2+1 flavor full QCD configurations by CP-PACS+JLQCD

The 2+1 flavor full QCD configurations listed below are generated by the CP-PACS and JLQCD Collaborations on various supercomputers including

the CP-PACS parallel computer (2048 nodes, peak 614 GFLOPS) and its frontend computer Hitachi SR-8000/G1 (12 nodes, peak 173 GFLOPS)
at the Center for Computational Sciences, University of Tsukuba,
Hitachi SR-8000/F1 (100 nodes, peak 1.2 TFLOPS)
at High Energy Accelerator Research Organization (KEK)
Fujitsu VPP-5000 (80 nodes, peak 768 GFLOPS)
at Academic Computing and Communucations Center, University of Tsukuba
the Earth Simulator (640 nodes, peak 40 TFLOPS)
at Japan Agency for Marine-Earth Science and Technology (JAMSTEC)

Simulation parameters
Configuration ensembles
Naming of Ensembles and Configurations (for ILDG users)
Location of Configurations (for JLDG users)

Simulation parameters

These gauge configurations are generated with a renormalization-group improved gauge action and a non-perturbatively O(a) improved clover quark action at three values of beta = 6/g², corresponding to lattice spacings of a \approx 0.12, 0.10 and 0.07 fm (a² \approx 0.015, 0.010 and 0.05 fm²). Note that the discretization errors are of the form O(a²) because the action is non-perturbatively O(a) improved. The scale a is fixed by experimental values of M_pi, M_rho, M_K (or M_phi) using a general quadratic chiral extrapolation in terms of VWI quark masses.

The sizes of lattice are chosen to be 16³x32, 20³x40 and 28³x56 so that the physical spatial size is kept constant at La \approx 2.0 fm. On each lattice, simulations are carried out at five values of light sea qaurk masses corresponding to m_PS(LL)/m_V(LL) \approx 0.78, 0.75, 0.7, 0.65 and 0.6, and two values of strange sea quark masses sandwitching the physical point m_PS(SS)/m_V(SS) \approx 0.68, where L and S stand for light and strange sea quarks.

Simulation parameters and values of a, La and m_PS/m_V are summarized in the table below.

beta	lattice size	c_SW	a [fm] K-input phi-input	La [fm] K-input phi-input		K_L
beta	lattice size	c_SW	a [fm] K-input phi-input	La [fm] K-input phi-input	K_S	m_PS(LL)/m_V(LL) m_PS(SS)/m_V(SS) #traj (thermalized) \delta\tau N_poly
1.83	16³x32	1.7610	0.1209(16) 0.1219(19)	1.934(26) 1.950(30)		0.13655	0.13710	0.13760	0.13800	0.13825
					0.13710	0.7772(13) 0.7522(15) 7000 1/80 80	0.7524(21) 0.7524(21) 7000 1/85 80	0.7076(18) 0.7414(17) 7000 1/100 100	0.6629(22) 0.7365(16) 8000 1/120 110	0.6213(24) 0.7343(15) 8000 1/140 120
					0.13760	0.7769(14) 0.7235(19) 7000 1/90 110	0.7448(14) 0.7128(16) 8600 1/100 110	0.7033(18) 0.7033(18) 8000 1/110 120	0.6525(23) 0.6941(20) 8100 1/120 130	0.6083(32) 0.6884(21) 8100 1/150 150
1.90	20³x40	1.7150	0.0982(19) 0.0983(19)	1.964(38) 1.966(38)		0.13580	0.13610	0.13640	0.13680	0.13700
					0.13580	0.7673(15) 0.7673(15) 5000 1/125 110	0.7435(18) 0.7647(17) 6000 1/125 110	0.7204(19) 0.7687(15) 7600 1/140 110	0.6701(27) 0.7673(17) 8000 1/160 110	0.6394(21) 0.7695(15) 7900 1/180 110
					0.13640	0.7667(16) 0.7211(21) 5200 1/125 140	0.7444(15) 0.7182(17) 8000 1/125 140	0.7145(16) 0.7145(16) 9000 1/140 140	0.6630(21) 0.7127(17) 9200 1/160 140	0.6243(28) 0.7102(20) 7900 1/180 140
2.05	28³x56	1.6280	0.0685(26) 0.0687(25)	1.918(73) 1.924(70)		0.13470	0.13510	0.13540	0.13550	0.13560
					0.13510	0.7757(26) 0.7273(29) 6000 1/175 200	0.7316(24) 0.7316(24) 6000 1/195 200	0.6874(30) 0.7395(23) 6000 1/225 200	0.6611(34) 0.7361(25) 6500 1/235 200	0.6337(38) 0.7377(28) 6500 1/250 200
					0.13540	0.7790(23) 0.6821(32) 6000 1/175 250	0.7341(29) 0.6820(39) 6000 1/195 250	0.6899(34) 0.6899(34) 6000 1/225 250	0.6679(45) 0.6899(43) 6500 1/235 250	0.6361(47) 0.6852(46) 6500 1/250 250

Configurations are generated with the Hybrid Monte Carlo (HMC) algorithm for the light sea quarks and the Polynomial HMC (PHMC) algorithm for the strange sea quark. See

Polynomial Hybrid Monte Carlo algorithm for lattice QCD with odd number of flavors
JLQCD Collaboration: S. Aoki, R. Burkhalter, M. Fukugita, S. Hashimoto, K-I. Ishikawa, N. Ishizuka, Y. Iwasaki, K. Kanaya, T. Kaneko, Y. Kuramashi, M. Okawa, T. Onogi, S. Tominaga, N. Tsutsui, A. Ukawa, N. Yamada, T. Yoshi\'e
Phys. Rev. D 65 (2002) ref.094507, pp.1-22 (hep-lat/0112051)

We adopt an even/odd preconditioned BiCGStab algorithm to invert quark matrices during the HMC+PHMC update.

For the clover coefficient c_SW we adopt a nonperturbatively improved value obtained by a Schroedinger functional method. See

Nonperturbative O(a) improvement of the Wilson quark action with the RG-improved gauge action using the Schrodinger functional method
CP-PACS, JLQCD Collaborations: S. Aoki, M. Fukugita, S. Hashimoto, K-I. Ishikawa, N. Ishizuka, Y. Iwasaki, K. Kanaya, T. Kaneko, Y. Kuramashi, M. Okawa, S. Takeda, Y. Taniguchi, N. Tsutsui, A. Ukawa, N. Yamada, T. Yoshie
Phys. Rev. D 73 (2006) ref.034501, pp.1-21 (hep-lat/0508031)

for details. The leading scaling violation with our action is O(a²).

Light hadron measurements are done every 10 unit-length HMC+PHMC trajectories. Errors are estimated by a jack-knife method with the bin-size of 100 trajectories. For more details, see

2+1 flavor light hadron spectrum and quark masses with the O(a) improved Wilson-clover quark formalism
CP-PACS/JLQCD Collaborations: T. Ishikawa, S. Aoki, M. Fukugita, S. Hashimoto, K.-I. Ishikawa, N. Ishizuka, Y. Iwasaki, K. Kanaya, T. Kaneko, Y. Kuramashi, M. Okawa, Y. Taniguchi, N. Tsutsui, A. Ukawa, N. Yamada, T. Yoshie
PoS (LAT2006) 181, pp.1-7 (hep-lat/0610063 )

Light quark masses from unquenched lattice QCD
CP-PACS/JLQCD Collaborations: T. Ishikawa, S. Aoki, M. Fukugita, S. Hashimoto, K-I. Ishikawa, N. Ishizuka, Y. Iwasaki, K. Kanaya, T. Kaneko, Y. Kuramashi, M. Okawa, Y. Taniguchi, N. Tsutsui, A. Ukawa, N. Yamada, T. Yoshie
Phys. Rev. D. 78 (2008) ref.011502(R), pp.1-5 (arXiv:0704.1937 [hep-lat])

Configuration ensembles

The table below gives a list of ensembles of configurations. A consecutive set of HMC+PHMC trajectories for a given value of beta, c_SW, K_L, K_S on a particular lattice size is called an ensemble. A series name (1, 2, 3, etc) is given to distinguish multiple HMC+PHMC runs.

ensemble #	beta	K_L	K_S	c_SW	lattice size	series	traj. range	missing traj	#config	total #config	size/config [MB]	total [GB]
1	1.83	0.13655	0.13710	1.7610	16³x32	1 2 3	000410-004400 000510-001500 000410-002400	-- -- --	400 100 200	700	72	49
2	1.83	0.13710	0.13710	1.7610	16³x32	1 2 3	000410-001800 000410-004600 000510-001900	-- -- --	140 420 140	700	72	49
3	1.83	0.13760	0.13710	1.7610	16³x32	1 2 3	000410-005400 000510-001500 000410-001400	-- -- --	500 100 100	700	72	49
4	1.83	0.13800	0.13710	1.7610	16³x32	1 2	000410-005900 000510-003000	-- --	550 250	800	72	56
5	1.83	0.13825	0.13710	1.7610	16³x32	1 2	000410-006500 000510-002400	-- --	610 190	800	72	56
6	1.83	0.13655	0.13760	1.7610	16³x32	1 2	000410-002400 000510-005500	-- --	200 500	700	72	49
7	1.83	0.13710	0.13760	1.7610	16³x32	1 2 3	000410-001400 000410-002400 000710-006300	-- -- --	100 200 560	860	72	60
8	1.83	0.13760	0.13760	1.7610	16³x32	1 2 3	000410-001400 000410-001400 000510-006500	-- -- --	100 100 600	800	72	56
9	1.83	0.13800	0.13760	1.7610	16³x32	1 2 3	000410-001400 000410-002700 000510-005300	-- -- --	100 230 480	810	72	57
10	1.83	0.13825	0.13760	1.7610	16³x32	1 2	000410-007300 000510-001700	-- --	690 120	810	72	57
ensemble #	beta	K_L	K_S	c_SW	lattice size	series	traj. range	missing traj	#config	total #config	size/config [MB]	total [GB]
11	1.90	0.13580	0.13580	1.7150	20³x40	1 2	000510-003500 000600-002600	002900 --	299 201	500	176	86
12	1.90	0.13610	0.13580	1.7150	20³x40	1	000410-006400	--	600	600	176	103
13	1.90	0.13640	0.13580	1.7150	20³x40	1 2 3	000600-005850 000410-001570 000410-001570	-- -- --	526 117 117	760	176	131
14	1.90	0.13680	0.13580	1.7150	20³x40	1 2 3	000410-006700 000410-001350 000260-001000	-- --	630 95 75	800	176	138
15	1.90	0.13700	0.13580	1.7150	20³x40	1 2	000510-004500 000610-004500	-- --	400 390	790	176	136
16	1.90	0.13580	0.13640	1.7150	20³x40	1 2	000510-003200 000410-002900	-- --	270 250	520	176	89
17	1.90	0.13610	0.13640	1.7150	20³x40	1 2	000410-006400 000410-002400	-- --	600 200	800	176	138
18	1.90	0.13640	0.13640	1.7150	20³x40	1 2 3 4	000590-006980 000410-001000 000510-001300 000510-001700	-- -- -- --	640 60 80 120	900	176	155
19	1.90	0.13680	0.13640	1.7150	20³x40	1 2 3	000570-007930 000510-001460 000510-001370	-- -- --	737 96 87	920	176	158
20	1.90	0.13700	0.13640	1.7150	20³x40	1 2 3 4	000510-003210 000560-003840 000410-001400 000510-001400	-- -- -- --	271 329 100 90	790	176	136
ensemble #	beta	K_L	K_S	c_SW	lattice size	series	traj. range	missing traj	#config	total #config	size/config [MB]	total [GB]
21	2.05	0.13470	0.13510	1.6280	28³x56	1 2	000510-001300 000510-005700	-- --	80 520	600	675	396
22	2.05	0.13510	0.13510	1.6280	28³x56	1 2	000510-004900 000110-001700	-- --	440 160	600	675	396
23	2.05	0.13540	0.13510	1.6280	28³x56	1 2	000510-004800 000110-001800	-- --	430 170	600	675	396
24	2.05	0.13550	0.13510	1.6280	28³x56	1 2	000510-005000 000110-002100	-- --	450 200	650	675	428
25	2.05	0.13560	0.13510	1.6280	28³x56	1 2	000510-004900 000110-002200	-- --	440 210	650	675	428
26	2.05	0.13470	0.13540	1.6280	28³x56	1 2	000510-001600 000510-005400	-- --	110 490	600	675	396
27	2.05	0.13510	0.13540	1.6280	28³x56	1 2	000510-004900 000110-001700	-- --	440 160	600	675	396
28	2.05	0.13540	0.13540	1.6280	28³x56	1 2	000510-004800 000110-001800	-- --	430 170	600	675	396
29	2.05	0.13550	0.13540	1.6280	28³x56	1 2	000510-005000 000110-002100	-- --	450 200	650	675	428
30	2.05	0.13560	0.13540	1.6280	28³x56	1 2	000510-004900 000110-002200	-- --	440 210	650	675	428

Naming of Ensembles and Configurations (for ILDG users)

An ensemble is identified by markovChainURI named e.g. as

mc://JLDG/CP-PACS+JLQCD/RCNF2+1/RC28x56_B2050Kud013560Ks013540C1628

A configuration is labeled by dataLFN named e.g. as

lfn://JLDG/CP-PACS+JLQCD/RCNF2+1/RC28x56_B2050Kud013560Ks013540C1628-2-002200

Basename of the dataLFN is formmated as
RgClover spatial size x temporal size _ \beta\kappa_ud\kappa_s\c_sw - series - trajectory number

To get full list of markovChainURIs and LFNs, please use QCDml faceted navigation or follow this link.

Location of Configurations (for JLDG users)

Configurations are placed in the gfarm directory
/gfarm/public/ILDG/JLDG/CP-PACS+JLQCD/RCNF2+1/
and classified in terms of β and κ_ud, κ_s. For example, the directory
/gfarm/public/ILDG/JLDG/CP-PACS+JLQCD/RCNF2+1/RC16x32_B1830/Kud013655Ks013710
contains configurations at β=1.830 and κ_ud=0.13655 κ_s=0.13710. Filename is the same as basename of dataLFN. All ensemble and configuration XML files can be found in
/gfarm/public/ILDG/JLDG/CP-PACS+JLQCD/RCNF2+1/QCDml

Last update: 03 Sept. 2009, T.Yoshie