1 / 26

The QCD transition temperature from simulations on BlueGene L supercomputer at LLNL

Rajan Gupta T-8, Los Alamos National Lab The HotQCD collaboration. The QCD transition temperature from simulations on BlueGene L supercomputer at LLNL. (All HotQCD data are preliminary). RHIC: What we don’t know a priori. The nature (order) of the transition between hadronic matter and QGP

cece
Download Presentation

The QCD transition temperature from simulations on BlueGene L supercomputer at LLNL

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. Rajan Gupta T-8, Los Alamos National Lab The HotQCD collaboration The QCD transition temperature from simulations on BlueGene L supercomputer at LLNL (All HotQCD data are preliminary)

  2. RHIC: What we don’t know a priori The nature (order) of the transition between hadronic matter and QGP The transition temperature Tc The equation of state (at < Tc- 5 Tc ) The transport properties of the QGP Behavior of bound states (onia) in the QGP Collective excitations The process of hadronization The process of thermalization

  3. QCD at finite temperature T. Battacharya (LANL) M. Cheng (Columbia) N. Christ (Columbia) C. DeTar (Utah) S. Gottlieb (Indiana) R. Gupta (LANL) U. Heller (APS) K. Huebner (BNL) C. Jung (BNL) F. Karsch (BNL/Bielefeld) E. Laermann (Bielefeld) L. Levkova (Utah) T. Luu (LLNL) R. Mawhinney (Columbia) P. Petreczky (BNL) D. Renfrew (Columbia) C. Schmidt (BNL) R. Soltz (LLNL) W. Soeldner (BNL) R. Sugar (UCSB) D. Toussaint (Arizona) P. Vranas (LLNL) HotQCD Collaboration:A unique US wide opportunity for simulating QCD at finite temperature using the BlueGene L Supercomputer at LLNL

  4. Timeline Developed collaboration with Ron Soltz and formulated project for Blue Gene L at LLNL (Nov-Dec 2005) Organized US wide meeting to explore doing RHIC phenomenology and Lattice QCD at LLNL (Feb 06) Developed white paper for calculation of Tc and EOS on the Blue Gene L (Feb – May 06) Got NNSA approval to run on Blue Gene L ( July 06) Ported Asqtad and p4fat codes to Blue Gene L (Sept 06) Started NT=8 lattice QCD simulations (Oct 06) Continual code Optimization ~10,000 trajectories simulated at 15 for AsqTad & p4

  5. Restrictive Environment Only T. Luu, R. Soltz and R. Gupta can access the machine or the data Minimal data can be brought into open Print results in ASCI or print figures Declassify Scan to convert to digital for distribution Blue Gene L at LLNL represents a 4-6 Teraflops sustained resource

  6. HotQCD Collaboration: Goals Nature of the transition Do deconfinement and S restoration “coincide”? Transition temperature Tc Equation of State (EOS) at (=0, 0) Spectral Functions Spatial and temporal correlators versus T Transport coefficients of the quark gluon plasma

  7. Order of the transitions and Tc • Look for discontinuities, singularities, peaks in thermodynamics quantities • Polyakov line and its susceptibility • Chiral condensate and its susceptibility • Quark number and its susceptibility • Finite size scaling analysis (1st versus 2nd) Previous Staggered simulations For 2+1 (ud, s) quark flavorsthe transition is a crossover

  8. Issues with previous calculations2005 Calculations on coarse lattices: NT=4,6 and/or small volumes. (a ~ 0.2- 0.13 fermi). Simulate NT=4,6,8, … and do a0 extrapolation Tc with different actions (different observables) differ by ~ 20 MeV Reduce uncertainty to ~ 5 MeVif unique transition Need simulations with 2+1 flavors with realistic light (up, down) quark masses holding strange quark mass fixed Mud / ms = 0.5, 0.2, 0.1  (physical ~ 0.04) Staggered fermions (detM1/4): Talks by Creutz/Kronfeld

  9. Reliable continuum limit results 2007 Control discretization errors (compare 2 different improved staggered actions) Asqtad Staggered (MILC collaboration) p4 staggered (RBC-Bielefeld collaboration) Explore Domain wall fermions New simulations at NT=8with 2+1 flavors (mlight 0.1 ms) Combine with ongoing/previous NT = 4,6 simulations to perform a0 extrapolation Precise determination of lattice scale

  10. Fodor etal: different observables give different Tc Important for RHIC

  11. P4 action N8 simulations: Status 7/21/07 Lattice Size 8323232 Quark mass ml/ms = 0.1 β a ml T [MeV] # trajectories 3.460 3.490 3.510 3.540 3.570 3.600 3.630 3.660 3.690 3.760 h3.525 h3.530 h3.535 h3.540 h3.545 h3.550 c3.525 c3.530 c3.535 c3.540 c3.545 c3.550 0.00313 0.00290 0.00259 0.00240 0.00212 0.00192 0.00170 0.00170 0.00150 0.00139 0.00240 0.00240 0.00240 0.00240 0.00240 0.00240 0.00240 0.00240 0.00240 0.00240 0.00240 0.00240 154 169 179 194 209 225 241 256 271 313 186 189 191 194 196 199 186 189 191 194 196 199 10000 10000 11280 11440 12460 11790 12070 11190 10760 10920 6510 5450 5140 5410 6280 5790 6120 5530 5750 6260 6250 6520

  12. Asqtad action data set as of 7/21/07 ═6/g2 t Acceptance s exp(-s/2)>103 Total # trajectories 6.4580 0.04170 0.66 -4.9 +/- 9.4 12965 6.5000 0.04170 0.72 3.7 7.6 12990 6.5500 0.04170 0.76 -4.5 5.8 12720 6.6000 0.04170 0.78 -1.5 5.9 12405 6.6625 0.04170 12365 6.6500 0.04540 0.76 2.8 6.8 12445 6.6675 0.04540 12660 6.7000 0.04540 0.81 -9.7 5.8 12320 6.7600 0.04540 0.83 2.0 4.2 12530 6.8000 0.04540 0.84 -3.4 4.3 12195 6.8500 0.04540 0.86 2.2 3.3 12405 6.9000 0.04540 0.86 3.4 3.2 12470 6.9500 0.04540 0.86 1.5 3.6 12625 7.0000 0.04540 0.87 -2.9 3.0 12595 7.0800 0.04540 0.86 3.6 3.0 12790 Lattice Size 8323232; Quark mass ml/ms = 0.1 0.79 8.1 5.2 0.75 3.9 6.1 About 70% of target 15000 trajectories (statistics) achieved

  13. P4: Line of constant physics (LCP) Use T=0 simulations Mss r0 = 1.58 (M r0 0.52  M 220 MeV) Quark mass ml/ms = 0.1 (real world ~0.04) Take a0 along this LCP varying just the gauge coupling

  14. Observables Polyakov loop Quark number susceptibility Chiral condensate Chiral susceptibility Deconfinement S Restoration

  15. P4: Setting the lattice scale See talk by Jan van der Heide r0=0.469(7) fm, r00.5 =1.104(5), r0/r1 =1.463(7)

  16. Renormalized Polyakov Loop Smooth behavior N = 4,6,8 don’t show large O(a2) effects (See talk by Kostya Petrov) In all figures the band corresponds to the range T=185-195 MeV

  17. Chiral Condensate: S restoration • Subtraction to eliminate additive renormalization • Sharp decrease in  reflects S restoration

  18. N=8: P4 consistent with AsqTAD

  19. S Restoration TDeconfinement

  20. Quark Number Susceptibility States carrying such quantum numbers(0 baryon number/strangeness)are heavy at low T and light at high T. Probe confinement Do not require renormalization at V Peak in fourth derivative!

  21. Quark Number SusceptibilityP4 data consistent with AsqTAD Determination of Tc from inflection point sensitive to • Range of data points included in the fit (8 points) • Fit function

  22. s,l and s/T2 • Smooth behaviour • N = 4,6,8 don’t show large O(a2) effects • Crossover at roughly the same T

  23. Chiral Susceptibility: N = 4, 6, 8 2 2     Susceptibility for light (l = u,d) and strange (s) quarks: - + Connected Disconnected

  24. Chiral Susceptibility: N = 8 Disconnected Connected

  25. Chiral Susceptibility : N = 8 Total Susceptibility for light (l = u,d) quarks: No renormalization: Bare results Need more work to quantify agreement and extract Tc(a=0)

  26. Future Complete N=8 simulations Finish analysis of all the variables Combine N=4,6,8 calculations Extract transition temperature at which bulk quantities show largest fluctuations Resolve Deconfinement versus S restoration

More Related