Slashdot Mirror

One glaring ommission from the FAQ is "Why participate in this?" I guess if you have to ask why, there's no point in asking.

You're right, compared to protein folding, cancer research, and other research projects there is no good reason to run GIMPS. Heck even SETI@home could return something useful. But there are a couple reasons to do it:

• One reason is money. The EFF put up a $100,000 prize for the discovery a 10 million digit prime.
• Another reason is developing new algorithms and enhancing existing ones. The source code for the GIMPS project is available here.
• Their deterministic stress test is great for checking stability of overclocked CPUs. Run ./mprime -t for a week with no errors and you're probably solid.

My only real complaint about GIMPS is the lack of optimized AMD64 clients.

One glaring ommission from the FAQ is "Why participate in this?" I guess if you have to ask why, there's no point in asking.

You're right, compared to protein folding, cancer research, and other research projects there is no good reason to run GIMPS. Heck even SETI@home could return something useful. But there are a couple reasons to do it:

• One reason is money. The EFF put up a $100,000 prize for the discovery a 10 million digit prime.
• Another reason is developing new algorithms and enhancing existing ones. The source code for the GIMPS project is available here.
• Their deterministic stress test is great for checking stability of overclocked CPUs. Run ./mprime -t for a week with no errors and you're probably solid.

My only real complaint about GIMPS is the lack of optimized AMD64 clients.

The FAQ for this endeavor can be seen here.

One glaring ommission from the FAQ is "Why participate in this?" I guess if you have to ask why, there's no point in asking.

GIMPS (Great Internet Mersenne Prime Search

They have Windows, Linux, FreeBSD, and OS/2 clients.

There are a *lot* of people crunching work for SETI@Home...

Gimps (http://www.mersenne.org/prime.htm, large prime number search) is also a popular distributed computing task, and I think there are a few other 'minor players' in the game of soliciting free CPU cycles from the public. Does anyone know offhand how many computers are running each task?

It's easy to use your P4 cpu more than 100% in this meaning. Just participate in the mersenne prime search. The P4 client is highly optimized assembly. When you look inside the source, you find comments like

The FDIV takes 39 clocks and only the last two can overlap with other float instructions. This gives us 37 clocks to do something useful with the integer units.

However, how did you measure your 112% value?

Maik

Not exactly. We know that 2^n-1 can only be prime if n is also prime (if I remember correctly). So while numbers in the form 2^n-1 generally may not have a significantly higher chance of being prime (I don't know), we can choose values of n that are easily 20x as likely to give us a prime number than if we used randomly chosen odd numbers.

The real reason why discovered Mersenne primes are way bigger than any other known primes, is because of the special algorithm for finding them. This is known as the Lucas-Lehmer test (google for it). It isn't feasible to run a brute-force trial factoring for numbers approaching 10 million digits. For a number on the form (2^P)-1 where P is a prime, the Lucas-Lehmer test requires only P iterations of the basic algorithm. It also works especially well for numbers on the form (2^P)-1, because the computer can get away with a lot of simple bit-shifting instead of costlier math operations in terms of CPU cycles.

I've done Lucas-Lehmer testing for the Great Internet Mersenne Prime Search for half a year now, and am currently running it on five different machines. The chances of actually finding a huge prime number are probably much less than of winning big-time in Lotto, but I for one think this is kind of fun.

The linked .txt file is 7.1 MB, so it probably will take a while. The Google cache doesn't do the number justice. In fact, the cached number is divisible by 2...

Personally I use memtest86 http://www.memtest86.com/ on all the tests, I found 5,7,10 very good at spotting my problems. You can burn and image on a cd or put it in lilo. When I got my latest PC, I had lots of problems which I traced back to the memory not working at DDR400 in dual channel mode. Another program is prime http://www.mersenne.org/prime.htm in a torture test mode (but in windows). I noted some else suggested GCC the problem with that is that it does not test ALL memory, so will find some bit problems but many bus related ones.

If you have a Dual channel system and have memory problems try clocking back one speed, it made my memory rock solid.

James

• ...will attempt to recreate a Shakespeare work from randomly selected words found in spam messages
• ...will use randomly selected characters from the spam, translated into numbers, to form pi
• ...will display randomly selected and coloured characters from the spam on the screen as a screensaver
• ...will extract all ACGT (that's DNA: Adenine, Cytosine, Guanine, Thymine) characters from all the spam, and randomly rearrange them trying to form the DNA or a large part of the DNA of some organism
• ...will randomly select characters and numbers from spam msgs to feed seeds to an RSA key generator
• a CPU-intensive combination of the above

Wouldn't you think the government has had some non-public serious hardware dedicated to this for the past decade? If these are the Top 500 "known" supercomputers (http://www.top500.org/list/2004/06/), where are all the Defense Department listings?

The NSA has "worked" closely with vendors supplying encryption equipment since the 1990s (http://www.cnn.com/TECH/computing/9807/27/securit y.idg/

I would believe the government's NSA hardware is probably around 6-10 years ahead of what is commercially available or even known. http://www.hpcc.gov/pubs/blue94/section.4.6.html

My lowly P4 2.4C is running overclocked at 270 MHz FSB (3.24 GHz) using Vcore of 1.65 Volts, air cooling via a Thermalright SP-94 heatsink (that I really should lapp) and Artic Silver 5 compound: 55C at load with 30C ambient, CPU fan at minimum (~2600 RPM). It benchmarks faster than the 3.6E, 3.4C, etc.

Using Sandra to benchmark, and Prime95 to torture-test for a minimum of 24 hours without errors.

You might also want to try undervolting your cpu. At it's rated speed, it probably does not need the default voltage specified, and each 0.025 Volt you can lower it (my Abit KR7A mainboard handles core voltage in steps of 0.025), you gain 1 or two degrees celsius. I can undervolt my old Athlon XP 1700+ to 1.60 volt from the standard 1.75 and it makes my processor about 7 degrees cooler than usual under load, and it's free. You do have to make sure your processor is still stable at the lower voltage, I use Prime95 to make sure my cpu does not make a single error even under stress.

There's this one, it's probably what you want:

http://www.stanford.edu/group/pandegroup/folding/

But I'm quite selfish (and actually interested in primes abd or at least know more about them than I do about protiens), and there are entities offering big prize money for big primes, and if one of my machines finds one, I'll get big bucks:

http://mersenne.org

Here it is: http://mersenne.org/prime7.htm

Uhhhh, a great burn-in program for CPU's for one thing.....money for another

The real reason is to show off how long our digital penis' are on the user page and the top producers pages.

What better way is there to pick up girls than to say you're the biggest GIMP on the planet?

The GIMPS Project found this prime. You too can contribute by downloading the client (for various OSes).

Thought I would drive the point home as this is a great DC project that doesn't receive half the attention of some of the more dubious DC projects...

I have an XP1600+ that overclocks to 2200+ at default voltage and works ideally under a torturous FAH load. In fact, you could view the chip as being underclocked at the factory instead of overclocked by the user!

Now that's a reasonable approach to O/C, taking advantage of a chance occurence where the chip was capable of more without spending extra cash on cooling.

My objection to O/C is where it's done solely for the bragging-rights, or people that spend way more cash then makes sense to gain a measly 5% gain.

BTW, another good O/C testing program is Prime95. It's proven to be so reliable as a stress test over the years, that they've even added a mode where it functions solely as a stress test.

You're clearly speaking for yourself. (I will speak up for myself below - personal opinion follows:)

Personally, I don't want anyone getting my bank account numbers, credit card numbers, tax information, etc. because a couple people who don't do these things don't mind if someone screws around on *their* computer. I will use my own judgment as to what is appropriate on *my* computer.

Basically, I will grant some programs that I trust the ability to "anonymously track bugs" or maybe allow a program like SETI or Mersenne Prime Search. I would be willing to fill out a survey at the time of download, from a web page - not an app running on my PC. I do not expect any ads to randomly come to my PC from the internet. The only company that has a right to do this would be my ISP (maybe), and I would quickly switch. None of my personal information should ever be transmitted without my express consent. Anything else amounts to identity theft or the enabling of such.

winzip
realvnc
ad-aware
html-kit
irfanview
zonealarm
kazaalite
subspace/continuum
prime95
weatherbug
google toolbar

Or, optionally, find a thirty-foot-long alien ray gun and turn it into Shockwave. Don't forget Unicron!

I think unicron is what I use to schedule jobs on my Sun workstation.

And isn't there a distributed computing project somewhere to find Optimus Prime?.

Hmmmmm...

This is slightly OT, but I've found one of the best ways to test (as opposed diagnose) hardware is to install FreeBSD then run "make buildworld" on it...

If you're worried about system stability or processor-cooling effectiveness, a couple of good stress tests are Prime95 (it has a "torture-test" mode) and cpuburn. Somebody complained that the Shuttle SFF boxes tend to overheat and flake out. I built one (with three hard drives...it was for a DVR setup), set it up outside in Las Vegas' summer heat, and fired up cpuburn. It ran without so much as a hiccup.

There are many tools which come with today's GNU/Linux distributions that can be used to test hardware. Knoppix and LiveCD distros are useful for their portability. Full-blown distros can be even more useful due to the sheer breadth of tools that they contain. GNU/Linux drivers and tools can usually be made to output large volumes of information on what they're doing and what they've found. Being open source, they can be more fully understood. Tools can be combined to perform all sorts of tasks and tests.

I have used The Linux Hardware Stability Guide at IBM DeveloperWorks (Part 1 | Part 2) to test and tweak a number of systems. I have found the best stress testers to be a looped kernel compilation, cpuburn and Memtest86. Mprime is good as well.

According to mersenne.org, a single iteration of a small exponent (6.52M-7.76M) takes 0.211 seconds on a PII-400 (not unreasonable for the average user), so a whole primality test will take, at the very least, 0.211*6.52e6 = 1375720 seconds, roughly 16 days, assuming the computer runs 24/7. Oops.

Of course, ignoring the fact that only 222k exponents need double-testing (yet), so the number of emails sent through this method is capped, without including the changes needed to mail servers, the effective destruction of mailing lists...

Here's a nice way. Before someone can send some mail, he has to get some exponent from mersenne.org which needs double-checking, run the primality test and report the low order 64 bits of the final S_{P-2} value, called a residue. If that value matches the value that mersenne.org expects, then the mail goes through.

Nice deterrent for spam, and as a side-effect one more Mersenne exponent has been double-checked.

Here's a nice way. Before someone can send some mail, he has to get some exponent from mersenne.org which needs double-checking, run the primality test and report the low order 64 bits of the final S_{P-2} value, called a residue. If that value matches the value that mersenne.org expects, then the mail goes through.

Nice deterrent for spam, and as a side-effect one more Mersenne exponent has been double-checked.

I have just applied for a patent on the first 10 million digit prime number. Much like SCO, I'm not going to tell you the number (but I can tell you the process by which I discovered it).

When the Gimps Project finds it, I am going to sue them for the $100,000 prize they will collect.

At that time, I will announce the 10 million digit number as 'exhibit A' in my lawsuit.

I like the idea of patenting math.

http://www.mersenne.org/prime.htm

SETI@Home is the most popular project, measured by number of participants, but it was actually the third large-scale, public domain, distributed computing project. A lot of its functionality and design is based on the second project, distributed.net, which in turn is based on some design ideas from the first project, GIMPS.

SETI@Home has definitely done a lot to popularize distributed computing, and has influenced many later projects, including protein folding projects like Distributed Folding and Folding@Home.

To see what other projects are out there, take a look at my site about distributed computing projects. And click on the links to past years (on my main index page) to see just how fast this field of science is growing.

Kirk

P.S. Somebody please /. my site so I don't have to keep plugging it in these SETI@home discussions :-)

And, for those curious, the largest prime curently known is the 40th Mersenne Prime 2 to the 20,996,011 -1, which is 6,320,430 decimal digits in length. If you're wondering what that looks like, and don't mind downloading 6.3 MB, wonder no more.

And, for those curious, the largest prime curently known is the 40th Mersenne Prime 2 to the 20,996,011 -1, which is 6,320,430 decimal digits in length. If you're wondering what that looks like, and don't mind downloading 6.3 MB, wonder no more.

Apparently you need a bit of education about mersenne primes.

A mersenne number is defined by 2^N-1 where N is prime. If N isn't prime, it isn't a mersenne number. (In your example, 2^4-1=15, you didn't have a prime in the N spot, so you aren't even looking at a mersenne number.) Not all mersenne numbers are prime. In fact, most are not. If they were, it would be incredibly easy to find large prime numbers.

2^N-1 is the basic math behind mersenne primes. It doesn't mean that you can plug any number into the N position and come up with a prime. However, if you go the other way, if 2^N-1 is prime, then N is prime, every time.

If N is prime, 2^N-1 is not always prime. However, the likelyhood that it *is* prime increases dramatically as compared to testing a random number with the same number of digits.

2^N-1 where N is not prime is meaningless, and has nothing to do with mersenne numbers.

The point of using mersenne numbers is that testing large numbers (in the millions of digits range) can-and-does take a long time. Mersenne numbers are more likely to be prime than other numbers. So instead of testing everything, if you test just mersenne numbers, you are more likely to find large primes. Testing only the mersenne numbers doesn't mean that you'll find every prime number. There are almost certainly a large number of unknown primes with less than 6.3 million digits - but finding that largest one was based on testing a large number of mersenne numbers.

http://www.mersenne.org/prime.htm will help if you want to know instead of spouting bullshit. There are links there which will explain the math, the history, etc. If you just want to talk without knowing what you are talking about, you are already on the right track.

the 40th Mersenne prime has been discovered 2-3 weeks ago and just proven to be correct. See http://www.mersenne.org/history.htm for more info.

From the website The largest known prime number, discovered by the Great Internet Mersenne Prime Search (GIMPS), has 909,526 digits. The Mersenne search just officially announced yesterday (it had come out to the group last month was was still being confimed by several re-checks) that Michael Shafer discovered the 40th known Mersenne prime, 2^20996011-1. Congratulations Michael. This prime is over 6.3 million digits, beating the previous world record prime by over 2 million digits. So a 6.3 million digit prime is known, replacing a 2 million digit one. The ten million one is not that far off.

You can get the software to try to find the next Mersenne prime here

It's not the most compact form, to be sure, but it is, as advertised, 6 megabytes of primey goodness.

Of course, it didn't occur to me to take a look at the Science section before submitting my own copy of this story (which, since it has several other useful links in it, follows):

Michael Shafer, a graduate student at Michigan State University, took time out for a "short victory dance" upon learning his computer had discovered the 40th known Mersenne prime as part of The Great Internet Mersenne Prime Search. The number itself is 2**20996011-1 and when expressed in base 10, has 6,320,430 digits (zipped copy). However, this is not necessarily the 40th Mersenne prime; there could be another between the previous largest known prime (M39=2**13466917-1, also discovered by GIMPS) and this one. Also worth noting is the still-standing USD$100,000 EFF prize for the discover of the first prime of at least 10 million (decimal) digits. GIMPS clients are available for various operating systems as well as information on how GIMPS would distribute the prize. A press release on the achievement is available as well as several articles. Of course, this also means there's a new largest known even perfect number in town.

Of course, it didn't occur to me to take a look at the Science section before submitting my own copy of this story (which, since it has several other useful links in it, follows):

Michael Shafer, a graduate student at Michigan State University, took time out for a "short victory dance" upon learning his computer had discovered the 40th known Mersenne prime as part of The Great Internet Mersenne Prime Search. The number itself is 2**20996011-1 and when expressed in base 10, has 6,320,430 digits (zipped copy). However, this is not necessarily the 40th Mersenne prime; there could be another between the previous largest known prime (M39=2**13466917-1, also discovered by GIMPS) and this one. Also worth noting is the still-standing USD$100,000 EFF prize for the discover of the first prime of at least 10 million (decimal) digits. GIMPS clients are available for various operating systems as well as information on how GIMPS would distribute the prize. A press release on the achievement is available as well as several articles. Of course, this also means there's a new largest known even perfect number in town.

Of course, it didn't occur to me to take a look at the Science section before submitting my own copy of this story (which, since it has several other useful links in it, follows):

Michael Shafer, a graduate student at Michigan State University, took time out for a "short victory dance" upon learning his computer had discovered the 40th known Mersenne prime as part of The Great Internet Mersenne Prime Search. The number itself is 2**20996011-1 and when expressed in base 10, has 6,320,430 digits (zipped copy). However, this is not necessarily the 40th Mersenne prime; there could be another between the previous largest known prime (M39=2**13466917-1, also discovered by GIMPS) and this one. Also worth noting is the still-standing USD$100,000 EFF prize for the discover of the first prime of at least 10 million (decimal) digits. GIMPS clients are available for various operating systems as well as information on how GIMPS would distribute the prize. A press release on the achievement is available as well as several articles. Of course, this also means there's a new largest known even perfect number in town.

Of course, it didn't occur to me to take a look at the Science section before submitting my own copy of this story (which, since it has several other useful links in it, follows):

Michael Shafer, a graduate student at Michigan State University, took time out for a "short victory dance" upon learning his computer had discovered the 40th known Mersenne prime as part of The Great Internet Mersenne Prime Search. The number itself is 2**20996011-1 and when expressed in base 10, has 6,320,430 digits (zipped copy). However, this is not necessarily the 40th Mersenne prime; there could be another between the previous largest known prime (M39=2**13466917-1, also discovered by GIMPS) and this one. Also worth noting is the still-standing USD$100,000 EFF prize for the discover of the first prime of at least 10 million (decimal) digits. GIMPS clients are available for various operating systems as well as information on how GIMPS would distribute the prize. A press release on the achievement is available as well as several articles. Of course, this also means there's a new largest known even perfect number in town.

Well yeah, that goes without saying.

In fact, it's fairly good evidence that humans *can't* solve those problems on large boards. There's a lot of problems humans can't readily solve that computers can solve easily (multiplying multiple-million-digit numbers without error, to start with).

Of course its cheating, but it does get the job done.

And you are correct about there being more pressing questions, such as prime numbers! :-)

The world needs no more than one computer, it's just very spread out.

Idle cycles that could go to other better projects.

1. Provided Microsoft uses a proper public key infrastructure, brute-forcing this thing could potentially take forever

2. This so that you can feel good subverting an X-Box by making it run Linux

3. By that time the hardware would be definitely obsolete, or X-Box 2 would be out with programs signed with a different key

4. And in any case, buying the X-Box already helps Microsoft. The more units sold, the more games developed.

5. There are tons of other worthwhile distributed computing projects to do out there - Folding@Home, SETI@Home, Mersenne Prime Search etc.

Grow up folks! Running Linux on a hacked X-Box is cool, yes, but this might be going too far...

I don't understand why companies don't include such things on new PCs as an option. Just include the .EXE file for Folding@Home (or one of the lesser projects :), a link on the desktop and an explanation of what the user can do with his/her idle CPU time. The number crunching power of millions upon millions of PCs wouldn't go to waste.

I'm not so sure about that. There's no reason why the CPU or any other device in a normal PC would fail. Programs like Prime 95 require over a year of CPU time for testing a single number for primality. If the algorithm and code is correct, then the only problem with calculating huge amounts of decimals of PI is that it'll require more and more disk swapping as you go along.

Actually, a distributed project for calculating PI decimals would be interesting (unfortunately not very realistic though as it's the kind of task that would be extremely hard to parallelize for separate units).

How about the search for Mersenne primes: prime numbers that take the form (2^n)-1. More information about them is available on that website, naturally.

Perhaps you can help by searching for factors of some obscenely large numbers.

That's why I use Prime 95. It's designed to find hideously large prime numbers, so while not as useful to the human race as the biochem ones my spare cycles aren't going to be making anyone else rich. And I'm a mathematics geek, so it's pleasing in that sense as well.

Anyone who doesn't use a cycle-sucker is scum. Think about it - how much power is wasted through PC idle time? How much money does that wasted time cost you, through your power bills? How many people will die today for the want of a few pence to buy some food or water?

Distributed computing helps me sleep at night.

Their online calculator provides for a bit of fun. It accepts any valued entry for your computer speed and number of hours in a day...
With a 98Ghz P4 on some distant planet with 285.6 hours in a day, I can crank out a test of 1 10 million digit Mersenne Prime every 24hrs!

See you suckers at the bank!