BOINC Projects

SZTAKI Desktop Grid

Revision as of 13:12, 30 June 2026 by Al Piskun (talk | contribs) (add logo)



SZTAKI Desktop Grid (also referred to as SZDG) was a Hungarian volunteer computing project that used the BOINC platform to crowdsource spare CPU cycles from volunteers' home computers for mathematics, physics, and digital humanities research. The project was developed and operated by the Laboratory of Parallel and Distributed Systems (LPDS) at the Institute for Computer Science and Control (MTA SZTAKI, now part of HUN-REN) of the Hungarian Academy of Sciences in Budapest.[1] Beyond hosting public BOINC applications, SZDG was itself a research vehicle: SZTAKI's developers extended the stock BOINC server and client software into a more flexible middleware platform, also called SZDG, that supported hierarchical organisation of desktop grids and an additional programming interface known as DC-API.[2] The project closed permanently in 2018.

SZTAKI Desktop Grid
Project
StatusCompleted
CategoryMathematics, Physics, Computer Science
ComputeCPU
RequiresNone
Development
DeveloperLaboratory of Parallel and Distributed Systems, MTA SZTAKI
AuthorPéter Kacsuk, Zoltán Balaton, Gábor Gombás, Attila Csaba Marosi
SponsorHungarian Academy of Sciences
Initial releaseMay 26, 2005  (21 years ago)
CompletedYes
DiscontinuedMay 31, 2018  (8 years ago)
Software
Written inC, C++
Operating systemWindows, Linux, macOS
BOINC statistics
Stats as ofDecember 17, 2017  (9 years ago)
Performance1157.819 GFlop/s
Total users1,001,774
Total hosts115,698
Metadata
Websitehttp://szdg.lpds.sztaki.hu/szdg/

History

SZDG was launched on 26 May 2005.[3] It grew out of earlier desktop grid research at SZTAKI's Laboratory of Parallel and Distributed Systems, which had been exploring grid computing and the P-GRADE portal, and was conceived from the outset both as a public volunteer computing project and as a testbed for new desktop grid middleware.[4]

A central research contribution of the project was the concept of the hierarchical desktop grid: in standard BOINC, every volunteer host connects directly to a single central server, whereas SZDG allowed an entire desktop grid, or even a compute cluster, to act as a single "client" of a higher-level desktop grid server, taking and redistributing batches of work among its own local resources.[5] This allowed academic institutions and companies to build internal, firewalled "Local Desktop Grids" that could, in turn, contribute idle capacity upward into larger desktop grids without exposing internal application code or data to the public Internet.[6] To make it easier for scientists to port existing, often sequential, programs onto the platform without learning the BOINC API directly, SZTAKI developed a higher-level programming interface called DC-API, which also gained support for Java applications.[7]

SZDG's middleware and DC-API were subsequently adopted by other BOINC-based projects, most notably SLinCA@Home, a materials-science project operated by the Institute for Metal Physics of the National Academy of Sciences of Ukraine, which ran on the SZTAKI Desktop Grid platform and DC-API alongside standard BOINC.[8] SZTAKI's desktop grid technology was also central to the European Union-funded EDGeS and EDGI projects, which built bridges between BOINC-based desktop grids and service-oriented grid infrastructures such as EGEE, allowing jobs to flow between the two kinds of systems.[9]

In its later years the project also operated a separate, smaller research and testing instance known as the "SZTAKI Desktop Grid Research Facility" (SZDGR), used by SZTAKI staff to test new BOINC applications with a small pool of volunteer hosts before wider release.[10]

Applications

Over its lifetime, SZDG hosted several scientific and digital humanities applications, most developed in-house at SZTAKI rather than by outside research groups.

BinSys

BinSys (also written BinSYS) searched for all generalized binary number systems up to dimension 11, where the "base" of the number system is a square matrix rather than an ordinary integer and the "digits" are vectors.[11] Formally, for an integer matrix M of dimension n, BinSys investigated when every integer vector in ℤn can be expressed uniquely as a finite digit expansion

v=∑i=0kMidi

where each digit di is drawn from a fixed finite digit set. Such generalized radix representations had been studied mathematically as canonical number systems, and BinSys used distributed computing to search the underlying parameter space (matrix dimension up to 11) for matrices and digit sets that yield a valid, exhaustive number system.[12] One scalar special case of this search, canonical number systems (CNS) defined by an integer polynomial p(x) rather than a full matrix, was investigated mathematically by Péter Burcsi and Attila Kovács, who developed an exhaustive search method for finding all expansive polynomials of a prescribed degree n and constant term c0, since expansivity is a necessary condition for a polynomial to define a valid number system; using this method, run on SZTAKI Desktop Grid, they extended earlier classifications of binary CNS polynomials to cover degrees 9, 10, and 11.[13] The application's source code and project description were maintained by the Eötvös Loránd University computer algebra group.[14] BinSys was additionally deployed on other desktop grids beyond SZDG itself through the EDGeS bridge infrastructure, including the Extremadura Desktop Grid in Spain.[15]

UC-Explorer

UC-Explorer ("Universality Classes Explorer") was a statistical physics application intended to help characterize the basic universality classes of non-equilibrium systems.[16] It was distributed for both Linux (i686) and Windows clients, with macOS support added later, and was tested through the smaller SZDGR research facility before wider rollout to the SZDG volunteer pool.[17] The application's code was later reused and adapted in the EDGeS project as the worker component of a fusion-physics simulation called ISDEP, illustrating how SZDG's infrastructure and existing applications were repurposed for unrelated scientific domains.[18]

UC-Explorer's computing power was put to direct scientific use by Géza Ódor and collaborators at the Research Institute for Technical Physics and Materials Science in Budapest, working with the Institute of Ion Beam Physics and Materials Research in Dresden. Using SZDG, they ran large-scale lattice gas simulations of nonequilibrium surface growth, mapping (2+1)-dimensional Kardar-Parisi-Zhang growth onto a driven lattice gas model of dimers and providing numerical evidence for the predicted scaling behaviour of the growth process.[19] A follow-up study extended the model with conserved, local exchange dynamics to simulate surface pattern formation, using SZDG's distributed CPU capacity to perform simulations at sizes and timescales that would otherwise have been impractical, and confirming that the Kardar-Parisi-Zhang scaling class remained stable under the addition of surface diffusion.[20] Attila Marosi, one of SZDG's own developers, separately described the practical experience of porting and running these nonequilibrium lattice gas simulations on SZTAKI Desktop Grid at the Third AlmereGrid Desktop Grid Experience workshop.[21]

KOPI Plagiarism Checker

The KOPI Plagiarism Checker, developed by SZTAKI's Distributed Systems Department (DSD), is described by SZTAKI as the best-known plagiarism detection service for the Hungarian language.[16] To detect translational plagiarism, in which a foreign-language source is translated and presented as original work, KOPI needed to maintain an up-to-date, searchable index of foreign-language reference text, including a monthly-refreshed dataset derived from Wikipedia dumps.[16] Because the English Wikipedia alone comprised nearly four million articles and around 30 gigabytes of text (excluding images and associated data) at the time, SZTAKI split the corpus into smaller chunks, converted them to plain text, segmented the text into sentences, and stemmed the words, distributing this processing load across SZDG's volunteer hosts.[16] KOPI is one of the clearer examples of SZDG being used for digital humanities and text-processing infrastructure rather than purely numerical science.

Closure

SZTAKI announced in early April 2018 that SZDG would close down officially on 31 May 2018, citing the project's age and the resources required to keep it running; the announcement noted that the server would be kept available as long as feasible afterward so that in-progress work units could finish, and that the project's home page would subsequently be replaced with a static page describing the project and its results.[22] The project's own retrospective static page states that the project ultimately closed on 21 June 2018.[23] SZDG's official website, hosted at szdg.lpds.sztaki.hu, remained reachable as a static archive afterward, while SZTAKI's institutional pages continued to describe SZDG's history and applications.[16]

Legacy

Although SZDG itself closed in 2018, the hierarchical desktop grid concept, the DC-API programming interface, and the EDGeS/EDGI bridge technologies it pioneered continued to influence later desktop grid and volunteer computing research, particularly work on interoperability between volunteer computing and traditional service grids.[24] SZTAKI's role in developing this middleware is also reflected in the BOINC project's own developer wiki, which cites SZDG as the originator of hierarchical desktop grid organization within the BOINC ecosystem.[25]

Scientific publications

The BOINC project maintains an official, continuously updated list of scientific publications arising from BOINC-based computing at boinc.berkeley.edu/pubs.php. That list applies a strict inclusion criterion: only publications whose underlying results were actually computed using the project's BOINC-distributed resources are listed. The five publications below are the ones credited to SZTAKI Desktop Grid there. A further set of publications about SZDG's own software architecture, middleware, and related bridge technologies, which describe the platform itself rather than reporting scientific results computed on it, is listed separately afterward.

Publications based on data computed on SZTAKI Desktop Grid

  1. Ódor, Géza, Bartosz Liedke and Karl-Heinz Heinig. "Surface pattern formation and scaling described by conserved lattice gases." Physical Review E, 2010. DOI: 10.1103/PhysRevE.81.051114
  2. Marosi, Attila. "Exploring University Classes in Nonequilibrium Systems on SZTAKI Desktop Grid." Third AlmereGrid Desktop Grid Experience workshop, 2010.
  3. Ódor, Géza, Bartosz Liedke and Karl-Heinz Heinig. "Mapping of (2+1)-dimensional Kardar-Parisi-Zhang growth onto a driven lattice gas model of dimers." Physical Review E, 2009. DOI: 10.1103/PhysRevE.79.021125
  4. Kacsuk, Peter, Jozsef Kovacs, Zoltan Farkas, Attila Csaba Marosi, Gabor Gombas and Zoltan Balaton. "SZTAKI Desktop Grid (SZDG): A Flexible and Scalable Desktop Grid System." Journal of Grid Computing, 2009. DOI: 10.1007/s10723-009-9139-y
  5. Burcsi, Péter and Attila Kovács. "Exhaustive search methods for CNS polynomials." Monatshefte für Mathematik, 2008. DOI: 10.1007/s00605-008-0005-y

Other project publications

These papers describe SZTAKI Desktop Grid's software architecture, middleware, and related bridge infrastructure, or provide mathematical background for its applications, rather than reporting results computed using SZDG itself.

  1. Balaton, Z., Gombás, G., Kacsuk, P., Kornafeld, Á., Kovács, J., Marosi, A. C., Vida, G., Podhorszki, N. and Kiss, T. "SZTAKI Desktop Grid: a Modular and Scalable Way of Building Large Computing Grids." 2007 IEEE International Parallel and Distributed Processing Symposium (IPDPS 2007), 2007, pp. 1–8. DOI: 10.1109/IPDPS.2007.370668
  2. Kacsuk, P., Marosi, A. C., Kovács, J., Balaton, Z., Gombás, G., Vida, G. and Kornafeld, Á. "SZTAKI Desktop Grid: a Hierarchical Desktop Grid System." Cracow '06 Grid Workshop, Kraków, Poland, 2006.
  3. Marosi, A. C., Gombás, G., Balaton, Z., Kacsuk, P. and Kiss, T. "SZTAKI Desktop Grid: Building a Scalable, Secure Platform for Desktop Grid Computing." In: Making Grids Work, Springer, Boston, MA, 2008, pp. 363–374. DOI: 10.1007/978-0-387-78448-9_29
  4. Marosi, A. C., Gombás, G. and Balaton, Z. "Enabling Java Applications for BOINC with DC-API." In: Kacsuk, P., Lovas, R. and Németh, Z. (eds.) Distributed and Parallel Systems: In Focus: Desktop Grid Computing, Springer, 2008, pp. 3–12.
  5. Marosi, A. C., Balaton, Z., Kacsuk, P. and Drótos, D. "SZTAKI Desktop Grid: Adapting Clusters for Desktop Grids." In: Davoli, F., Meyer, N., Pugliese, R. and Zappatore, S. (eds.) Remote Instrumentation and Virtual Laboratories: Service Architecture and Networking, Springer, 2010, pp. 133–144.
  6. Fedak, G., He, H., Lodygensky, O. et al. "EDGeS: A Bridge between Desktop Grids and Service Grids." 2008 Third ChinaGrid Annual Conference, 2008. DOI: 10.1109/ChinaGrid.2008.44
  7. Akiyama, S., Borbély, T., Brunotte, H., Pethő, A. and Thuswaldner, J. M. "Generalized Radix Representations and Dynamical Systems I." Acta Mathematica Hungarica, vol. 108, 2005, pp. 207–238. DOI: 10.1007/s10474-005-0221-z
  8. Kacsuk, P., Farkas, Z., Kovács, J., Lovas, R., Marosi, A. C. et al. "Desktop Grid in the Era of Cloud Computing." 2015.007/s00605-008-0005-y]

References

  1. ↑ SZTAKI Desktop Grid - Discontinued. HUN-REN SZTAKI. Retrieved 2026-06-29.
  2. ↑ (2007).SZTAKI Desktop Grid: a Modular and Scalable Way of Building Large Computing Grids. 2007 IEEE International Parallel and Distributed Processing Symposium. pp. 1–8. DOI: 10.1109/IPDPS.2007.370668.
  3. ↑ SZTAKI Desktop Grid. Wikidata. Retrieved 2026-06-29.
  4. ↑ (2009).SZTAKI Desktop Grid (SZDG): A Flexible and Scalable Desktop Grid System. Journal of Grid Computing. pp. 439–461. DOI: 10.1007/s10723-009-9139-y.
  5. ↑ (2006).SZTAKI desktop grid: a hierarchical desktop grid system. Cracow '06 Grid Workshop.
  6. ↑ (2008).SZTAKI Desktop Grid: Building a Scalable, Secure Platform for Desktop Grid Computing. Making Grids Work. pp. 363–374. DOI: 10.1007/978-0-387-78448-9_29.
  7. ↑ (2008).Enabling Java applications for BOINC with DC-API. Distributed and Parallel Systems: In Focus: Desktop Grid Computing. pp. 3–12.
  8. ↑ SLinCA@Home. Wikipedia. Retrieved 2026-06-29.
  9. ↑ (2008).EDGeS: A Bridge between Desktop Grids and Service Grids. 2008 Third ChinaGrid Annual Conference. DOI: 10.1109/ChinaGrid.2008.44.
  10. ↑ Sztaki Desktop Grid. BOINC Confederation forum. Retrieved 2026-06-29.
  11. ↑ SZTAKI Desktop Grid. BOINC Wiki. Retrieved 2026-06-29.
  12. ↑ (2005).Generalized radix representations and dynamical systems I. Acta Mathematica Hungarica. pp. 207–238. DOI: 10.1007/s10474-005-0221-z.
  13. ↑ (2008).Exhaustive search methods for CNS polynomials. Monatshefte für Mathematik. pp. 421–430. DOI: 10.1007/s00605-008-0005-y.
  14. ↑ (2010).SZTAKI Desktop Grid: Adapting Clusters for Desktop Grids. Remote Instrumentation and Virtual Laboratories. pp. 133–144.
  15. ↑ ISDEP, a fusion application deployed in the EDGeS project. Retrieved 2026-06-29.
  16. ↑ 16.0 16.1 16.2 16.3 16.4 SZTAKI Desktop Grid - Discontinued. HUN-REN SZTAKI. Retrieved 2026-06-29.
  17. ↑ Sztaki Desktop Grid. BOINC Confederation forum. Retrieved 2026-06-29.
  18. ↑ ISDEP, a fusion application deployed in the EDGeS project. Retrieved 2026-06-29.
  19. ↑ (2009).Mapping of (2+1)-dimensional Kardar-Parisi-Zhang growth onto a driven lattice gas model of dimers. Physical Review E. DOI: 10.1103/PhysRevE.79.021125.
  20. ↑ (2010).Surface pattern formation and scaling described by conserved lattice gases. Physical Review E. DOI: 10.1103/PhysRevE.81.051114.
  21. ↑ Exploring University Classes in Nonequilibrium Systems on SZTAKI Desktop Grid. Third AlmereGrid Desktop Grid Experience workshop. Retrieved 2026-06-29.
  22. ↑ Sztaki Desktop Grid. BOINC Confederation forum. Retrieved 2026-06-29.
  23. ↑ SZTAKI Desktop Grid. Retrieved 2026-06-29.
  24. ↑ (2015).Desktop Grid in the Era of Cloud Computing.
  25. ↑ DesktopGrid. BOINC/boinc GitHub Wiki. Retrieved 2026-06-29.
Retrieved from "https://boincsynergy.ca/wiki/index.php?title=SZTAKI_Desktop_Grid&oldid=1663"