LHC@home

From BOINC Projects
Revision as of 22:26, 12 June 2026 by Al Piskun (talk | contribs) (update)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search



LHC@home
LHC@home SixTrack screensaver
Project
StatusActive
CategoryParticle physics
ComputeCPU & GPU
RequiresVirtualBox (for most applications), Docker (optional for Theory tasks)
Development
DeveloperCERN
AuthorBen Segal and François Grey (founders); Frank Schmidt (SixTrack)
SponsorCERN
MaintainerLHC@home team
Initial releaseSeptember 1, 2004  (22 years ago)
Repositoryhttps://github.com/cern-it
Software
Written inC, C++, Python, Fortran
Operating systemWindows, Linux, macOS
SizeVaries by application
BOINC statistics
Stats as ofMay 21, 2026  (0 years ago)
Performance52 TFLOPS
Active users1,260
Total users178,244
Active hosts3,633
Total hosts577,548
Metadata
Websitehttps://lhcathome.cern.ch/lhcathome/
LicenseMixed free software licenses

LHC@home is a volunteer computing project operated by CERN using the BOINC platform. The project invites volunteers worldwide to donate spare CPU and GPU cycles on their personal computers to help physicists run computationally intensive simulations related to the Large Hadron Collider (LHC) and its experiments.[1] As of May 2026, LHC@home counts more than 178,000 registered users and 577,000 participating hosts, collectively delivering around 52 teraFLOPS of computing power.[2]

Wikipedia page

LHC@home

History

Origins

The Large Hadron Collider tunnel at CERN, near Geneva, Switzerland

LHC@home was conceived in 2004 by Ben Segal and François Grey, both then members of CERN's IT Department, as an outreach activity for CERN's 50th anniversary celebrations.[3] Grey had been an early volunteer for SETI@home and recognised that a similar model could engage the public in the computational challenges facing the LHC. He and Segal contacted David Anderson at the University of California, Berkeley, who was at that time developing BOINC as an open platform to generalise the SETI@home approach.[4]

The application chosen for the launch was SixTrack, a particle tracking code that had already been ported to Windows by CERN's Accelerators and Beams group and was in active use for beam dynamics studies. Since over 90% of BOINC volunteers at the time ran Windows machines, this was an ideal fit.[4]

Launch and early growth

LHC@home opened on 1 September 2004 as a beta test. Within 24 hours more than 1,000 participants had signed up, and over 7,000 had joined by the end of the first week, overwhelming the project's initial server capacity.[4] What had been planned as a three-month anniversary event proved so popular with CERN's accelerator physicists that the decision was taken to make it a permanent facility.[4]

Before LHC@home existed, SixTrack was run only on desktop computers at CERN using a proprietary tool called the Compact Physics Screen Saver (CPSS). The BOINC-powered platform opened the simulations to volunteers worldwide and delivered unprecedented computing capacity for accelerator beam tracking studies.[4]

Expansion through virtualisation

The project remained centred on SixTrack until 2011, when virtualisation technology was introduced to allow high-energy-physics (HEP) software from the LHC experiments to run on volunteer machines. Because each experiment's software stack is large, constantly updated, and built for Linux, direct porting was impractical.[5] The solution was CernVM, a compact Linux virtual machine developed at CERN beginning in 2008, which allowed experiment code to run inside a lightweight virtual environment on volunteer Windows, macOS, or Linux hosts. The Test4Theory application, launched in 2011, was the first BOINC project anywhere to use virtual machine technology.[6]

Over time, the platform evolved further. Applications began using Docker containers and the CVMFS (CernVM File System) to deliver application software on demand, reducing the size of the base virtual machine image to around 20 MB.[7] A substantial portion of contributed computing capacity also now comes from opportunistic backfill from data centres with spare capacity, supplementing individual volunteers.[8]

Goal

The primary goal of LHC@home is to provide CERN physicists with additional computing capacity for simulations that cannot be handled solely by the Worldwide LHC Computing Grid or CERN's own batch clusters.[9] These simulations are computationally intensive but involve small data sets, making them well suited to volunteer computing.

Project objectives include:

  • Studying the long-term dynamic aperture and stability of particle beams inside the LHC
  • Running Monte Carlo detector simulations for the ATLAS and CMS experiments
  • Simulating high-energy particle collisions to compare theoretical predictions with experimental measurements
  • Supporting research into dark matter, antimatter asymmetry, and fundamental particles
  • Preparing beam dynamics simulations for the High-Luminosity LHC (HL-LHC) upgrade

Applications

SixTrack

The ATLAS particle detector at CERN

SixTrack is the original LHC@home application and the longest-running BOINC application at CERN. It was developed by Frank Schmidt of CERN's Accelerators and Beams Department and performs multi-turn symplectic tracking of proton trajectories through the magnetic lattice of the LHC.[10]

The central quantity SixTrack evaluates is the dynamic aperture (DA), defined as the boundary in transverse phase space within which particle orbits remain stable over a large number of turns. Because no analytic theory can predict the long-term behaviour of particles in the strongly nonlinear magnetic fields of the LHC with sufficient accuracy, the DA must be determined numerically. Each volunteer's computer tracks hundreds of particles step by step through the thousands of LHC magnets for up to one million turns and checks whether each particle remains confined or escapes its orbit.[11]

In simplified terms, the transverse motion of a proton in a circular accelerator can be described by the Courant-Snyder (Twiss) parameterisation. The betatron oscillation amplitude in the horizontal plane is related to the action variable Jx by:

x=2βxJxcos⁡(ϕx)

where βx is the local beta function and ϕx is the betatron phase. SixTrack tracks particles in six-dimensional phase space (x,px,y,py,σ,δ), accounting for the longitudinal coordinate σ and relative momentum deviation δ=Δp/p0, using a symplectic thin-lens element-by-element map through each magnet.[7] The dynamic aperture is then reported as the minimum stable amplitude, expressed in units of the RMS beam size σbeam.

Each SixTrack work unit is submitted at least twice to two different volunteer hosts and results are cross-validated to eliminate hardware-induced numerical errors. Peaks of over 350,000 simultaneously running tasks on 24,000 hosts have been recorded during intensive simulation campaigns.[12]

Lyn Evans, head of the LHC project during its construction, stated that the results from SixTrack were genuinely making a difference, providing new insights into how the LHC would perform.[13] SixTrack results played an essential role in the design of stable beam conditions for the LHC, and the application continues to be used for simulations relevant to the High-Luminosity LHC upgrade.[4]

ATLAS@home

The ATLAS@home application allows volunteers to run detector simulations for the ATLAS experiment, one of the two general-purpose particle detectors at the LHC.[14]

ATLAS tasks use CernVM virtualisation and the CVMFS file system to deliver the ATLAS software environment to volunteer machines. Simulations include the creation and decay of hypothetical supersymmetric bosons and fermions as well as Standard Model processes required for background studies.

CMS@home

CMS@home supports Monte Carlo detector simulations for the Compact Muon Solenoid (CMS) experiment.[15]

CMS tasks process full detector simulation and event reconstruction workloads. Integration between CMS@home and the CMS global submission infrastructure was described in detail in a 2018 publication by Field et al., which demonstrated how volunteer cloud computing could be transparently merged with high-throughput computing workflows.[16]

Test4Theory (Theory)

The CMS detector

Test4Theory, now simply called the Theory application, was launched on 1 August 2011 and was the first BOINC project in the world to run applications inside a virtual machine.[17] Volunteers run theoretical Monte Carlo simulations of high-energy proton collisions using models based on the Standard Model of particle physics. Results feed into the MCPLOTS database, a publicly accessible repository that allows both experimental and theoretical physicists to compare Monte Carlo generator predictions against existing measurements from accelerator experiments.[18]

In March 2018 the Theory application passed the milestone of four trillion simulated collision events, and by September 2023 it had surpassed six trillion events.[19]

Beauty (LHCb)

The Beauty application studied the decay properties of beauty (bottom) quarks and CP violation, or matter-antimatter asymmetry, relevant to the LHCb experiment.[20] Volunteer submission for LHCb workloads was paused indefinitely in November 2018.[21]

Xtrack / Xsuite

Xtrack, part of the broader Xsuite modular beam dynamics framework, is a next-generation replacement for SixTrack currently in testing on the BOINC platform.[22] Unlike SixTrack, Xtrack can also run on GPUs, potentially giving LHC@home access to a much broader range of volunteer hardware. A wider variety of beam simulations can be submitted through the Xtrack framework, and it is expected to become the principal beam dynamics application for the HL-LHC era.[23]

Technology

LHC@home uses the BOINC middleware platform originally developed at the University of California, Berkeley to power SETI@home.[24] CERN's BOINC server software is maintained by Laurence Field and a dedicated team at CERN, who have integrated the volunteer computing service with other CERN batch and grid computing infrastructure, notably HTCondor, for unified job submission and accounting.[25]

Key infrastructure components include:

  • CernVM: a compact Linux virtual machine (~20 MB base image) developed at CERN, used as the execution environment for virtualised applications. Application-specific software is streamed on demand via CVMFS rather than bundled in the image.[7]
  • VirtualBox: the local hypervisor used to run CernVM on volunteer Windows, macOS, and Linux machines for the ATLAS, CMS, and Theory applications.
  • Docker containers: used alongside or in place of VirtualBox for some Theory workloads.
  • CVMFS (CernVM File System): a content delivery network that streams experiment software to volunteer machines without requiring large local installations.[7]

SixTrack is a native application and does not require virtualisation, making it accessible to volunteers with simpler hardware configurations. The BOINC client also supports task replication: each SixTrack work unit is run on at least two independent hosts and results are compared to guard against hardware-induced floating-point errors, an approach that has made LHC@home a valuable tool for investigating numerical reproducibility across diverse computing architectures.[7]

Many virtualised workloads require several gigabytes of disk space and memory. GPU support is being introduced through the Xtrack application.

Team and administration

LHC@home was founded at CERN by Ben Segal and François Grey. The original SixTrack application was authored by Frank Schmidt of CERN's Accelerators and Beams Physics Group. Massimo Giovannozzi, also of that group, has contributed to the interpretation of SixTrack results for understanding particle dynamics in the LHC.[4] Igor Zacharov of the Particle Accelerator Physics Laboratory (LPAP) at EPFL has provided BOINC integration support for SixTrack.[4]

Laurence Field leads the team at CERN responsible for maintaining the BOINC server software and integrating the volunteer computing service with CERN's wider computing infrastructure.[26] Other contributors to recent LHC@home papers include David Cameron, Nils Høimyr, Frederik Van der Veken, and Ben Segal.[27]

Scientific impact

Thanks to computing power contributed by volunteers, numerous accelerator beam physics studies have been carried out, yielding improved understanding of charged particle dynamics in the LHC and its planned upgrades.[7] In particular:

  • SixTrack results were essential during the design of the LHC to establish magnet quality tolerances and safe beam conditions, ensuring beams would remain stable rather than fly off course into the vacuum pipe walls.[4]
  • Simulation campaigns during 2008-2010 used real magnetic measurement data from LHC tunnel installations to build more realistic tracking models, leading to better understanding of beam-beam interactions at collision points.[28]
  • The Theory application's Monte Carlo event database at MCPLOTS serves as a shared resource for experimental and theoretical physicists working on LHC analyses as well as earlier accelerator experiments.[29]
  • Volunteer computing resources supplement the Worldwide LHC Computing Grid for simulation workloads, which were difficult to accommodate within CERN's fully loaded batch clusters.[12]

The project hosted the first Pan-Galactic BOINC Workshop at CERN in 2005, with two further workshops organised jointly with the University of Geneva in 2006 and 2007, fostering collaboration among BOINC developers, scientists, and volunteers.[30]

Project statistics

As of May 2026, LHC@home reports the following statistics:[31]

  • Over 178,000 registered users
  • More than 577,000 participating hosts
  • Approximately 1,260 active users on around 3,633 active hosts
  • Combined computing performance of approximately 52 TFLOPS

Community

LHC@home maintains active message boards on its BOINC portal where volunteers discuss workunit availability, virtualisation setup, troubleshooting, and scientific background.[32] Community discussions on Reddit and external BOINC forums frequently cover VirtualBox compatibility issues, Docker container behaviour, Linux configuration, and long-running Theory tasks.[33]

Scientific publications

The following peer-reviewed publications are directly associated with LHC@home and its applications. A full list of papers arising from BOINC projects is maintained at https://boinc.berkeley.edu/pubs.php.

  • Barranco, J. et al. "LHC@Home: a BOINC-based volunteer computing infrastructure for physics studies at CERN." Open Engineering 7(1): 379-393 (2017). DOI: 10.1515/eng-2017-0042.
  • Cameron, D., Field, L., Giannakis, N. and Høimyr, N. "Extending CERN computing to volunteers: LHC@home consolidation and outlook." EPJ Web of Conferences 214: 03016 (2019). DOI: 10.1051/epjconf/201921403016.
  • Cameron, D. et al. "All grown-up; 18 years of LHC@home." EPJ Web of Conferences 295: 04004 (2024). DOI: 10.1051/epjconf/202429504004.
  • Karneyeu, A., Mijovic, L., Prestel, S. and Skands, P.Z. "MCPLOTS: a particle physics resource based on volunteer computing." European Physical Journal C 74: 2714 (2014). DOI: 10.1140/epjc/s10052-014-2714-9.
  • Field, L., Spiga, D., Reid, I., Riahi, H. and Cristella, L. "CMS@home: Integrating the Volunteer Cloud and High-Throughput Computing." Computing and Software for Big Science 2 (2018). DOI: 10.1007/s41781-018-0006-z.
  • Dykstra, D., Bockelman, B., Blomer, J. and Field, L. "The Open High Throughput Computing Content Delivery Network." EPJ Web of Conferences 214: 04023 (2019). DOI: 10.1051/epjconf/201921404023.
  • Anderson, D.P. "BOINC: A platform for volunteer computing." Journal of Grid Computing 18(1): 99-122 (2020). DOI: 10.1007/s10723-019-09497-9.

See also

LHC@home SixTrack 4.67 Screensaver
LHC@home SixTrack screensaver

External links

References

  1. ↑ LHC@home information portal. CERN. Retrieved 2026-05-21.
  2. ↑ LHC@home server status. CERN. Retrieved 2026-05-21.
  3. ↑ History of LHC@home. CERN. Retrieved 2026-05-21.
  4. ↑ 4.0 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 History of LHC@home. CERN. Retrieved 2026-05-21.
  5. ↑ Barranco, J. et al..(2017).LHC@Home: a BOINC-based volunteer computing infrastructure for physics studies at CERN. Open Engineering. pp. 379–393. DOI: 10.1515/eng-2017-0042.
  6. ↑ Test4Theory. CERN. Retrieved 2026-05-21.
  7. ↑ 7.0 7.1 7.2 7.3 7.4 7.5 Barranco, J. et al..(2017).LHC@Home: a BOINC-based volunteer computing infrastructure for physics studies at CERN. Open Engineering. pp. 379–393. DOI: 10.1515/eng-2017-0042.
  8. ↑ Cameron, D. et al..(2024).All grown-up; 18 years of LHC@home. EPJ Web of Conferences. pp. 04004. DOI: 10.1051/epjconf/202429504004.
  9. ↑ Welcome to LHC@home. CERN. Retrieved 2026-05-21.
  10. ↑ SixTrack. CERN. Retrieved 2026-05-21.
  11. ↑ SixTrack and numerical simulations. CERN. Retrieved 2026-05-21.
  12. ↑ 12.0 12.1 LHC@Home: A Volunteer computing system for Massive Numerical Simulations of Beam Dynamics and High Energy Physics Events. Retrieved 2026-05-21.
  13. ↑ LHC@home. Wikipedia. Retrieved 2026-05-21.
  14. ↑ ATLAS@home. CERN. Retrieved 2026-05-21.
  15. ↑ CMS@Home. CERN. Retrieved 2026-05-21.
  16. ↑ Field, L., Spiga, D., Reid, I., Riahi, H. and Cristella, L..(2018).CMS@home: Integrating the Volunteer Cloud and High-Throughput Computing. Computing and Software for Big Science. DOI: 10.1007/s41781-018-0006-z.
  17. ↑ Test4Theory. CERN. Retrieved 2026-05-21.
  18. ↑ Karneyeu, A., Mijovic, L., Prestel, S. and Skands, P.Z..(2014).MCPLOTS: a particle physics resource based on volunteer computing. European Physical Journal C. DOI: 10.1140/epjc/s10052-014-2714-9.
  19. ↑ LHC@home news. CERN. Retrieved 2026-05-21.
  20. ↑ Beauty. CERN. Retrieved 2026-05-21.
  21. ↑ (2018-11-19).Pausing Submission of LHCb Applications. CERN. Retrieved 2026-05-21.
  22. ↑ Cameron, D. et al..(2024).All grown-up; 18 years of LHC@home. EPJ Web of Conferences. pp. 04004. DOI: 10.1051/epjconf/202429504004.
  23. ↑ All grown-up; 18 years of LHC@home (CHEP 2023 paper). Retrieved 2026-05-21.
  24. ↑ BOINC. University of California, Berkeley. Retrieved 2026-05-21.
  25. ↑ Cameron, D., Field, L., Giannakis, N. and Høimyr, N..(2019).Extending CERN computing to volunteers: LHC@home consolidation and outlook. EPJ Web of Conferences. pp. 03016. DOI: 10.1051/epjconf/201921403016.
  26. ↑ Cameron, D., Field, L., Giannakis, N. and Høimyr, N..(2019).Extending CERN computing to volunteers: LHC@home consolidation and outlook. EPJ Web of Conferences. pp. 03016. DOI: 10.1051/epjconf/201921403016.
  27. ↑ Cameron, D. et al..(2024).All grown-up; 18 years of LHC@home. EPJ Web of Conferences. pp. 04004. DOI: 10.1051/epjconf/202429504004.
  28. ↑ SixTrack. CERN. Retrieved 2026-05-21.
  29. ↑ Karneyeu, A. et al..(2014).MCPLOTS: a particle physics resource based on volunteer computing. European Physical Journal C. DOI: 10.1140/epjc/s10052-014-2714-9.
  30. ↑ All grown-up; 18 years of LHC@home (CHEP 2023 paper). Retrieved 2026-05-21.
  31. ↑ LHC@home server status. CERN. Retrieved 2026-05-21.
  32. ↑ LHC@home forums. CERN. Retrieved 2026-05-21.
  33. ↑ r/BOINC. Reddit. Retrieved 2026-05-21.
Retrieved from "https://boincsynergy.ca/wiki/index.php?title=LHC@home&oldid=1604"