What should I do with this information? (New Method of Generating Primes)

[DarkDarkly]

So for the past couple months I have been trying to find a NEW method of generating prime numbers. From what I could find for existing methods online, i succeeded, because this didn't come up anywhere I could see. I currently do not know if the method is faster or slower than current generators (Though my assumption is slower), but the major difference is that it does NOT need to compare each new value to all previous values in the sequence to check if it is prime. I have some methods that I plan to test that will hopefully speed it up a lot, but for now just know that it took roughly 30 seconds or less to generate all primes less than 20000 on my laptop. (and that also includes the extra time it took to run another calculation verifying that each value was in fact prime, so it was really generating all the primes two times. One with mine, and one with a second method that is already known, but allows me to compare to my list.)

I don't want to share too much related to my final solution of obtaining this until I have a general idea of what I should do with it. Is there some organization I should contact? Does this really matter if this is just shared publicly here, even if it's a newly discovered method? (and does the answer change knowing I'm 17 ._. ?) I do plan to use it for an Internal Assessment I have to write for my math class, but I want it to do something else with it too. It may not be super important (or maybe it is, who knows), but I want the information to reach the right person first.

Thanks!

 

[Steven G]

Very interesting. My gut feeling is that your method is already known or it was not looked into then it was foreseen to be a slow method. Having said that, who knows maybe you have a great method to find primes.

What should you do with it? First finish it up then go to your local university and have someone in the math or comp sci look at it. I doubt that they will steal your work. If that worries you, then go straight to publication in a math journal. After publishing the paper at 17! please come back and post your method here.

If you get the paper published make sure you point that out in all your college applications

Good luck!

 

[DarkDarkly]

Very interesting. My gut feeling is that your method is already known or it was not looked into then it was foreseen to be a slow method. Having said that, who knows maybe you have a great method to find primes.

What should you do with it? First finish it up then go to your local university and have someone in the math or comp sci look at it. I doubt that they will steal your work. If that worries you, then go straight to publication in a math journal. After publishing the paper at 17! please come back and post your method here.

If you get the paper published make sure you point that out in all your college applications

Good luck!

Thanks for the information! I do have to write a paper on it for class, so with a bit of fine-tuning and guidance from my teacher I could probably have something worthy of actually publishing. It is definitely different than a prime sieve, and doesn't involve comparing a new value to old ones in the list, and doesn't use any method I've seen for proving a number is prime, so I'm definitely hoping it's new. I'll be back to show what I have, since it is definitely an unconventional way of making them, but it works and has a valid reason to within the math from which I derived the program I made to test it.

 

[joshuaa]

you said that you run your own method with a second method that is already known in 30 s

does this mean the second method is as fast as your method?

Or for example, you run your own method in 5 seconds, and the other took 25 s, so total 30 s?

if the latter is true. i would offer you a million bucks (just kidding)

but it is worth (you're a genius)

 

[DarkDarkly]

you said that you run your own method with a second method that is already known in 30 s

does this mean the second method is as fast as your method?

Or for example, you run your own method in 5 seconds, and the other took 25 s, so total 30 s?

if the latter is true. i would offer you a million bucks (just kidding)

but it is worth (you're a genius)

I would estimate that mine is a little bit slower than a prime sieve in it's current state, mostly because of how unoptimized it is. If i rework my code a little bit I can probably make it equally as fast, if not a little faster. The main issue now is that I do have to make a minor change to my equations, since I learned after last night that one of my steps caused it to essentially be a rewritten prime sieve. Removing the specific step that does this causes my generator to only be 90% accurate or so, meaning only 90% of the values will be prime (however, all primes should still appear)

Having said that, as long as I can make it work again without the part I mentioned, it still should be entirely unique as I originally hoped. It would definitely be different than a prime sieve, but i couldn't say if it'd be faster.

 

[Steven G]

You're amazing! Keep in mind that almost all famous mathematicians become famous a few years after 17 so you are running out of time!

 

[DarkDarkly]

You're amazing! Keep in mind that almost all famous mathematicians become famous a few years after 17 so you are running out of time!

Thanks! I just came back to mention that I figured out an alternate solution that turned out to be 100 times easier than I was expecting. The program does function much like a prime sieve after thinking about it, but the math behind it is represented as if I was expressing the prime as a sum of two numbers that share no factors. Now, that isn't the ONLY thing that will work to create a prime, as realistically all odd numbers can be written that way with "n + (n-1)", but I have some extra steps involved, as well as a repeating shape that I remove some values from in said shape. This finally resulted in me being able to generate all primes between 1 and 64001 in about 8 seconds or so. The software I'm using limits me to that many, but if I rewrote it in another language I could probably fix that.