This latest version of prime95 has been optimized for Intel's new AVX instruction set. Owners of Sandy Bridge and Ivy Bridge CPUs (Core i3/i5/i7 2xxx and 3xxx) will see a substantial performance boost. Any modern personal computer with Windows, Mac OS X, Linux, or FreeBSD can participate.

How it works:

This page describes what you will need to participate in GIMPS and how it works on your computer. This page is aimed at people with Intel-compatible personal computers, although in many cases it is possible for UNIX and Mac users to participate as well.

In brief, you need:

  1. a relatively modern computer - though very old computers can still be useful to help other computers find primes.
  2. your computer on and running more often than off - at least most of the time - it's OK to go on holidays, vacations, etc.
  3. time and patience - testing for a prime can take weeks even on the fastest computers and months on very old computers.
  4. an Internet connection available at least once every two months - more often is preferred, weekly or daily is ideal .

GIMPS requires a modern PC that is on most of the time. The program runs at the lowest possible priority. You should not see any impact on your system's performance. The program will use about 32MB of memory and about 50MB of disk space. WARNING: Running the program continuously will use about 40 watts of additional power, about the same as an energy-saving compact florescent lamp - your electric bill will go up just a little bit.

Most importantly, you will need a lot of patience. Roughly speaking it will take about a month to run a single primality test - visit the benchmark page for a more accurate estimate on your computer.

Setup Instructions for New Users:

Joining GIMPS is usually as simple as downloading and running the program, answering a few questions, and the program does the rest. There are cash awards for discovering a new Mersenne prime!

  1. If you have not done so, CREATE YOUR USERID. It's optional, but required to check your account details, computer status and performance statistics and to assign computers to your user ID.
  2. Download the appropriate free program for your OS
  3. Create a directory and decompress the file you just downloaded. Windows 7, Vista, and XP have built-in unzip features. Other Windows users can choose from a variety of decompression programs. We use 7-zip. Linux and FreeBSD users should use the standard tar and gzip decompression utilities.
  4. Start the program! (Linux and FreeBSD users should run the program from the command line with a -m switch, i.e. "./mprime -m"). Enter your optional userID created on the website in Step 1, and optionally name your computer. We recommend Windows users select Options, Start at Bootup or Start at Logon.

That's all you need to do! The program contacts a central server called PrimeNet to get some work to do. Usually the program and PrimeNet know the best work to assign, but it's up to you!

You can administer your account and computers on your userID's account page. Once you complete a workunit you can track your standings on the competitive stats pages the server updates every hour (see Top Producers in the menu, left, for more stats). You can monitor each of your computers' progress, even remote-control the work assignments they request using your userID's CPUs page!

Linux and FreeBSD versions can also be set up to run every time you restart your computer. Ask for help at the Mersenne Forum.

Questions and Problems:

Please consult the readme.txt file for possible answers. You can also search for an answer, or ask for help in the GIMPS forums. Otherwise, you will need to address your question to one of the two people who wrote the program. Networking and server problems should be sent to Scott Kurowski. Such problems include errors contacting the server, problems with assignments or userids, and errors on the server's statistics page. All other problems and questions should be sent to George Woltman, but please consult the forums first.


See GIMPS Terms and Conditions. However, please do send bug reports and suggestions for improvements.

What's New:

  • PRP proofs. This allows GIMPS to double-check a PRP test at less than 1% of the cost of a full PRP test!
  • PRP proofs require lots of temporary disk space.
  • PRP proofs require uploading a large proof file.
  • PRP proof verifications require downloading a modest verification file.
  • Proofs automatically uploaded to server in 30.2.
  • First time LL, World-record LL, 100M-digit LL work preference is deprecated.
  • New resource limits menu choice and dialog box. Consult readme.txt before making changes to these settings. Some options previously in Test/Worker Windows and Options/CPU are moved to the resources dialog box.
  • LL-DC and PRP-DC combined into a single work preference.
  • Warning raised if temporary disk space is less than 1.5GB -- you may not get first time prime tests.
  • Thanks to Mihai Preda, the P-1 probability calculator has been improved. This change results in a lower optimal B1 value and higher optimal B2 value.
  • P+1 factoring. A worktodo.txt entry looks like this:
  • Pplus1=k,b,n,c,B1,B2,nth_run[,how_far_factored][,"known_factors"]
  • Unlike P-1, the fact that factors of Mersenne numbers is 1 mod 2p is of no value.
  • Thus, P-1 is vastly more effective at finding factors. A P+1 run is about as valuable as running one ECM curve. P+1 stage 1 is 50% slower than P-1 stage 1 but several times faster than ECM stage 1. P+1 stage 2 is a little faster than P-1 stage 2 which in turn is a little faster than ECM stage 2.
  • Unlike P-1, P+1 has only a 50% chance of finding a factor if factor+1 is B1/B2 smooth. Thus, it makes sense to do 1 or 2 (maybe 3) runs. That is what the nth_run argument is for. There are two special starting values for P+1 that have a slightly higher chance of finding a factor. These special starting values correspond to nth_run=1 and nth_run=2.
  • Like P-1, if how_far_factored is specified, prime95 will ignore B2 and calculate the best B2 value for the given B1.
  • Faster ECM stage 1 when factoring exponents near the limit of an FFT size.
  • Gwnum library new functions give a programmer more tools to reduce normalized adds.
  • A tutorial.txt file is now available to help a programmer use the gwnum library.