CP-PACS Collaboration

CP-PACS Collaboration
Name	CP-PACS Collaboration
Formation	1996
Dissolved	2006
Type	scientific collaboration
Location	Japan
Fields	Particle physics, Computational physics, Lattice gauge theory
Leader	Yoshinobu Kuramashi
Notable projects	PACS-9; PACS-CS

CP-PACS Collaboration

The CP-PACS Collaboration was a Japanese computational physics consortium established to perform large-scale lattice QCD simulations using specialized parallel computers, driven by researchers from University of Tsukuba, University of Tokyo, KEK, and other institutions; it collaborated with projects such as PACS project, influenced efforts at RIKEN, and interfaced with international groups like the MILC collaboration and UKQCD. The collaboration operated during an era marked by developments at Fermilab, CERN, Brookhaven National Laboratory, and engagement with supercomputing centers including Waterloo-area facilities and national centers in Japan.

History

The collaboration formed in 1996 amid rising interest in nonperturbative calculations in Quantum Chromodynamics and builds on precedents set by groups at Los Alamos National Laboratory, Brookhaven National Laboratory, and Argonne National Laboratory; its timeline intersects with milestones at KEK, the commissioning of the Earth Simulator, and advances from the PACS project. Early phases featured system design influenced by architectures from Cray Research, IBM, and research at Riken, while subsequent phases coordinated with funding agencies such as the Ministry of Education, Culture, Sports, Science and Technology (Japan) and the Japan Society for the Promotion of Science. During its operational lifetime the collaboration announced results alongside conferences like the International Symposium on Lattice Field Theory and the Annual Meeting of the Physical Society of Japan.

Objectives and Research Programs

CP-PACS aimed to compute hadron spectra, quark masses, and matrix elements in Quantum Chromodynamics using lattice discretizations, addressing problems relevant to experimental programs at KEK, CERN, SLAC, and J-PARC. Programs targeted light hadron spectroscopy, heavy-quark phenomenology related to experiments at Belle, BaBar, and LHCb, and investigations of the QCD phase diagram pertinent to work at RHIC and ALICE. The collaboration set milestones for precision benchmarks comparable to those from the HPQCD collaboration, RBC and UKQCD collaborations, and JLQCD efforts, and coordinated software development with initiatives like QDP++ and Chroma.

Computational Methods and Infrastructure

CP-PACS developed and deployed the CP-PACS supercomputer, an architecture inspired by SIMD and MIMD designs used at Cray Research and IBM, and drew on VLSI and node design expertise akin to projects at NEC and Hitachi. Algorithms emphasized iterative solvers such as conjugate gradient and multigrid, gauge actions including Wilson and clover improved schemes tied to formulations studied by Kenneth Wilson and Sheldon Glashow, and fermion discretizations like domain wall and staggered fermions connected to work by David Kaplan and Howard Georgi. Code optimization targeted cache hierarchies similar to those in Fujitsu systems and leveraged parallel I/O strategies used at National Institute of Informatics (Japan) and Center for Computational Sciences, University of Tsukuba.

Key Results and Publications

The collaboration produced high-precision determinations of the light hadron spectrum, quark masses, and strong coupling parameters, publishing results in leading journals alongside comparisons with data from Particle Data Group summaries and experiments at CERN and KEK B. Major publications addressed finite-volume effects, continuum extrapolations, and nonperturbative renormalization techniques referenced against methods developed at ALPHA Collaboration, JLQCD, and HPQCD. CP-PACS papers were presented at venues such as the Lattice Field Theory conference, the American Physical Society meetings, and the International Conference on High Energy Physics, influencing reviews by authors affiliated with Princeton University, University of California, Berkeley, and MIT.

Collaboration Members and Organization

Membership spanned faculty and researchers from University of Tsukuba, University of Tokyo, Osaka University, Nagoya University, Kyoto University, KEK, and research centers including RIKEN and the National Institute for Fusion Science; notable scientists included leaders linked to laboratories at KEK and universities with ties to Yukawa Institute for Theoretical Physics. The organizational structure featured principal investigators, technical staff for hardware design reminiscent of teams at Hitachi and NEC, and graduate students collaborating with international partners such as Columbia University, Cornell University, and University of Edinburgh. Funding and oversight involved agencies like the Ministry of Education, Culture, Sports, Science and Technology (Japan), coordination with consortia comparable to JST and engagement with computational facilities at Tsukuba.

Impact and Legacy

CP-PACS advanced lattice QCD precision, influenced successor machines such as PACS-CS, and informed algorithmic developments adopted by the MILC collaboration, HPQCD collaboration, and contemporary efforts at RIKEN and KEK. Its hardware and software contributions impacted supercomputing practices at University of Tsukuba and inspired collaborations with groups at CERN and Brookhaven National Laboratory; alumni from CP-PACS moved to positions at University of Tokyo, Osaka University, Princeton University, and national laboratories influencing projects at J-PARC and LHCb. The collaboration's legacy persists in citations across reviews by the Particle Data Group, methodological references in lattice toolkits like Chroma, and in training a generation of computational physicists active at institutions such as RBC and UKQCD, HPQCD, and JLQCD.

Category:Lattice gauge theory Category:Computational physics collaborations Category:Research institutes in Japan