Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userReversibility Violation in the Hybrid Monte Carlo Algorithm
150
0
0.0
(
0
)
Authors
Carsten Urbach
Ask ChatGPT about the research
No Arabic abstract
We investigate reversibility violations in the Hybrid Monte Carlo algorithm. Those violations are inevitable when computers with finite numerical precision are being used. In SU(2) gauge theory, we study the dependence of observables on the size of the reversibility violations. While we cannot find any statistically significant deviation in observables related to the simulated physical model, algorithmic specific observables signal an upper bound for reversibility violations below which simulations appear unproblematic. This empirically derived condition is independent of problem size and parameter values, at least in the range of parameters studied here.
rate research
Read More
UKQCDs dynamical fermion project uses the Generalised Hybrid Monte-Carlo (GHMC) algorithm to generate QCD gauge configurations for a non-perturbatively O(a) improved Wilson action with two degenerate sea-quark flavours. We describe our implementation of the algorithm on the Cray-T3E, concentrating on issues arising from code verification and performance optimisation, such as parameter tuning, reversibility, the effect of precision, the choice of matrix inverter and the behaviour of different molecular dynamics integration schemes.
We present a polynomial Hybrid Monte Carlo (PHMC) algorithm as an exact simulation algorithm with dynamical Kogut-Susskind fermions. The algorithm uses a Hermitian polynomial approximation for the fractional power of the KS fermion matrix. The systematic error from the polynomial approximation is removed by the Kennedy-Kuti noisy Metropolis test so that the algorithm becomes exact at a finite molecular dynamics step size. We performed numerical tests with $N_f$$=$2 case on several lattice sizes. We found that the PHMC algorithm works on a moderately large lattice of $16^4$ at $beta$$=$5.7, $m$$=$0.02 ($m_{mathrm{PS}}/m_{mathrm{V}}$$sim$0.69) with a reasonable computational time.
We apply the UV-filtering preconditioner, previously used to improve the Multi-Boson algorithm, to the Polynomial Hybrid Monte Carlo (UV-PHMC) algorithm. The performance test for the algorithm is given for the plaquette gauge action and the $O(a)$-improved Wilson action at $beta=5.2, c_{mathrm{sw}}=2.02, M_{pi}/M_{rho}sim 0.8$ and 0.7 on a $16^3times 48$ lattice. We find that the UV-filtering reduces the magnitude of the molecular dynamics force from the pseudo fermion by a factor 3 by tuning the UV-filter parameter. Combining with the multi-time scale molecular dynamics integrator we achieve a factor 2 improvement.
We present a polynomial hybrid Monte Carlo (PHMC) algorithm for lattice QCD with odd numbers of flavors of O(a)-improved Wilson quark action. The algorithm makes use of the non-Hermitian Chebyshev polynomial to approximate the inverse square root of the fermion matrix required for an odd number of flavors. The systematic error from the polynomial approximation is removed by a noisy Metropolis test for which a new method is developed. Investigating the property of our PHMC algorithm in the N_f=2 QCD case, we find that it is as efficient as the conventional HMC algorithm for a moderately large lattice size (16^3 times 48) with intermediate quark masses (m_{PS}/m_V ~ 0.7-0.8). We test our odd-flavor algorithm through extensive simulations of two-flavor QCD treated as an N_f = 1+1 system, and comparing the results with those of the established algorithms for N_f=2 QCD. These tests establish that our PHMC algorithm works on a moderately large lattice size with intermediate quark masses (16^3 times 48, m_{PS}/m_V ~ 0.7-0.8). Finally we experiment with the (2+1)-flavor QCD simulation on small lattices (4^3 times 8 and 8^3 times 16), and confirm the agreement of our results with those obtained with the R algorithm and extrapolated to a zero molecular dynamics step size.
In Hybrid Monte Carlo(HMC) simulations for full QCD, the gauge fields evolve smoothly as a function of Molecular Dynamics (MD) time. Thus we investigate improved methods of estimating the trial solutions to the Dirac propagator as superpositions of the solutions in the recent past. So far our best extrapolation method reduces the number of Conjugate Gradient iterations per unit MD time by about a factor of 4. Further improvements should be forthcoming as we further exploit the information of past trajectories.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
