The η′ and cooling with staggered fermions

Embed Size (px)

Text of The η′ and cooling with staggered fermions



    Nuclear Physics B (Proc. Suppl.) 47 (1996) 358-361


    The and Cooling with Staggered Fermions G. Kilcup, J. Grandy and L. Venkataraman ~

    ~Department of Physics, The Ohio State University, 174 West 18th Ave, Columbus, Ohio 43210

    We present a calculation of the mass of the *7' meson using quenched and dynamical staggered fermions. We also discuss the effects of "cooling", and suggest its use a quantitative tool.

    1. q' Lore

    One of the more curious inhabitants of the low energy spectrum of QCD is the q, meson. Deriving part of its mass from its kinship with the Goldstone boson pseudoscalars, the q' owes most of its anomalous weight to its connection to the topological susceptibility. Specifically, in an SU(3) symmetric world, one would write

    2 _-m~+ra0 2 (1)

    where ms is the mass common to all of the octet meson, and m0 is peculiar to the rf. In the chi- ral limit, the octet mass ms vanishes like x/-m-q, while m0 remains finite. Identifying ms 2 with the average mass-squared of the physical rr's, K's and the r/, one derives the "experimental" number of m0 -- 860 MeV for N/ = 3. To complete the cal- culation of the ground state hadron spectrum, one would like to compute this number from a first principles lattice simulation, free of assumptions such as the validity of the large N expansion.

    In the language of quarks, the rf is understood to gain its mass from "hairpin" diagrams, where the quarks annihilate into glue and then recon- stitute themselves. In the language of mesons, this phenomenon can be described by ascribing a strength m~ to this interaction of flavor singlet mesons. Iteration of this process leads to a ge- ometric series which shifts the pole in the prop- agator from ms 2 to m~ + m02. In the quenched approximation, the series terminates with the sec- ond term. The quenched ,7' propagator then has the unphysical form of a sum of a pole and a dou- ble pole, the infrared divergences of which lead to extra quenched chiral logs. It is therefore of inter- est to see if one can distinguish pole and double pole behavior in a lattice simulation.

    R(O =

    0 t

    0 t

    Figure 1. Diagrams for the q' propagator.

    The quantity of interest is the ratio of the dis- connected and connected diagrams.

    R(t) = 2-'P (0'(0)~'(t)>l-,oop (2)

    If we compute the ratio using Nv~l valence quarks and Ndyn dynamical quarks, with the same quark mass for valence and dynamical, then R(t) should asymptote to

    n( t ) - - , gw, [1 - Z' Ndy---'-~ ~- exp(--tAm)] (3)

    ' and Z' and Z are the where Am = rn 0 --ms, residues for the creation of the singlet and octet particles, respectively. For Ndyn ---- O, i.e. in the quenched approximation, R(t) should never asymptote, but instead rises linearly with slope 2m ~o / ms.

    This approach has been taken with quenched configurations and valence Wilson fermions in ref. [1]. Here we extend the analysis to staggered fermions, and use quenched and dynamical con- figurations.

    2. Imp lementat ion

    This calculation used three ensembles of lat- tices of size 163 32, one quenched and two in-

    0920-5632/96/$15.00 1996 Elsevier Science B.V. All rights reserved. Plh S0920-5632(96)00073-4

  • G. Kilcup et aL /Nuclear Physics B (Proc. Suppl.) 47 (1996) 358-361 359

    cluding dynamical quarks. The dynamical ensem- bles were "borrowed" from the Columbia group, and have been used previously in our BK analysis [4], and the calculation of fB by the MILC collab- oration [5]. These were generated using molec- ular dynamics with N l = 2 staggered fermions with quark masses mdy n : .01 and .025. The quenched lattices were generated on the Cray- T3D at the Ohio Supercomputer Center (OSC), using 3-subgroup SU(2) updates, with 4 to 1 mix- ture of over-relaxed and heatbath. We saved con- figurations with 2000 sweeps, and obtained an av- erage update time of 5 microseconds per link on 16 processors.

    ]'able 1 The Statistical Ensemble

    fl: 6.0 5.7 5.7 N/: 0 2 2

    mdv,~ : oc .025 .01 Ns~mp: 32 35 49

    agonal of the hypercube and putting the phase appropriate for the r/' operator. Thus the opera- tor which created the rf was the spatial sum over a double timeslice of the hypercube operator, plus noise terms which vanish on average. We took two noise samples per configuration.

    For the two-loop contractions we used source ~/ with U(1) noise at every site on the lattice. Then solving (~+ m) = r/we estimate the propagator as G~ v = (mCx~v). We note that this estimator is orders of magnitude better than the alterna- tive G, v = {*Yv*), but it only available when the sites x and y are separated by an even dis- tance. For this reason we restricted our attention to the pseudoscalar operator, and neglected the axial vector (which is distance 3). For each color we used 16 noise samples on a doubled lattice, or an effective number of 96 noise vectors per mass.

    3. Resu l ts

    We computed staggered propagators on the 32 node T3D at OSC and on the 256 node T3D at Los Alamos's Advanced Computer Laboratory. We ran our fixed size problem on sets of pro- cessors ranging from 16 nodes to 128 nodes, and found the per node performance to be essentially independent of the partition size over this range. By hand coding the inner loops in CAM, the Cray T3D assembly language, we obtained a sustained performance of 45 Mflops per node, including I /O time.

    To create and destroy the ~' we used the stag- gered flavor singlet operator 0(75 I )Q. This is a distance 4 operator, which we made gauge invariant by putting in explicit links, and aver- aging over the 24 paths across the edges of the hypercube.

    To obtain the one-loop contractions of this op- erator (the denominator of fig. 1), we computed two types of propagators. One type used a noisy source which was nonzero on two timeslices t = 0 and t = 1. This source was a random phase r/x = e i(x) for each color and each site, such that in the average over noise samples, (q, rjvt ) = 5zv. The second type of propagator used a source ob- tained by transporting the noise r/, across the di-

    mdyn =; royal ~-~

    1 .........

    0 0 2 4 0


    Figure 2. Quenched R(T).

    Figure 2 shows the measured ratio R(T) in the quenched ensemble at a valence quark mass mq = .01. In in principle, the sickness of the quenched approximation should manifest itself in an unending linear trend in the data. By contrast, the dotted curve is the exponential one would ex- pect if there were N$ = 4 active dynamical fla- vors. Clearly the data rule out Nf = 4, and are consistent with the quenched form, but from the data alone one couldn't rule out a nonpathological behavior with Nf = 2 or fewer flavors. Extract- ing the slope from such curves for several quark masses, we obtain the values for m0 plotted in figure 3

  • 360 G. Kilcup et al./Nuclear Physics B (Proc. Suppl.) 47 (1996) 358-361



    I . . . . I . . . . I ' ' ' Quenched

    0.0 I . . . . I . . . . 0.00 0.02


    Figure 3. Quenched m~.


    0.04 0 .06



    '1 . . . . I . . . . i . . . . l ' ' m.. ,~=.O I ; m,, , , l=.O 1

    Am - . l~* .03 . . . . .......

    0 2 4 8 Ta

    Figure 4. Dynamical R(r).

    Extrapolating linearly to mq = 0 and rescaling to the most relevant case of N! = 3 degenerate flavors, we obtain-


    rn~(g/ = 3) = (1050 4- 170 MeV)2(2-~e-V )2, (4)

    which is compatible with the "experimental" number quoted above.

    In the presence of iV] = 2 dynamical fermions, we expect R(r) to asymptote to the constant 2. Figure 4 shows the result for mdyn = m,~az = .01. To fit it to the exponential form may be putting more weight on the data than it should bear, but its behavior does stand in clear contrast to the quenched data. If we ignore our theoretical prejudice and simply press ahead to repeat the linearized analysis as in the quenched case, we find the result


    m~(g] = 3) = (780 + 50 MeV)2(2-~e'V )~, (5)

    On the other hand, if we do the correct job and fit to the exponential form, we can extract mn, a at two points, mdyn : mvaZ = .01 and rndu~ = mvat = .025, finding .488 4-.030 and .676 4- .040 for the bare lattice numbers respec- tively. Extrapolating to mq = 0 and rescaling to N! = 3 we end up with


    m~(N] = 3) = (730 4- 250 MeV)2(~) 2 (6)

    4. Cool ing

    In a closely related study, we have "cooled" the dynamical (mq = .01) to count large instantons.

    Dividing the configurations into bins according to their cooled topological charge, we compute the qJ mass on two subensembles, one with n = 0,1 or 2 instantons, and one with n = 3,4 or 5 instan- tons. Like the authors of ref. [2] we find that m02 mass is larger (by a factor 2) in the subensem- ble with larger topological charge, confirming the idea that the rf gets its mass from instantons. This result can be understood by looking at the pseudoscalar expectation value in particular con- figurations. Figure 5 shows what happens to the qJ operator as one cools. The correlation evident in the plot persists as a function of cooling time, even out to the point where all that remains are smooth instantons. We conclude that even in the hot configuration, the large features in the rf are due to underlying instantons.



    ~v- lO

    I . . . . I . . . . I m I _ m mo

    [] e 6 0

    ~lm "


    m~ m m ~

    m e m mW [] hot mmgmmmmmmm o aoo]

    I . . . . I . . . . I . . . . I

    I0 20 30 T

    Figure 5. Effect of 50 cooling steps.

    This observation invites one to explore a method which could speed up this and other cal- culations. Any obser