Thread: 45th Known Mersenne Prime Found???

1. 45th Known Mersenne Prime Found???

On August 23rd, a computer reported finding a new Mersenne prime to the server! Because I was on vacation, verification did not begin until the 26th. Verification should take roughly 2 weeks to complete. Check back here for updates!
Source

2. Ooooh!

3. Will the number have more than 10 million digits? That's the question...

4. Because I was on vacation, verification did not begin until the 26th. Verification should take roughly 2 weeks to complete. Check back here for updates!
2 weeks puts it around the 10th. /me wonders if he fudged the timetable so people wouldn't bug him about it as much.

Either way, in 2 weeks we'll know.

5. Myself and the other person doing the verifications are around 40% complete now, so a little more than another week left to go...

6. I don't know how that Mersenne Prime Project exactly works, but how do you get a WU/prime to verify? "per accident" or voluntary? I'm just curious...

7. Originally Posted by the-mk
I don't know how that Mersenne Prime Project exactly works, but how do you get a WU/prime to verify? "per accident" or voluntary? I'm just curious...
Mersenne numbers are numbers of the form 2 to the power of x minus 1, so the first Mersenne prime is 2^2-1, or 3, the second one is 2^3+1, or 7. The workunits are semi-random in that most people just connect to the server and ask for a specific kind of work, and then crunch whatever they're handed. In the case of running a primality test, they're simply handed one or more n-values, which means bandwidth for running the project is measured in the 10s of bytes(not kilobytes, bytes, 10s of bytes), but when you actually start the test, you might have workunits that are about 10MB to deal with, since it has to actually create the number and then perform the iterations necessary to run the test.

I can probably answer any questions anyone has about this project, just so long as it doesn't deal with the math involved, or particulars about the Windows client. If you run Linux, I can probably help you with the client.

8. The verification runs are not done by the project itself but by different computers, AFAIK even with a different program/method, just to be sure they really are prime. You are not selected to verfy by accident.

9. that's what I wanted to know...

10. Yes, the verification of a suspected prime is not "handed" out by the server, there are specific people on the team that are available to perform such a task.

As was mentioned, to verify, we use completely different code on a different hardware architecture just to be sure. In this case, there are two verifications underway. The original suspected prime was found using Prime95 on an x86 system, it is being verified using MLucas on a Sparc system, and GLucas on an Itanium2 system.

11. Just because I'm anal(love me, or hate me, but I'll never change ) I'd like to point out that verifications of compositeness is done on the server. It's actually possible to ONLY do double-checks. When a test is finished, the computer sends the last 64 bits of the result back to the server for verification. If it's a first-pass test, then verification might not happen for 3-5 years. But 64 bits isn't very much at all, so it's no big deal to store it. If a second run doesn't match the result of the first run, then they keep sending out the test until they get matching residues. But the server only hands out one instance of the test at a time, and a test might take anywhere from a week to years to complete, depending on the exponent and type of computer running the test.

12. Speed up that box DP

46th seem to be found as well

13. Originally Posted by cwhyl
Speed up that box DP
I'm trying captain! I'm givin er all she's got!

Originally Posted by cwhyl
46th seem to be found as well
Yes, but the real question is, is this the 46th or the 45th??? Which came first, the chicken or the egg?

14. Originally Posted by Digital Parasite
Yes, but the real question is, is this the 46th or the 45th??? Which came first, the chicken or the egg?
[off-topic]
Round here, a favorite joke is that the chicken came first, 'cause God don't lay no eggs.
[/off-topic]

15. The first prime has been verified! Second one should be finished verifying on the weekend so the plan is to release all the details and press release on the 15th.

16. 2^43112609-1 (12978189 digits) was found on August 23rd and
2^37156667-1 (11185272 digits) on September 6th.

www.mersenne.org

17. UCLA get's \$100,000 for doing some number crunching that should pay for the utilities bill for it.

18. Originally Posted by gopher_yarrowzoo
UCLA get's \$100,000 for doing some number crunching that should pay for the utilities bill for it.