• Quick note - the problem with Youtube videos not embedding on the forum appears to have been fixed, thanks to ZiprHead. If you do still see problems let me know.

Largest Prime Number Discovered

I don't know how these prime finding algorithms work. Has every prime number up to this new one been discovered and now there are only larger primes left?

I doubt it, this project is only looking for Mersenne primes so would miss any other primes between this and the last highest known prime.

There's a pretty big gap between these, but I'm not sure how many primes would be expected to be in that gap, if any.
 
I doubt it, this project is only looking for Mersenne primes so would miss any other primes between this and the last highest known prime.

There's a pretty big gap between these, but I'm not sure how many primes would be expected to be in that gap, if any.

http://en.wikipedia.org/wiki/Prime_number_theorem

or: why you need to study complex analysis to count prime numbers. :)
 
Primes have very direct application in music theory -- tuning theory. But anything much over 23, you're probably kidding yourself. Each small prime has a characteristic sound.

Primes of all sizes also can be used to make special self-similar sequences, but again, only the smaller ones are certain to be useful in music theory.

And -- here I'm really getting out of my territory -- they can be used in experimental designs.

Costas sequences, generated from primes, are used to make optimal radar signals.

Sounds fascinating - I'd love to know more.
What units are we talking here - semitones, Hertz...?

http://en.wikipedia.org/wiki/Just_intonation

The primes occur in frequency ratios, which are usually written like fractions. So we're talking ratios of hertz (cps) which are typically converted back and forth into cents.

Because we're used to 12 pitches per octave, intervals which differ from those intervals have a novel sound. As you go up the series of primes -- 5,7,11,13 -- you get more and more out of tune with 12-tone. 12-tone (or 12EDO) tuning has "fifths" (G-C) at 700 cents, which are around 2 cents narrow if compared to the frequency ratio of 3:2, which is 701.955 cents.

http://www.sengpielaudio.com/calculator-centsratio.htm

So 12EDO is a very good tuning for ratios involving primes no higher than 3 (or "3-limit" tuning). (Note that 9:8 (or approximately D-C) is still within the 3-limit, because 9 is 3x3, and 8 is 2x2x2.)

The next set of EDO cycles that approximate a 3:2 are 17 and 19 tones per octave. These tunings have their own flavors, but, unfortunately, don't do everything that 12-tone can do, so there's a trade-off. For example, 17EDO doesn't have a "minor second" or something approximating a 16/15 that is in the vicinity of 100 or 100+ cents.

http://xenharmonic.wikispaces.com/17edo

There's always a trade-off between practicality and novelty or richness.

Because of advances in technology, it's much easier to explore new tunings than ever before. There's a fairly large community of people making music with non-standard tunings. However, advocates of alternate tunings exaggerate when they sometimes claim that making music in non-standard tunings is as easy and practical as making music in good old 12EDO.

When I say that each prime has its own flavor, I mean that a frequency ratio like 7:2 (2168.826 cents) has a "harmonic" sound, but is around 31 cents narrow compared to 12EDO, 11:8 is around 49 cents narrow, and so on. In other words, each new prime-generated ratio deviates from 12EDO by an approximately characteristic amount.
 
Last edited:
Where did they excavate it ?
Intersting question. There can be no doubt that Pompeii existed before it was excavated, but in what sense does such a monstrous number exist before it is defined. What are such numbers, numbers "of"?
 
I'm pretty sure none of the applications of primes mentioned so far would need a 17 million digit prime. :)

It might be more accurate to say that none of them need such a big prime yet. The trick with most cryptography is that data is not actually completely secure, it would just take an infeasible length of time to break the encryption by brute force. But the problem with that is that computers keep getting faster. At the moment, a 17 million digit prime number is completely pointless since even the best computers would take millions of years to break encryption using much shorter primes. However, modern computers can easily break encryption that a computer from 30 years ago would have taken millions of years to break.

To give some perspective, in 1990 the fastest supercomputer in the world was the Cray-2 which had a maximum computing power of 1.9 gigaflops - 1.9 billion calculations per second. In 2013, a single graphics card in a home PC can be capable of just under 4 teraflops, while the fastest supercomputer has been measured at over 27 petaflops - 14 million times faster. In other words, something that would have taken a million years to break in 1990 could today be broken in under a month*.

For another comparison, this prime would take up about 22MB of memory. In 1990, you'd struggle to fit that in a home PC at all. Today that's the size of a single photo, and you can happily install games and videos thousands of times bigger while barely noticing any different to how full your hard drive is.

So sure, this prime may seem ridiculously big for now, but give it a few decades and it's going to start looking a lot smaller.


* Based on assuming similar improvements in general computing power; obviously these supercomputers aren't actually used for that, and don't run at peak speed continuously.
 
I don't know how these prime finding algorithms work. Has every prime number up to this new one been discovered and now there are only larger primes left?

No, there are a mindbogglingly huge number of prime numbers between that one and the second largest prime that we know.

There's a weird theorem stating that the number of prime numbers that are smaller than a number n is approximately n/ln n where ln is the natural logarithm of the number.

The new prime has over 17 million decimal digits, which means that its natural logarithm is around 25 million or so. Dividing a number of a bit over 17 million digits by 25 million gives a number with a slightly smaller bit over 17 million digits.

That's a huge enough number that if you wanted to store all those prime numbers and you could use all material in the universe as your storage space, you would need to some serious cramming there. The number of prime numbers that you would need to store in every atom would still have over 17 million digits.
 
Last edited:
I don't know how these prime finding algorithms work. Has every prime number up to this new one been discovered and now there are only larger primes left?
For Mersenne numbers M(p) = 2p - 1 for p a prime, one can take some shortcuts, like the Lucas–Lehmer primality test.

The longest continuous sequence of primes found has been up to about 1018, for testing the Goldbach conjecture (List of prime numbersWP).
 
The actual prime found may not be very useful. But as usual when investigating, the side effects can be more useful than the original aim.

The Great Internet Mersenne Prime Search was one of the first distributed computing projects. A lot was learned about how to initialise them, rewarding contributors, breaking up problems to run on lower speed machines and generally controlling the whole process. Other systems have now superseded GIMPS but it is still cool to contribute directly to it rather than the nebulous procedure of paying taxes and seeing most of your money being wasted before it reaches the investigators.

For those interested, JREF have a team that contributes to a similar project, Folding@home.
http://www.internationalskeptics.com/forums/showthread.php?t=154&highlight=folding+home&page=46
 
Last edited:
Anyway it's all nonsense. This isn't the biggest prime.
Why , I know of one twice as big!
 

Back
Top Bottom