiple CPUs
  and memory and communication logic built in.
• Bluegene will focus on protein folding in
  particular.
• Modelling heart and brain function, organ and
  arterial simulation - chaos, heart fibrillation,
  epileptic seizures, etc.
• The virtual human project - Oak Ridge National
  Laboratory.
• DNA and protein sequence analysis and
  classification.
• Genome data and bioinformatics - data farming
  for efficient storage, recovery, searching and
  matching of vast biological data, e.g., for
  efficient drug design, etc.

46
Conclusions and Outlook

• This is an exciting time for theoretical
  subatomic physics in Australia.
• The CSSM is continuing to build on its successes
  and is highly productive.
• The Australian lattice QCD program is now
  well-established and runs a world-class
  supercomputing facility.
• With the establishment of a cross-disciplinary
  CHPCA the path to a Teraflop computer and a
  world-class resource for the coming years is
  becoming a reality.
• We look forward to new discoveries and new
  opportunities in subatomic physics over the next
  few years.
• Cross-disciplinary activity in High-Performance
  Computing research is becoming increasingly
  important for the HPC field in order that
  research areas that depend on it continue to
  thrive. - We are working on that.
• Thank you for your attention.

47
End of Presentation
48

• Fdhhsh
• sfhg
• sdfgsdgsdfg
• sdfgsdfgsdfg
• sdfgsdfgsdf
• sdfgsdfgsdfg

49
Equations pasted from latex
50
Equations pasted from latex
51
Animation