BOINC Central

BOINC Central
Project
Status	Active
Category	Multi-project
Compute	CPU
Requires	BUDA
Development
Developer	David P. Anderson, University of California, Berkeley Space Sciences Laboratory
Initial release	November 26, 2021  (5 years ago)
Completed	Boolean Chains, Cislunar Orbit Stability Analyzer
Software
Written in	C, C++
Operating system	Windows, Linux, macOS, Android
Size	~50 MB
BOINC statistics
Stats as of	February 25, 2026  (0 years ago)
Performance	5721.89 GigaFLOPS
Active users	603
Total users	953
Active hosts	1,139
Total hosts	2,235
Metadata
Website	https://boinc.berkeley.edu/central/
License	Open-source software (LGPL)

BOINC Central is a BOINC project that provides a shared service for volunteer computing, allowing people around the world to donate computing power to science research.

Overview

BOINC Manager, the client program that volunteers install to receive and process BOINC tasks.

BOINC Central lets scientists in multiple research areas use volunteer computing without creating their own BOINC project. The project was publicly launched on 26 November 2021 and is operated by the University of California, Berkeley BOINC project.

Unlike a typical BOINC project, which serves a single research team, BOINC Central acts as a shared service. Several independent scientists can make their applications available to volunteers through one website.

Scientists who need high-throughput computing but cannot afford commercial clouds may apply for computing. The project states that it provides access to thousands of computers at no charge, and that it serves independent researchers as well as researchers at academic institutions.

Goal

The project aims to make volunteer computing available to scientists who lack the funding or technical staff to run their own BOINC project.

How It Works

Infrastructure

• BOINC Central is a BOINC project. It is operated by the UC Berkeley BOINC project, and volunteers attach to it in the same way they attach to other BOINC projects.
• Application catalogue. The project's Apps page lists each application with its science keywords, a description, a web page, and whether Docker is required.
• Docker support. The project supports any application packaged with Docker. Several applications on the Apps page are marked as requiring Docker.
• Applying for computing. Scientists apply for computing through the project's web form, or they can contact David P. Anderson directly.

BUDA: BOINC Universal Docker App

BOINC Central describes BUDA as BOINC's new framework for Docker-based applications. The framework lets scientists run Docker-packaged applications across the volunteer network. Both completed BOINC Central sub-projects, Boolean Chains and Cislunar Orbit Stability Analyzer, used BUDA.

Applications

The Apps page lists the applications that BOINC Central currently distributes. As of October 2026 it lists eight entries.

Application	Science keywords	Docker required (per listing)
AutoDock	Biology and Medicine; Drug discovery	Not marked
AutoDock Vina	Biology and Medicine; Drug discovery	Not marked
AutoDock Vinardo	Biology and Medicine; Drug discovery	Not marked
D48-24 Starday Dime	Mathematics and Computer Science; Number Theory	Yes
Boolean chains	Mathematics and Computer Science	Yes
Nova API	Not listed	Yes
Prime Queen Attack	Mathematics and Computer Science	Yes
Symplectic Attention & Error Envelope	Physics; Mathematics and Computer Science; Artificial Intelligence and Cognitive Science	Yes

AutoDock family

AutoDock predicts how small molecules, such as substrates or drug candidates, bind to a receptor such as a protein. AutoDock is a suite of automated docking tools, and it is especially effective for protein-ligand docking.

AutoDock Vina is a version of AutoDock that improves speed and accuracy. It is the successor to AutoDock and has an improved local search routine that uses multiple CPU cores.

AutoDock Vinardo is another version of AutoDock that improves speed and accuracy. It is described in a 2016 paper in PLOS ONE, which reports that the Vinardo scoring function outperforms Vina in the tests the authors ran. Vinardo is available as an option in smina, a fork of Vina released under the GNU General Public License v2.0.

D48-24 Starday Dime

D48-24 Starday Dime is a Mathematics and Computer Science application, listed under Number Theory, that searches for runs of consecutive positive integers that all have the same number of divisors. A forum post by the project scientist, dated 9 October 2026, describes the application's goal as "The Hunt for D48-24".

Background

The number of divisors of $n$ is written $\tau (n)$. If $n=\prod _{i=1}^s{p}_i^{{\alpha }_i}$ is the prime factorization of $n$, then $\tau (n)=\prod _{i=1}^s({\alpha }_i+1)$. Two integers $m$ and $n$ are called equidivisible when $\tau (m)=\tau (n)$. For example, $\tau (242)=\tau (243)=\tau (244)=\tau (245)=6$.

For each $k$, let $M(k)$ be the maximal length of a run of consecutive equidivisible integers with exactly $k$ divisors each. $M(k)$ is bounded for every $k$. The OEIS entry for the longest run of consecutive integers with exactly $n$ divisors records a run of 20 integers with 48 divisors each. Combined with an upper bound, this gives $20\le M(48)\le 31$.

History

The problem was first addressed in the work of Ivo Düntsch and Roger Eggleton, who systematically described such "equidivisible integers". According to the forum post, their work, published in 1989, set a world record for that time with a run of 9 consecutive numbers. R. K. Guy's problem collection separately reports that Düntsch and Eggleton discovered several sequences of 7 numbers, two of 8, and one of 9, each with 48 divisors, in 1990.

Further theoretical foundations and algorithmic constraints were developed by participants in the "Dream Pentadecathlon" thread on the dxdy forum, including Vladimir Letsko. Letsko's results are preserved on arXiv, and Letsko and Dziubenko published further work in Russian in 2016.

Current records

In 2022 the D12-15 Pentadecathlon was discovered. It is a continuous run of 15 consecutive integers, each with exactly 12 divisors. It has stood since then as the only such sequence of length greater than 7 for which the absolute maximum possible length has been achieved. Several more runs of this kind have since been found, and they are listed on page 100 of the dxdy thread. The smallest known D12-15 sequence is a 35-digit number.

Whether this run is the absolute minimal one remains an open question. Minimality has also not been established for the shorter runs D12-14 and D12-13.

The forum post states that 48 is the most promising number of divisors for finding the longest runs, and that the theoretical ceiling for a run length can reach 31 numbers in a row. A run of length 20 was first found in 2022. About three years later, the project reports runs of 21, 22, and 23 consecutive integers, the latest of which is a 54-digit number. The post says that exceeding the 31 ceiling is theoretically possible.

Challenges

The project's scientist has invited participants to pursue two further goals:

• Half-length record. Find a consecutive run whose length is exactly half the number of divisors of each of its integers, that is, $L=\tau (n)/2$.
• Swallows and bustards. A 24-element run that contains exactly two integers with an incorrect number of divisors is called a "swallow" (lastochka). A run with only one such integer is called a "bustard" (drofa). The first participant to spot a swallow or a bustard in their logs will receive special recognition.

Other listed applications

• Nova API is listed as a Docker-based application. The listing gives no description or science keywords at the time of writing.
• Prime Queen Attack is listed under Mathematics and Computer Science, with no description given. Its name matches a recreational problem posed by G. L. Honaker Jr. in 1998. In that problem, a queen is placed on a board carrying a knight's tour numbered in visiting order, and the aim is to attack as many prime numbers as possible. The listing does not say whether the application is connected to that problem.
• Symplectic Attention & Error Envelope generates multi-precision trajectories of chaotic systems, such as the double pendulum, using fixed-step symplectic integrators with optional Kahan compensation. It extracts attention features and energy-envelope slopes, and it studies the relationship between numerical error growth and attention entropy. Its stated aim is to build physically interpretable attention mechanisms for long-term chaotic prediction.

Completed Sub-Projects

BOINC Central volunteers have completed the following sub-projects.

Boolean Chains (completed)

Boolean Chains was the first project hosted by BOINC Central. It searched for short boolean chains, a topic that Donald E. Knuth analyses in section 7.1.2 of The Art of Computer Programming, Volume 4A.

A boolean chain on inputs $x_1,\dots ,x_n$ is a sequence of operations $x_i=x_{j(i)}{\circ }_i{x}_{k(i)}$ for $n+1\le i\le n+r$, where each ${\circ }_i$ is a binary operation. The goal is to compute a set of target functions with as few operations as possible, because a shorter chain corresponds to a smaller circuit.

Knuth's example is a seven-segment display for hexadecimal digits. The four input bits encode the digits 0 to 15, and each of the seven segments is controlled by a boolean function of those inputs. Knuth's method gives a chain of length 22 for the seven-segment decimal display, and David Stevenson later found a 21-step chain. The project's author, Oliver Runge, reports several chains of length 20 for the hexadecimal display, one step shorter than the previously shortest chain in the book.

The seven-segment display problem: a classic combinatorics challenge that motivated the Boolean Chains project.

Milestones and Results

On 31 May 2025 the project reached a milestone. The search space for $N=15$ and $L=21$ was completed, with $37,444,981,252,103,000$ chains generated using 2,139 days of computing. The project was subsequently completed. Volunteers supplied the equivalent of $450$ CPU-years of computing power. The next phase was planned to explore the case $N=16$ and $L=22$.

Cislunar Orbit Stability Analyzer (completed)

The Cislunar Orbit Stability Analyzer project studied the stability of orbits in the Earth/Moon system. Like Boolean Chains, it used the BUDA framework.

Results

The project has been completed. The BOINC Central news of 24 March 2026 states that the project led to a paper that was then in review. The team reported that, like Boolean Chains, the project involved scaling challenges, including batches of 10,000 jobs and multi-gigabyte output file downloads.

Technical Infrastructure

AutoDock Vina

AutoDock Vina is used for molecular docking and virtual drug screening.

AutoDock Vina is an open-source program for molecular docking, developed by Oleg Trott and Arthur J. Olson. It is maintained by Scripps Research, specifically the Center for Computational Structural Biology (CCSB), which Olson leads. Vina is widely used in computational drug discovery to predict how small molecules, such as potential drug candidates, bind to protein receptors.

The 2010 paper that introduced Vina describes a new scoring function, efficient optimization, and multithreading that together improve the speed and accuracy of docking. BOINC Central's distributed infrastructure allows researchers to run large virtual screening campaigns, docking many compounds against a protein target, using volunteer computing power.

The BOINC Platform

BOINC (Berkeley Open Infrastructure for Network Computing, pronounced /bɔɪŋk/, rhyming with "oink") is an open-source middleware system for volunteer computing, developed at the UC Berkeley Space Sciences Laboratory. It was first released on 10 April 2002, and it was originally developed to support SETI@home.

As of 13 August 2026, BOINC brings together 25,314 active participants with 130,853 active computers worldwide, processing on average 15.220 PetaFLOPS per day. BOINC is funded by the National Science Foundation.

Researchers

Name	Research interests	Location	Project
Oliver Runge	Computer science	Germany	Boolean Chains
Lezhe Gao	Cislunar orbit stability	Not stated	Cislunar Orbit Stability Analyzer

Project Team / Sponsors

David P. Anderson leads the BOINC project, under which BOINC Central is operated. The BOINC project is operated by the B'erkeley Open Infrastructure for Network Computing team at the University of California, Berkeley. Anderson also led SETI@home. Anderson's homepage is linked from the project website.

How to Participate

Volunteers can contribute computing power as follows:

1. Download and install the BOINC client.
2. Attach to BOINC Central using its URL: https://boinc.berkeley.edu/central/
3. Choose which applications to run within the project.
4. The BOINC client receives, processes, and returns tasks automatically.

BOINC Central is one of the projects currently available on the Android BOINC app.

Scientists who wish to submit workloads can apply for computing or contact David P. Anderson.

See Also

• wikipedia:Berkeley Open Infrastructure for Network Computing
• Volunteer computing
• wikipedia:AutoDock
• Molecular docking
• David P. Anderson
• SETI@home
• World Community Grid
• Yoyo@home
• BOINC Alpha Test
• WUProp@Home

External Links

• BOINC Central official website
• BOINC Central applications
• BOINC Central forum thread 93: The Hunt for D48-24
• BOINC Central news archive
• BOINC main website
• Boolean Chains project website
• BOINC Central computing projects
• Publications by BOINC projects
• David P. Anderson's homepage
• Dream Pentadecathlon thread on dxdy
• OEIS A119479, longest run of consecutive integers with n divisors
• BUDA framework documentation