BOINC FAQ: Introduction To BOINC
From Unofficial BOINC Wiki
If you have an FAQ not listed here, add it by edting the page, and adding your question to the bottom surrounded by double equals signs, like:
==Question==
and somebody will answer it, when it has been found.
[edit] What does BOINC mean?
Berkeley Open Infrastructure for Network Computing.
[edit] So, what is "BOINC"?
The BOINC Software is not a specific application program designed to perform a particular science project. It is an architecture that can support many different applications that are doing scientific research. It is a host that supports the Science Applications and performs common tasks so that each and every Project does not have to keep "reinventing the wheel" and in effect makes each of the research projects less expensive.
Also this will make it easy for us to run multiple computations simultaneously (like AstroPulse, SETI@Home, and a new SETI for the southern hemisphere search), and to release new versions of these applications without requiring you to manually download and install separate or additional software once you have joined a project.
At the moment we are seeing a significant number of new projects that are in testing and production. Each of these projects is doing science for universities around the world and I am sure that there are more projects waiting in the wings to later show up in active production. The projects that I know about include:
[edit] Active BOINC Projects
These BOINC Powered Projects are in the production phase. They are considered reasonably stable and are doing valid science. Alpha and Beta Projects are less stable and not recommended for general Participation.
| Wiki Project Page | Project Web Site | Project Goal or Comments |
|---|---|---|
| ABC@Home | ABC@Home Web Site | This project will try to find abc triples related to the ABC conjecture. |
| BBC Climate Change Experiment | BBC Climate Change Experiment Web Site | Study of transient climate change model predictions.
As of 10.12.2006, this project is closed for new users. New users should join the main Climateprediction.net (CPDN) project instead. |
| Climateprediction.net (CPDN) | Climateprediction.net Web Site | Study of Climate Models to answer some of the important Climate Change Questions.
Note: Only works with Version 5.x.x or later of the BOINC Client Software. |
| Seasonal Attribution Project | Seasonal Attribution Project Web Site | High resolution variant of CPDN, taking a more in depth look at certain regional areas. As of December 2007, out of work again with no current plans of a 3rd. extension. |
| Einstein@Home | Einstein@Home Web Site | Search for gravitational waves to reveal the presence of neutron stars (pulsars). |
| Leiden Classical | Leiden Classical Web Site | Desktop Computer Grid dedicated to general Classical Dynamics for any scientist or science student. |
| LHC@Home | LHC@Home Web Site | Model and improve the design of the CERN, LHC particle accelerator. Has work infrequently. |
| MalariaControl.net | MalariaControl.net Project Web Site | The MalariaControl.net project is an application that makes use of network computing for stochastic modelling of the clinical epidemiology and natural history of Plasmodium falciparum malaria. Also see: Africa@Home |
| Predictor@Home | Predictor@Home Web Site | Investigate protein behavior. As of February 2007 is beta-testing a new application. |
| Rosetta@Home | Rosetta@Home Web Site | Predict and design protein structures, and protein-protein and protein-ligand interactions using the Rosetta program. |
| SETI@Home | SETI@Home Web Site | Look for radio evidence of extraterrestrial life. |
| SIMAP@Home | SIMAP@Home Web Site | Study of protein similarities. Has work infrequently, check home page for more info. |
| SZTAKI Desktop Grid | SZTAKI Web Site | Hungarian project. This project is exploring a mathematical domain looking for solutions to a particular problem space. |
| World Community Grid | World Community Grid Web Site | Runs multiple sub-projects, users can choose which project(s) to run.
Note: Only works with Version 5.x.x or later of the BOINC Client Software. |
[edit] Projects, Science Applications, and Platforms
Below is a table of the known BOINC Powered Projects, their Science Applications vs. support by Platform.
| Project | Science Application | Operating System | Installed Memory-requirement (MB) | Disk Space Required (MB) | Process time (CPU dependant) | Download Size (MB) | Upload Size (MB) | Deadline (days) |
|---|---|---|---|---|---|---|---|---|
| ABC@Home | abc-finder | Windows, Linux, MacOS Intel | 104.91 | 48 | 0.5 h | 0.01 | 0.01 | 7 |
| BBC Climate Change Experiment | hadcm3l | Windows, Linux | 96.00 | 600 | 5 months | 0.02 | 5-6 (16x) | 347 |
| ClimatePrediction.net | hadam3 | Windows, Linux | 1464.85 | 1024 | 4 weeks | 13 | 30 | 180 |
| hadcm3l | Windows, Linux, MacOS Intel | 96.00 | 600 | 5 months | 0.02 | 5-6 (x16) | 347 | |
| hadsm3 | Windows, Linux, MacOS Intel | 61.1 | 573 | 3 weeks | 0.02 | 5-10 (x3) | 345 | |
| Seasonal Attribution Project | hadam3 | Windows, Linux | 256 (uses 430, recommended 1024) | 1024 | 4 weeks | 13 | 30 | 180 |
| Einstein@Home | einstein-S5R3 | Windows, Linux, Mac OS X, Other(*3) | 57.22 | 96 | 24 h | 4-30 | 0.16 | 14-21 |
| Leiden Classical | Classical | Windows, Linux, Mac OS X, FreeBSD | ? | ? | ? | ? | ? | 7 |
| trajtou-cu111 | Windows, Linux | ? | ? | ? | ? | ? | ? | |
| trajtou-pd110paw | Windows, Linux | ? | ? | ? | ? | ? | ? | |
| trajtou-pt111 | Windows, Linux | ? | ? | ? | ? | ? | ? | |
| LHC@Home | garfield | Windows, Linux | ? | ? | ? | ? | ? | ? |
| sixtrack | Windows, Linux | 57.22 | 29 | 1-10 h | 0.024 | 0.034 | 4-8 | |
| MalariaControl.net | malariacontrol | Windows, Linux, MacOS Intel | 85.84 | 191 | 2h | 0.07 | ? | 3.5 |
| Prediction of Malaria Prevalence | Windows | ? | ? | 0,5 h | ? | ? | ? | |
| malariacontrol test version | Windows, Linux, MacOS Intel | ? | ? | ? | ? | ? | ? | |
| Estimation of parameters of infection dynamics | Windows | ? | ? | ? | ? | ? | ? | |
| Rosetta@Home | rosetta | Windows, Linux, Mac OS X | 95.37 (recommended 256) | 96 | 3 h (by default, but configurable) | 2.8-4 | 0.02 | 10 |
| SETI@Home | astropulse | Windows, Linux, Mac OS X, Other(*4) | 256 | 128 | 5? days | 8 | ? | 14 |
| setiathome-enhanced | Windows, Linux, Mac OS X, Solaris, Other(*1) | 31.00 | 32 | 1.5-36 h | 0.36 | 0.02 | 4.3-55.1 | |
| SIMAP@Home | Hmmer | Windows, Linux, Mac OS X, Other (*2) | 28.62 (peak 60+) | 48 | 3 h | 5.5 | 0.1 | 8 |
| simap | Windows, Linux, Mac OS X, Other(*2) | 95.37 | 96 | 1.5 h | 2 | 0.7-1.8 | 8 | |
| SZTAKI Desktop Grid | search | Windows, Linux, Mac OS X | 9.54 | 0.95 | 0.9 h | 0.01 | 0.01 | 10 |
| World Community Grid
(uses 'on the fly' up & download compression in BOINC v5.8.xx) | African-Climate@ Home | Windows, Linux | 512 ? | 1024 | 9 h | 77 | ? | 5 d |
| dddt (Discovering Dengue Drugs – Together) | Windows, Linux, Mac OS X | 250 | 600 total, 50 available | 4-13 h | ? | ? | 14 | |
| faah (aids) | Windows, Linux, Mac OS X | 119.21 (recommended 250) | 72 | 5-12 h | 0.1-0.2 | 0.3 | 9 | |
| hcc (cancer) | Windows, Linux, Mac OS X | 95.37 (recommended 250) | 48 | 5-12 h | 0.1-0.2 | 0.3 | 9 | |
| HPF2 (Rosetta) | Windows, Linux, Mac OS X | 190.74 (recommended 250) | 96 | 2-24 h | 1.4 | 1.9 | 11 | |
| Project | Science Application | Operating System | Installed Memory-requirement (MB) | Disk Space Required (MB) | Process time (CPU dependant) | Download Size (MB) | Upload Size (MB) | Deadline (days) |
- Note on memory, this is computers total installed memory, but remember a little bit is always "lost" in bios so a computer needs more memory than noted in the table to get any work.
- Mac-support, projects marked Mac OS X (Not Intel) does not support Intel-Macs.
- Some projects do not provide Work Units for their projects continuously, but rather periodically (yellow) or in irregular intervals (red)
- Some projects have Tasks with various Process-times, these are marked yellow.
- SETI@Home Enhanced has various deadlines, depending on expected Process-times.
- LHC@Home also varies the deadline of Tasks, based on how time-critical they are.
- Einstein@Home uses Locality Scheduling, this means same large download is normally re-used for many Tasks.
- Process Times is for roughly a 2 GHz computer, and is by no means exact. Also, other things like cpu-type, cache-size and memory-speed can greatly influence Process Times, so same computer can perform better in one project while significantly worse in another. Still, the table should hopefully give a rough idea how a computer will perform in different projects.
- Other(*1) : SETI@Home Enhanced binaries are available for AIX, FreeBSD on several architectures, HPUX (PARISC+IA64), IRIX, Linux (Alpha, ARM, HP PARISC, IA64, MIPS, MIPS EL, PPC/Power, PS3,SPARC, zSeries), OpenBSD (i386, PPC/Power and x86_64), OS/2, NetBSD (i386, MIPS EL, PPC/Power), Solaris (SPARC, x86 and AMD64) via the BOINC Third Party Site.
- Other(*2) : SIMAP binaries are available at the SIMAP applications for i386 CPUs and UNIX Site. Binaries of the SIMAP app are available for AIX, FreeBSD i386, HPUX (PARISC and IA64), Linux (Alpha, HP PARISC, IA64, PPC/Power/PS3 and SPARC), NetBSD i386, OpenBSD i386, Solaris (SPARC and x86) and TRU64. The HMMER app is available for FreeBSD i386, HPUX (PARISC and IA64), Linux (HP PARISC, IA64, PPC/Power/PS3 and SPARC), Solaris (SPARC and x86).
- Other(*3) : Einstein@Home offers Windows, Linux x86, FreeBSD MacOS PPC and Intel optimized version via the Beta Testing Site. Irix, Solaris, Linux Alpha,... applications are planed.
- Other(*4) : SETI@Home Astropulse binaries are available for FreeBSD i386, HPUX (PARISC), Linux (IA64, PPC/Power, PS3), MacOS (Intel and PPC), OpenBSD (i386), NetBSD (i386), Solaris (SPARC, x86 and AMD64) via the BOINC Third Party Site.
[edit] Other BOINC Projects
These BOINC Powered Projects are also available for general participation. Because these sites may not have information in the English language, unfortunately the most universal language in use (not my fault, honest!), we must caution that it may be difficult to get good information about the project.
| Project | Project Web Site | Project Goal or Comments |
|---|---|---|
| Cell Computing | Cell Computing Web Site | Japanese; currently this Project requires nonstandard BOINC Client Software. |
[edit] Alpha Test/Beta Test Projects
These BOINC Powered Projects are potentially high-risk to your computer. Choose to Participate only if you are confident of your ability to cope with potential problems which may occur.
[edit] Completed and Inactive Projects
[edit] What is this networked computing you mentioned?
Networked computing, also called Distributed Computing, grid, or mesh computing is a system whereby many small machines are connected to perform work. Instead of one large computer working on the entire problem using all the data at the same time we "eat-the-elephant" one bite at a time. The problem space and data set is divided up into smaller "chunks" and those are processed one small bit at a time.
One of the interesting developments from the SETI@Home Classic project was the realization of the effectiveness of using PC class machines to perform large-scale computing projects. The success of SETI@Home has inspired other projects; and stretching the research dollar at universities through the use of projects like this are turning idle PCs into a resource that can be doing useful work at all times.
[edit] Why not use a "supercomputer"?
Very simply it comes down to cost. Typically, Supercomputer machines are at the high end of price spectrum. Getting time on these machines is hard because they are so expensive to buy, that there are just not very many available. Additional costs related to these machines comes from the specialized environments these machines require. Since only some problems flat-out require a machine like that there was some research done to see if there were alternatives.
The primary alternative of course is to just use simpler machines that just do not have those stringent requirements and to possibly accept slower processing; or, divide the problem into smaller "chunks" so they do not overwhelm the smaller machines. This has been proven to work so well that, interestingly enough, almost all of the new super-computer machines are actually constructed from a room full of typical PCs.
The race still goes on, though, to make bigger and faster machines. If the BOINC System works out the way that I expect that it will, I believe it will create one of the largest capacity computers in the world. I do not believe, however, that it will be mentioned in anyone's listing of the biggest supercomputers.
[edit] Also See
- Top 500, a list of the top supercomputers.
[edit] I wish I could see a super-computer. Is there one nearby?
Closer than you think. As soon as you download the BOINC Client Software and any project's Science Application, well then you are looking at a super-computer!
Well, at least one node in what is the worlds largest super-computer.
[edit] How can you say SETI@Home or BOINC is the worlds largest super-computer?
SETI@Home Classic had over 5,000,000 people that joined with a huge number of machines that returned Results. Let us say that there were only 500,000 active machines contributing results. Well, the typical super-computer that uses PC class machines as its processing elements usually have about 1,000 to 5,000 nodes/CPUs. Well, they still have a ways to go to catch up to us.
We can expect this to rise as the BOINC Client Software takes hold and starts to serve as the main "backbone" of university distributed computing activity. As a matter of fact, with the mingling of the Projects we can expect to see even more interest in the various projects because of more projects "advertising" for the group.
What that will mean is that those that have no interest in space aliens might just be interested in making cancer drugs. Or, any one of the future projects that start on BOINC. My personal expectation is that with the BOINC Client Software reducing the total cost of a Project, there will be new projects that in the past could not get the chance to go operational because to the development costs. Now, with the BOINC Client Software taking care of a large portion of the total cost we can spend the project's development budget on just the creation of the Science Application.
[edit] So are you saying SETI@Home was the only project that had merit?
Not at all. Forgive the earnest enthusiasm for a project I have been involved with. Many Projects stretching back to the earliest computers all contributed. But the SETI@Home Classic project does have a little bit of a spot as a major force in the popularization of distributed computing on PCs utilizing only the PC's idle time on a world-wide basis.
And I have to admit, that at the moment, I am starting to lean towards other BOINC Powered Projects as more of them come on-line. A comment today (as I write this) on one of the Forums talked about two Projects one trying to develop climate prediction models (strangely enough named: Climateprediction.net (CPDN)) is working to save the world in one way, and another Project just entering Beta Testing (Orbit@Home) is also going to be working to that end on another aspect of protecting the Earth.
[edit] Why do I need BOINC anyway?
The BOINC Client Software is the main support system, the Science Applications rely on the BOINC Client Software for all the support tasks. The BOINC Client Software is a "helper" program that provides an environment for a Science Application the same way that Microsoft Windows® provides an environment for programs like the BOINC Client Software to run on your computer. The BOINC Client Software is a critical component in the whole scheme of Distributed Computing.
[edit] What does the BOINC Software do?
The original project that is the ancestor to the BOINC Client Software was SETI@Home Classic. This was a program that was looking for possible signals from outer space from other civilizations. We really were looking for E.T. to see if his parents were answering his phone call. During the course of that Project some drawbacks from the design of that program became evident. Also Berkeley wanted to do additional science projects and came to the realization that they might be duplicating a lot of infrastructure elements.
So, for every dollar the spent on duplication is one less dollar to spend on the actual science.
[edit] What were the drawbacks to the original SETI@Home?
SETI@Home Classic consisted of many interconnected programs - Screen Saver, Data Server, web-page scripts, and so on. This software served the original project very well. But, limitations in the design and implementation illustrated that there might be a better way to perform some tasks.
Specific flaws included:
- Inability to automatically update the Science Application to the latest Version.
- The inability of the program to buffer work without the support of another application.
- Insufficient work might idle the client machines "wasting" the donation of the Participants.
- "Cheating" to gain "Credit" when in fact the submitted Results were "forged".
- Participation in multiple Distributed Computing projects was difficult.
- The program could only use one processor on a Multi-Processor machine.
[edit] You mentioned that a client may become "idle" how does could that happen?
For example, suppose you have joined the SETI@Home Project and the radio telescope used by SETI@Home is shut down for repairs and we temporarily run out of data to analyze. With no new data to process no new data is sent to the Participant's Computers and they will no longer perform calculations.
Other possible reasons that the Participant's Computers might run out of work include:
- Data gathering is halted because of machine failures or maintenance.
- Data can not be distributed because of network failures, outages, or overloading.
- Temporary losses of funding.
- Data distribution is halted due to errors or failures of project servers.
- Data distribution is halted while the results are examined and conclusions are formulated.
- Data distribution is halted when it is apparent that a new Science Application is needed and the project has to be stopped while it is developed and distributed.
[edit] How can someone "Cheat"?
With out protections built in, some participants will attempt to "forge" results to increase the "Credit" earned for participating in a BOINC Powered Project. This seems like strange behavior when the whole point is to donate time to perform science for the benefit of society at large. But, there is a small class of Participants that will get into a project to obtain the false gratification of producing a large number of Results even if they have to "forge" them.
We talk about this in great detail in the BOINC FAQ: Credit.
[edit] So what is this "credit" thing you talked about?
Well, mankind seems to have an innate competitive streak in him. And to be honest there is a sense of satisfaction in doing something well and gaining some "bragging rights" that we can use at the next party. In the original SETI@Home, project credit was based solely on the number of Work Units that the Participant processed. The only flaw in this was that the Work Units varied some in the amount of time it takes to process them. This created some angst because to some of the competitors, someone else seemingly got more credit because they did more "faster" Work Units. And since for these Participants the Credit earned was the important thing, well, it was a problem.
For the third time already; go to read the BOINC FAQ: Credit. Good grief, I don't want to talk about it here.
[edit] SETI@Home had a screen saver, what happened to it in BOINC?
The Screensaver capability is still there for those projects that want to use it. The BOINC Client Software version of SETI@Home will have a Screensaver available for those that want to use it. As a matter of fact it is a vastly improved version that will allow the Participant control over many of the display parameters. This will allow even greater amounts of Participant customization.
We do have to take this moment to stress that it is possible that the Project you wish to support will not have a Screensaver mode. The capability is a part of the BOINC Client Software, but it is not a requirement. So it is possible that the project might not have the ability (or money) to make a screen saver. In these cases, the BOINC Client Software can, and does, act as a screen saver with a bouncing BOINC Logo on the screen.
This is not too likely, a Screensaver is just to good of an advertisement for the project that it is unlikely to occur. But if I did not make the disclaimer, well, then the next 26 projects that make a BOINC capable version of their program will not include a screen saver. So, it is possible that the screen saver will not be available. Not likely, but possible.
Note that the BOINC Owner's Manual has more information on setting up the screen savers. You can also look at the related FAQ topic: "The constant movement of the SETI@Home screen saver are driving me crazy; how can I stop it?" for another problem!
[edit] Why shouldn't I install BOINC on every machine I can find?
Every program running on a computer uses resources on that machine. The BOINC Client Software's use of these resources means that they are not available to be used on programs that are authorized and intended to run on that machine.
And, depending on the specific company or governmental agency, these are things that can happen to you if you try to run the BOINC Client Software on computers you shouldn't have:
- Dismissal; yep, you can get fired.
- Fines; (use of computers belonging to government agencies may violate public law).
- Imprisonment; (use of computers belonging to government agencies may violate public law).
- Other penalties and sanctions up to and including 40 lashes with a wet noodle.
So, watch it!
[edit] I pay by the hour for Internet access. How much connect time will BOINC need?
The BOINC Client Software connects only when transferring data to and from the Data Server (to send up Results, Result Data Files, and to get new Work Unit Data Files to process) and from the Project's Web Site only when the Science Application changes. Connections will only occur once every few hours to days (depending on the speed of your computers and the processing time required to complete Work Units) and lasts for about 10 seconds to 5 minutes. When and how the BOINC Daemon connects is part of the Preferences that you are able to set for your computers. All of the actual processing is done while the computer is off-line.
We are not trying to put you off, but it is really hard to give you exact numbers. The problem we are facing here is the complexity of all of the interrelated and linked factors. To give you the number of factors involved, here is a partial list of things that will affect this timing:
- The processing speed of your computer
- The speed of your Internet connection, 56K Dial-Up, Cable Modem, or DSL; heck some offices have a T1 line
- The size of the work units
- The size of the Result Data Files
- The length of you Work Buffer (your "connect every 'x' days" setting)
- The number of the Results to be sent (how long since your last upload)
- Is there a new Science Application that needs to be down loaded
- The reliability of the connection, loss of packets, noise, etc.
- The phase of the moon (Just kidding)
Now, once again I am going to qualify what I just said. Most projects running on the BOINC Client Software have fairly small Work Unit Data Files. That is not a situation that anyone can guarantee for the future. There may be projects that will have larger data transfer requirements. The good news again is that the developers of these projects are well aware of your issues and make every effort to minimize your "out of pocket" costs.
Selecting a project may depend on awareness of things like the size of the Work Unit Data Files downloaded and Result Data Files uploaded. See the table for estimated sizes.
Work Unit Data File Size
| Project | Science Application | Approximate Size of Work Unit Data Files | Approximate Size of Result Data Files |
|---|---|---|---|
| Climateprediction.net | hadsm | 330 Mega-Bytes | |
| Einstein@Home | einstein | 14.1 Mega-Bytes | |
| LHC@Home | |||
| Predictor@Home | charmmB1 | ||
| mfoldB | |||
| SETI@Home | setiathome | 356 Kilo-Bytes |
[edit] What about security?
The primary mechanism used distribute worms and virus programs is e-mail. More to the point, the "infection" is done using an executable program which the Participant runs on their computer. Part of the problem is that many of our e-mail programs all easy routes because they allow the Participants to initiate actions by a simple double click. This automation is a convenience feature that has be hijacked by the creators of the malicious programs.
How does the BOINC Client Software fit into this? Well, quite simply the potential for the normal distribution of a malicious program is limited in that the majority of the communications between your local BOINC Client Software is quite simply a transfer of data, not an exchange of programs.
The Project's Data Server doesn't download any executable code to your computer.
The BOINC Client Software and many of the Science Applications are likely to be Open Source programs and this, in fact, can be a very positive feature for enhancing security. For more details look at the BOINC FAQ on Open Source.
All in all, the BOINC Client Software and Science Applications are much safer than the Browser you're running right now to read this!
[edit] I have tried to find my question in this FAQ but I cannot find anything that makes sense. Am I out of luck?
Not at all. The first suggestion is to read all of the questions. Sometimes we may not have a very good question or answer simply because we missed it. Other cases it is because we call things by other names than what you expect.
If you still did not find an answer there is a link that you can use to put your question in for consideration to be included when the FAQ is updated by sending a message to the author, the great, the wonderful, and magnificent, and incredibly talented Paul D. Buck; lets here some applause … "cheep …, cheep …, neeeek …, breek …". Oh well, darn … Hopefully over time we will get closer to having more of those questions answered than ever before.
In the mean time, you can go to the Forums for the specific BOINC Powered Project and ask your question there.
By the way, there is no such a thing as a "stupid" question.
Keep in mind these points:
- There are things we know, and we know we know them.
- There are things we know, but we don't know we know them.
- There are things we don't know, and we know we don't know them.
- There are things we don't know, but we don't know we don't know them .
Your question falls into category #3.
The only other problem is that most people spend a lot of time in category #4, all the while thinking they are in category #1.
Ask your question, then you can move into category #1, and by the way, the person next to you probably has the same question as you. They are also just as scared of asking, maybe even more scared, than you are. So, help them out. Ask away, and help your fellow traveler; they will appreciate it (well, they should appreciate it, the ungrateful bums).
Also, in this new database supported system, you can use the search box to try to find the key word that will get you where you want to go. You can also look in the "Category: BOINC Glossary" to see if the word you are looking for is in there also.
As a last resort, send me an e-Mail and I will get back to you.
[edit] My electric bill has gone up since I started running BOINC! What is up with this?
By definition, the BOINC Client Software is going to run your computer 100% of the time at 100% of its capabilities. This means that your computer will not enter a "sleep" mode nor enter any other power saving mode.
All of this translates into an increased consumption of power by the computer which in turn will make it run hotter, to the point where a laptop will be very hot.
