1 / 14

Dynamical overlap fermions in the ε-regime

Dynamical overlap fermions in the ε-regime. Hidenori Fukaya (RIKEN), T.Kaneko, S.Hashimoto, K.Kanaya,H.Matsufuru, K.Ogawa,M.Okamoto, T.Onogi, N.Yamada,for JLQCD collaboration. Contents. Introduction Numerical cost Static quark potential with almost massless sea quarks

valin
Télécharger la présentation

Dynamical overlap fermions in the ε-regime

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Dynamical overlap fermions in the ε-regime Hidenori Fukaya (RIKEN), T.Kaneko, S.Hashimoto, K.Kanaya,H.Matsufuru, K.Ogawa,M.Okamoto, T.Onogi, N.Yamada,forJLQCD collaboration

  2. Contents • Introduction • Numerical cost • Static quark potential with almost massless sea quarks • Low eigenmodes • Pion correlators in the ε-regime • Summary and discussion

  3. Introduction • Standard lattice QCD simulations require • V→∞, m→0, a→0 limits. • In the ε-regime ( 1/ΛQCD < L < 1/mπ), • Finite V effects are calculated on ChPT side. • m→0 can be reached in small V. • Low-energy constants of ChPT can be extracted without chiral extrapolation. • m→0 ⇒ Chiral symmetry is crucial ⇒ The overlap fermions Neuberger,Phys.Lett.B417(1998)141,Phys.Lett.B427(1998)353

  4. Introduction • JLQCD’s dynamical overlap fermion project • Iwasaki gauge action + Nf=2 overlap fermions • Q=0 topological sector only. • β~2.3-2.4 (1/a~1.69 -1.9GeV) . • Lattice size = 163×32 (L~1.8fm with β=2.35). • Hasenbusch acceleration. [Hasenbusch,Phys.Lett.B519(2001)177] • Sea quark mass is 2MeV – 120MeV (with an assumption Zm ~ Zmquenched=1.8 ). • Details →Matsufuru’s talk (Algorithm part2 15:40 today). • m~2.0MeV on L~1.8fm is in the ε-regime.

  5. Introduction – references - • ε-regime in ChPT • Gasser & Leutwyler, Phys.Lett.B188(1967)477, Nucl.Phys.B307(1988)763 • Hansen, Nucl.Phys.B345(1990)685 • Hansen & Leutwyler, Nucl.Phys.B350(1991)201 • Leutwyler & Smilga, Phys.Rev.D46(1992)5607 • Sharpe, Phys.Rev.D46(1992)3146 • Damgaard, Diamantini,Hernandez & Jansen, Nucl.Phys.B629(2002)445, • Damgaard,Hernandez,Jansen,Laine & Lellouch,Nucl.Phys.B656(2003)226 • Hernandez & Laine, JHEP301(2003)063, hep-lat/0607027

  6. Introduction – references - • Quenched lattice • Prelovsek & Orginos,Nucl.Phys.Proc.Suppl.119(2003)822 • Giusti,Hoelbling,Luscher & Wittig,Compt.Phys.Commun.153(2003)31 • Giusti,Luscher,Weisz & Wittig,JHEP0311(2003)023 • Bietenholz,Chiarappa,Jansen,Nagai & Shcheredin, JHEP0402(2004)023 • Giusti,Hernandez,Laine,Weisz & Wittig,JHEP0401(2004)003 • Giusti,Hernandez,Laine,Weisz & Wittig,JHEP0404(2004)013 • Ogawa & Hashimoto,Prog.Theor.Phys.114(2005)609 • HF,Hashimoto & Ogawa,Prog.Theor.Phys.114(2005)451 • Mehen & Tiburzi,Phys.Rev.D72,014501(2005) • Bietenholz & Shcheredin,PoSLAT2005,138(2006) • Giusti & Necco,PoSLAT2005,132(2006) • Damgaard,Heller,Splittorff & Svetitsky,Phys.Rev.D72,091501(2005) • Bietenholz & Shcheredin, hep-lat/0605013

  7. Introduction – references - • Nf=2 overlap fermion (m~35-100MeV) • DeGrand & Schaefer, Phys.Rev.D72,054503 • Nf=2 staggered fermion with isospin chemical potential • Damgaard,Heller,Splittorff,Svetitsky, & Toublan,Phys.Rev.D73,074023(2006) →Plenary talk by Splittorff. • Related talks –overlap fermions in JLQCD- • Kaneko -> overview of JLQCD overlap fermion project. • Yamada-> Locality of dynamical overlap quarks. • Hashimoto-> dynamical overlap at a fixed topology. • Okamoto-> pion mass, decay constant. • Matsufuru-> algorithm

  8. Numerical cost • CG iteration ∝|λmax + m|2/ |λmin+ m|2. • In large V regime,λmin ~ 1/(ΣV) ~ 0 →CG iteration ∝1/m2. • In the ε-regime, λlow feels repulsive force from 0; Chiral Random Matrix Theory (RMT) indicates 〈 λmin 〉ΣV ~ 4.34 >> mΣV [Nf=2,Q=0] →CG iteration ∝ 1/ λ2minRMT (independent of m). • But auto-correlation time seems longer. • On BlueGene/L 512node (2.8Tflops), 1trj / 1h , autocorrelation ~ 100 trj • We’ve generated 1400trj (100 confs) in Q=0 sector. (400trj was discarded for thermalization.)

  9. Static quark potential with almost massless quarks • String breaking at m~2MeV? → No indication. • Too small V for two static-light mesons? • Wilson loop has small overlap with static-light meson operators? • Anyway, a ~ 0.11fm [Sommer scale r0~0.49fm].

  10. Nf=2,Q=0 (Nf=0,Q=2) Low eigenmodes - compared with ChRMT - • λ’ ≡ Im λ/ (1- Re λ/2) (λ is eigenvalue of Dov ) is compared with ChRMT. [λ was calculated by Lanczos algorithm with Chebyshev acceleration] • The lowest mode should feel stronger repulsive force from 0 (Nf=2, m=0); 〈λmin〉ΣV ~ 4.34. than quenched case; 〈λmin〉ΣV ~ 1.77. • The ratio 〈λi〉/〈λj〉 (λi: i-th eigenvalue) shows a good agreement in both quenched and Nf=2 cases. • m dependence is also interesting… • From λmin, Σ1/3= 228.9(3.6) MeV [bare lattice value] is obtained.

  11. Pion correlators in the ε-regime • Pion correlator in the ε-regime ChPT is not exponential but quadratic; • Low-mode averaging DeGrand & Schaefer,Comp.Phys.Commun. 159 (2004)185, Giusti,Hernandez,Laine,Weisz & Wittig,JHEP0404(2004)013 is very effective when m→0 since ~90% contribution is from 100 lowest-modes, which are obtained with Lanczos algorithm. Higher mode contribution is calculated with a point source, using low-mode preconditioning.

  12. Pion correlators in the ε-regime • With an input; Σ= 228.9(3.6) MeV , from lowest eigenvalue, quadratic fit (fit range=[10,22],β1=0) worked well. [χ2 /dof ~0.25.] • Fπ = 86.2(6.9)MeV is obtained [preliminary]. • Error from Σ is not taken into account. • Consistency with the other correlators (axial, scalar etc.) would be important. • Fitting range is appropriate? • O(a2) effects? • Higher order ChPT contribution?

  13. Summary and discussion • Nf=2 dynamical overlap fermion at fixed topology can reach m~2MeV on a 16332 lattice (a~0.11fm, L~1.8fm). CG iteration ≠1/m2 but 1/(λRMT+m)2 • Wilson loops shows no indication of string breaking. • The eigenvalue distribution is consistent with ChRMT in the ε-regime⇒ Σ • Sea quark mass dependence of low-modes was also observed. • Low-mode averaging is effective to evaluate the meson correlators. ⇒ Fπ

  14. Summary and discussion For future works, • The other connected meson correlators and condensates with higher accuracy. • Q dependence. In fact, Q≠0 sector would be easier. ( λmin is higher after projecting out 0-modes.) • Disconnected correlators with low-mode averaging. • Baryon correlators. • Including chemical potentials to valence quark to extract Fπ. • String breaking ?? • ρ→ππ decay ??? • …

More Related