(M48) NEW MERSENNE PRIME! LARGEST PRIME NUMBER DISCOVERED!

[Prime95]

Originally Posted by dbaugh

Some rough calculations show that $1,000,000 was spent on electricity to run GIMPS over the last four years.

Ouch. Don't let the bean counters at University of Central Missouri find this out!

[ewmayer]

Originally Posted by Prime95

Ouch. Don't let the bean counters at University of Central Missouri find this out!

Just spin it as "GIMPS has saved you a $million in heating costs".

[dbaugh]

They are only 10% of GIMPS and probably pay a third of what the average GIMPster pays for electricity. They also probably have their own cooling plant for the campus which is another electricity cost savings. OTOH they are not running the most efficient computers.

[dbaugh]

I had my gas service disconnected two years ago as it was only used for central heat.

[Dubslow]

Originally Posted by Prime95

Ouch. Don't let the bean counters at University of Central Missouri find this out!

That's why I haven't actively pursued the "let's use these wasted cycles!" line of attack (or any other for that matter) here at my place. The last thing it needs is more expense. (Though the heating may offset it, who knows.)

[chalsall]

Originally Posted by Dubslow

That's why I haven't actively pursued the "let's use these wasted cycles!" line of attack (or any other for that matter) here at my place. The last thing it needs is more expense.

That's the elephant in the room. In the "olden days" CPUs and OSs busy-waited on unused cycles. That is no longer the case.

Originally Posted by Dubslow

(Though the heating may offset it, who knows.)

Only for those who need heating. Many of us very much don't!

[Prime95]

Originally Posted by davieddy

GIMPS:
Average gap from expo 21M to 43M ~1.11
Max was ~1.15
Theory says 1.31 is 50/50.
Average of a random selection of gaps less than this is ~ 1.11.
Mean/expected gap is ~1.47.

Annoyingly, 1.34 is just "tails" (assuming ~58M for our new baby).

I originally called M#48 a "heads" because it was less than the mean/expected gap of 1.47. I now accept davieddy's 1.31 definition of heads/tails. My incorrect definition would result in many more heads than tails in the long run.

So, my apologies for calling M#48 a heads. Maybe, we'll find a new prime below M#48 and my call will then be premature rather than plain wrong.

Note: Applying the 1.31 rule means we have a 50-50 chance of finding a prime below 76M (ignoring untested exponents below M#48 and already tested exponents above M#48).

[ixfd64]

I wonder if Curtis Cooper is going to send us those pins again...

[J.F.]

I get the impression the wait was worth it - news is catching on nicely!

Here at least, it is. In The Netherlands it is covered on a very popular news site www.nu.nl, but in category "other" rather than "tech"... hm. Also, a tech site tweakers.net spends some attention.

... on the latter one, I got a bit frustrated and carried away with the "what is the practical use of this" reactions, so I had to take a stand... arguing
* because we can (paraphrased from: "omdat het kan", because it can (be done) (simply))
* algorithmic / implementation / infrastructure improvements
* possible new Mathematical insights, should for instance New Mersenne Conjecture suddenly break or slowly become statistically stronger. The stronger such a conjecture gets, the more one wants to break it or prove it, right.

[davieddy]

Originally Posted by davieddy

Reminds of the last discovery "Another success?" (Lycorn with enough info for us to know what the expo was).

George was predictably on vacation, but Tony Reix asked us to tell him the exponent so that he could verify it. I (inter alia) duly did that.

Prematurely you spilled the beans and were duly ticked of by me and Tony.
Nice to see a bit of restraint this time around

D

Originally Posted by retina

I don't remember ever breaking the unwritten rule of premature exponent exposure. Perhaps you are thinking of someone else?

I didn't see this post before. Do you have a time machine for posting in the past?

IIRC (unlikely) it was just before the official announcement, and you gave the perfectly reasonable excuse of "everybody knows already".

As for the small print, it has got something to do with the gerbils, and is a pain in the arse when it comes to civilized conversation.

D

[aketilander]

According to the Lenstra–Pomerance–Wagstaff conjecture we are supposed to find 5.92 Mersenne primes between 10M-digits Mersennes and 100M-digits Mersennes. We have already found 4 between 10M-digits and 18M-digits and many below 18M-digits untested.

That's what I call heads!!!

... or there might be something wrong with the conjecture???