Leading the World QCD from QCDOC

1
Leading the World QCD from QCDOC

• Richard Kenway

2
the computational problem

• path integral over anti-commuting quark fields
  and SU(3) gauge link variables

• we cannot measure experimentally quark masses or
  mixings, only hadron masses and decays

• simulations at physical parameter values are too
  expensive
• use effective theories to extrapolate simulation
  results from parameter regimes where systematic
  errors can be controlled to the physical regime

3
state of the art
wrong!
light and heavy hadron results are consistent
with experiment when u, d and s sea quarks are
included
CTH Davies et al. (UKQCD HPQCD/MILC),
PRL92022001 (2004)
4
determination of QCD parameters
Q Mason (Lat05)
Q Mason et al., PRL95052002 (2005)
Strong coupling
5
experimental impact

• calculations of the weak decays of b quarks
  inside B mesons are critical for determinations
  of CKM matrix elements

6
the computer
Gert Aarts Chris Allton Gunnar Bali Biagio
Lucini Ken Bowler Peter Boyle Christine
Davies Luigi Del Debbio Ian Drummond Jonathan
Flynn Simon Hands Alistair Hart Ron Horgan Roger
Horsley Tony Kennedy Richard Kenway Alan
Irving Chris Michael Brian Pendleton Paul
Rakow Chris Sachrajda Mike Teper John Wheater
Edinburgh 16 Nov 2004 14,336 processors, 11.5
teraflops
7
actions and algorithms

• there is freedom to choose different lattice QCD
  actions
• all have the same classical continuum limit
• at non-zero lattice spacing, symmetries and
  discretisation effects may differ
• as well as the cost
• we are using two choices of lattice action with
  21 flavours
• improved staggered to extend state-of-the-art
  results to smaller quark masses, finer lattice
  spacings and more physical quantities
• domain wall to have correct chiral and flavour
  symmetries at non-zero lattice spacing and to
  compute a wider range of matrix elements
• we have developed an exact algorithm (RHMC) for
  odd numbers of flavours

8
first results

• domain wall parameter search
• for finite 5th dimension, chiral symmetry is
  broken by an additive quark mass

UKQCD and RBC (Lat05)

• production L 3 fm, a-1 1.8(1) GeV, mud
  ms/4, mres mud/5
• ms, BK, pion and nucleon structure, nucleon
  decay, c and b physics

9
leading the world

• improved staggered results are consistent with
  experiment and are driving towards few percent
  uncertainties
• comparison with domain wall results will test our
  control of chiral and flavour symmetries at
  non-zero lattice spacing
• a wider range of matrix elements will then be
  available because mixing and renormalisation will
  be easier to control
• falsification of the Standard Model will then be
  possible

we are entering unknown territory