CAMPA, presently a consortium of three existing efforts, has the vision of coordinating and integrating the development, maintenance, distribution and support of state-of-the-art accelerator computer codes. Under the aegis of CAMPA, we will make progress in accelerator science through advanced computation, in support of the Department of Energy’s mission. Advancement of modeling and simulation of accelerators will help all those who design and build them, and in turn the tens of thousands of researchers throughout the physical and life sciences who make use of the beams that accelerators provide, and ultimately all of society.
- Push the frontier of accelerator science through advanced simulation and modeling, and, in turn, the frontier of computational methods in accelerator science through algorithmic advances.
- Provide the scientific community with a comprehensive and integrated toolset of state-of-the-art simulation codes for multi-scale, multi-physics accelerator modeling, in support of the mission of the Office of Science within the DOE.
- Develop and maintain the codes on DOE’s supercomputing facilities; distribute and support codes for installation on smaller scale clusters, desktops or laptops.
- Support users.
- Apply the codes as education tools to train students and young researchers on the science and the modeling of accelerators.
- Promote collaboration and re-use of accelerator simulation codes and data through common interfaces, data standards, and integrated visualization and analysis capabilities.
Existing codes will continue to be available in “toolkit” fashion; CAMPA will add value by making it easier to combine their inputs and results. Data sharing, development of common user interfaces, modularization, and consolidation are among the focus areas for CAMPA.
The Need for What CAMPA Can Do
The importance of these activities is underscored by the findings and recommendations of blue-ribbon panels, such as the Particle Physics Project Prioritization Panel (“P5”), June 2014 ; the Topical Panel Meeting on Computing and Simulations in High Energy Physics; and DOE’s Community Summer Study (Snowmass 2013), as well as inputs given to panels such as
- “White Paper on DOE-HEP Accelerator Modeling Science Activities,” J.-L. Vay et al., input to Snowmass 2013.
- Input to the P5 panel on DOE-HEP Computer Modeling of Accelerators,”J.-L. Vay et al. (Dec. 20, 2013).
- Input to the HEPAP Accelerator R&D subpanel on DOE-HEP Accelerator Modeling Science Activities,” J.-L. Vay et al. (Sept. 12, 2014).
We detail below the relevance of CAMPA with regard to the related P5 recommendations.
Recommendation 29: Strengthen the global cooperation among laboratories and universities to address computing and scientific software needs, and provide efficient training in next-generation hardware and data-science software relevant to particle physics. Investigate models for the development and maintenance of major software within and across research areas, including long-term data and software preservation.
- For DOE-HEP accelerator activities that rely on complex computer modeling, CAMPA will provide integrated suites of software that will speed up design and innovation. Unlike existing software that are often developed by individuals inside projects, the CAMPA software suite will be developed and supported by a specialized and experienced team. This will consolidate scattered efforts to enable routine virtual prototyping of accelerator components on much larger scales than would be possible otherwise, leading to high payoffs that will maximize DOE-HEP return on investment.
- “To address the science Drivers, increasingly higher demands are being placed on the performance in all three areas (accelerators, instrumentation and computing), at reduced cost. This necessitates the ongoing pursuit of innovation.”
CAMPA’s code development teams have an outstanding track record in innovative algorithms and use of high-performance computing. As a result, CAMPA’s codes contain a remarkably large number of novel algorithms (including the “Lorentz boosted frame” approach) that were introduced by the developers and adopted elsewhere (see the “Algorithms” tab on this site).
The culture of innovation of the CAMPA team will be a key asset for addressing the multiple challenges that will arise with the consolidation of code suites, the addition of new physics, and the porting of codes to emerging platforms. CAMPA’s capabilities and focus are all the more timely as computer architectures are transitioning to new technologies that require the adaptation of existing algorithms and codes and the development of new ones.
- “the use of high-performance computing, combined with new algorithms, is advancing full 3-D simulations at realistic beam intensities of nearly all types of accelerators. This will enable “virtual prototyping” of accelerator components on a larger scale than is currently possible.”
CAMPA’s consolidation of modeling efforts and improved user support will enable routine virtual prototyping and modeling of virtual accelerators on a scale that is currently unavailable. Advances in accelerator computing advance both of these enablers: computing and accelerators. As advances in algorithms and codes will speed up advances in accelerator research, funding accelerator computing activities will have compounding effects.
CAMPA In Context
The DOE’s Advanced Scientific Computing Research (ASCR) program is a major source of support for computer-science and networking research, as well as of computing facilities, including LBNL’s National Energy Research Supercomputing Center.
The High Energy Physics Forum for Computational Excellence reaches out across the HEP community. Its three principal activities are incubating R&D projects that result from certain DOE HEP calls; making specific computational expertise that exists within the HEP community more widely accessible and broadly useful; and promoting a collaborative culture across HEP domains. CAMPA is their topical POC for accelerator modeling activities, and HEP-FCE helps us disseminate the outcome of our activities.
The NERSC Exascale Science Applications Program (NESAP) brings together code teams, library and tools developers, and hardware vendors Cray and Intel to prepare for the manycore architecture of Cori, which is being commissioned as NERSC’s flagship system. The BLAST codes Warp and Impact are part of the top-tiers projects included in the NESAP partnership with NERSC, Cray and Intel (NESAP projects). The Berkeley Lab Accelerator Simulation Toolkit (BLAST) codes Warp and Impact are part of the 20 top-tier projects included in NESAP.
Through CAMPA, the knowledge accumulated in porting the BLAST codes Warp and Impact to Cori is being shared with the ACE3P and Synergia team, who have early access to the Cori machine. The knowledge will then be disseminated to the broader community through HEPFCE.
Benefits Beyond Accelerators
The establishment of CAMPA will provide an outreach platform to introduce or enhance the use of high-performance computing (HPC) in electromagnetic and beam physics simulation to potential customers beyond the DOE accelerator complex. The tools that are developed for accelerator science are applicable to environment, industry, medicine, material science and other areas of national interests. The efforts of code integration, novel functionality development, and easy access for users endorsed by CAMPA will benefit application areas as diverse as plasma physics applications in fusion, astrophysics, spacecraft charging and propulsion, and applications of electrical discharges, as well as electromagnetic modeling for power sources, communication systems, antennas, radar, etc.
Who We Are
The leaders of the consortium proposal — Jean-Luc Vay and Ji Qiang of LBNL; Cho-Kuen Ng and Zenghai Li of SLAC; and James F. Amundson and Eric G. Stern (Fermilab) — all have extensive experience not only with the phenomena they are modeling, but also in collaborating with the computer scientists and working with the HPC facilities of their respective laboratories.