CP-PACS: A massively parallel processor at the University of Tsukuba

Nakazawa, Koyomi; Nakamura, Hiroshi; Boku, Taisuke; Nakata, Ikuo; Yamashita, Yoshiyuki

doi:10.1016/s0167-8191(99)00078-2

Cited by 15 publications

(6 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This is an endeavor demanding considerable computing resources, which we hope to meet with the use of the CP-PACS parallel computer with a peak speed of 614 GFLOPS developed at the University of Tsukuba [19,20]. We explore, as a first step toward a realistic simulation of QCD, the case of dynamical up and down quarks, which are assumed degenerate, treating the strange quark in the quenched approximation.…”

Section: Introductionmentioning

confidence: 99%

Erratum: Light hadron spectroscopy with two flavors of dynamical quarks on the lattice [Phys. Rev. D65, 054505 (2002)]

Khan¹,

Aoki²,

Boyd³

et al. 2003

Phys. Rev. D

View full text Add to dashboard Cite

We present results of a numerical calculation of lattice QCD with two degenerate flavors of dynamical quarks, identified with up and down quarks, and with a strange quark treated in the quenched approximation. The lattice action and simulation parameters are chosen with a view to carrying out an extrapolation to the continuum limit as well as chiral extrapolations in dynamical up and down quark masses. Gauge configurations are generated with a renormalization-group improved gauge action and a mean field improved clover quark action at three values of β = 6/g 2 , corresponding to lattice spacings of a ≈ 0.22, 0.16 and 0.11 fm, and four sea quark masses corresponding to mPS/mV ≈ 0.8, 0.75, 0.7 and 0.6. The sizes of lattice are chosen to be 12 3 × 24, 16 3 × 32 and 243 × 48 so that the physical spatial size is kept constant at La ≈ 2.5 fm. Hadron masses, light quark masses and meson decay constants are measured at five valence quark masses corresponding to mPS/mV ≈ 0.8, 0.75, 0.7, 0.6 and 0.5. We also carry out complementary quenched simulations with the same improved actions. The quenched spectrum from this analysis agrees well in the continuum limit with the one of our earlier work using the standard action, quantitatively confirming the systematic deviation of the quenched spectrum from experiment. We find the twoflavor full QCD meson masses in the continuum limit to be much closer to experimental meson masses than those from quenched QCD. When using the K meson mass to fix the strange quark mass, the difference between quenched QCD and experiment of 2.6 −0.031 in two-flavor full QCD, approaching the experimental value J ≈ 0.48. We take these results as manifestations of sea quark effects in two-flavor full QCD. For baryon masses full QCD values for strange baryons such as Ξ and Ω are in agreement with experiment, while they differ increasingly with decreasing strange quark content, resulting in a nucleon mass higher than experiment by 10% and a ∆ mass by 13%. The pattern suggests finite size effects as a possible origin for this deviation. For light quark masses in the continuum limit we obtain m −11 MeV (φ-input), which are reduced by about 25% compared to the values in quenched QCD. We also present results for decay constants where large scaling violations obstruct a continuum extrapolation. Need for a non-perturbative estimate of renormalization factors is discussed.

show abstract

Section: Introductionmentioning

confidence: 99%

Erratum: Light hadron spectroscopy with two flavors of dynamical quarks on the lattice [Phys. Rev. D65, 054505 (2002)]

Khan¹,

Aoki²,

Boyd³

et al. 2003

Phys. Rev. D

View full text Add to dashboard Cite

show abstract

“…The original prototype HMCS consists of two HPC platforms; the CP-PACS [1] with 2048 processing units (PUs) for general-purpose computation, and a GRAPE-6 [7] cluster with 256 GRAPE-6 chips for gravity calculation. CP-PACS is equipped with 128 additional I/O units (IOUs), each connected to individual RAID-5 disk arrays to provide very high file I/O throughput.…”

Section: Hmcsmentioning

confidence: 99%

HMCS-G: grid-enabled hybrid computing system for computational astrophysics

Boku

Sato

Onuma

et al. 2003

CCGrid 2003. 3rd IEEE/ACM International Symposium on Cluster Computing and the Grid, 2003. Proceedings.

View full text Add to dashboard Cite

The authors have developed a hybrid computing system called HMCS-G, a Grid-enabled Heterogeneous Multi-Computer System, that provides a multiple cluster environment centered around a dedicated machine for gravity calculation. The purpose of HMCS-G is to provide an ideal computational environment for astrophysical study involving multiple physical phenomena. The worker cluster may comprise general-purpose PCs to perform tasks such as hydrodynamics computations, while the specialpurpose machine, in this case a GRAPE-6 cluster, performs gravity calculations for all pairs of particles in the system. These systems are connected by OmniRPC, a grid-enabled RPC system that supports Globus and ssh for authentication. HMCS-G effectively provides worldwide access to a GRAPE-6 cluster, thereby securing several TFLOPS performance for intensive computations such as gravity calculation. All participating PC-clusters share this resource in a time-based manner using grid technology. The actual turn-around response time was measured for a system implemented over a number of institutions, and it was confirmed that HMCS-G provides acceptable real-world application performance. Precise simulations of galaxy formation are currently being performed on clusters in several institutes, involving smoothed particle hydrodynamics and radiative transfer in the context of complete gravity calculation as the first real application of HMCS-G.of the 3rd IEEE/ACM International Symposium on Cluster Computing and the Grid (CCGRID'03) 0--7695-1919-9/03 $17.00 © 2003 IEEE Proceedings of the 3rd IEEE/ACM International Symposium on Cluster Computing and the Grid (CCGRID'03) 0--7695-1919-9/03 $17.00 © 2003 IEEE Proceedings of the 3rd IEEE/ACM International Symposium on Cluster Computing and the Grid (CCGRID'03) 0--7695-1919-9/03 $17.00

show abstract

“…The prototype of HMCS (for short, we call it simply HMCS hereafter) consists of two HPC platforms, the CP-PACS [1] with 2048 processing units (PUs) for general purpose computation and the GRAPE-6 [4] cluster with 256 GRAPE-6 chips for gravity calculation. CP-PACS is equipped with 128 additional I/O processors (IOUs) as well as 2048 PUs, and each IOU is connected to individual RAID-5 disk array to provide very high throughput on file I/O.…”

Section: Concept and Design Of Hmcs Prototypementioning

confidence: 99%

“…CP-PACS [1,2] is an MPP for large scale scientific calculations equipped with 2048 PUs and 128 IOUs which started the operation in 1996 under the collaborative work of University of Tsukuba and Hitachi. The processor on each PU or IOU is based on HP PA-RISC 1.1 Architecture with special feature named PVP-SW (Pseudo Vector Processing with Sliding Window) [3] which enables to carry out very high performance for vector processing as well as pure vector processors.…”

Section: Cp-pacsmentioning

confidence: 99%

“…The basic concept of PIO is to distribute a large amount of data transfer between an MPP and its surrounding computational resources like front end workstations in both sense of I/O processors and network links. The basic function of PIO is similar to trunking technology ordinarily used for Ethernet switches, 1 It is given with the number of pipeline calculation elements and the circuit frequency. It corresponds to the FLOPS when the same floating point operations are performed on ordinary general purpose processors.…”

Section: Pio Networkmentioning

confidence: 99%

See 1 more Smart Citation

Heterogeneous multi-computer system: a new paradigm of parallel processing

Boku

Makino

Susa

et al.

Proceedings. International Conference on Parallel Computing in Electrical Engineering

View full text Add to dashboard Cite

HMCS (Heterogeneous Multi-Computer System) is anew parallel processing platform combining massively parallel processors for continuum simulation and particle simulation to realize multi-scale computational physics simulations. We constructed a prototype system of HMCS with a general purpose scientific parallel processor CP-PACS and a gravity calculation parallel processor GRAPE-6 connecting them via commodity-base parallel network.On the prototype of HMCS, a microscopic gravity calculation on GRAPE-6 and a macroscopic radiation hydrodynamics calculation on CP-PACS are performed simultaneously to realize detailed simulation on computational astrophysics. In this paper, we report the overall concept, design and implementation of HMCS as well as the result of a novel computational simulation for galaxy formation.

show abstract

CP-PACS: A massively parallel processor at the University of Tsukuba

Cited by 15 publications

References 23 publications

Erratum: Light hadron spectroscopy with two flavors of dynamical quarks on the lattice [Phys. Rev. D65, 054505 (2002)]

Erratum: Light hadron spectroscopy with two flavors of dynamical quarks on the lattice [Phys. Rev. D65, 054505 (2002)]

HMCS-G: grid-enabled hybrid computing system for computational astrophysics

Heterogeneous multi-computer system: a new paradigm of parallel processing

Contact Info

Product

Resources

About