[QUOTE=Batalov;514943]<snip>
Please keep this thread on the subject of "[U]the formula[/U]". There are separate threads to discuss hardware - I foresee that there will be two dozen of hardware related posts here in a span of a couple days; and they will be detracting from the progress of "[U]the formula[/U]". Please.[/QUOTE] Formula? What formula? First, the OP said [url=https://www.mersenneforum.org/showpost.php?p=514568&postcount=1]here[/url] that [quote]actually the formula does not guarantees mersenne prime as it has two variables. it returned 1/3 of proven composite mersenne numbers or tested numbers.[/quote]

Then, [url=https://www.mersenneforum.org/showpost.php?p=514573&postcount=4]here[/url] that[quote]the formula produced result as follows, composite = composite number, either the expooents is not prime, or it is divisible by something.



m11-m20
composite
composite
m21-25
composite
composite
m26-m29
composite
m30-m32
composite
m33 (misses m34)
composite
m35
composite x 2
m36
composite x 3
m37
composite
m38
composite
m39-m41
composite
m42
m43
composite
m44
composite x 4
m45
composite x 2
m46
composite x 2
m47
composite x 3
m48
composite
m49
m50
composite x 8
m51[/quote]
Then [url=https://www.mersenneforum.org/showpost.php?p=514574&postcount=5]here[/url] the OP said,[quote]the formula is a recurrence relation on two arbirtary sequences and values are taken on this sequence every 5 terms


an=f(bn x g(cn)) where bn,cn are sequences and f,g are functions acting on these sequences.


a1,a2,a3,a4,a5,a6,...
then my mersenne prime generating sequence would be something like
a2,a7,a12,a17,...[/quote]
"Two arbirtary [[i]sic[/i]] sequences" means two sequences lacking definition. The formula with two functions f and g is [i]not[/i] a recurrence relation. The only thing that appears to be a "variable" is n. If neither sequence b[sub]n[/sub] or c[sub]n[/sub] is defined in terms of n, there are two more "variables." If neither function f or g is well-defined, that's yet two more. In any case, there is no apparent way to name exactly [i]two[/i] variables.

It is also not clear what the terms of the output sequence a[sub]n[/sub] are. Are the a[sub]n[/sub] numbers of the form 2[sup]p[/sup] - 1? The exponents?

And if the OP is only taking every fifth term a[sub]5k-3[/sub] anyway, why not go back and redefine the sequences b[sub]n[/sub] and c[sub]n[/sub] as every fifth term b[sub]5k-3[/sub] and c[sub]5k-3[/sub]?

In short, >>plonk<<

[QUOTE=samuel;514980]
i have i7 too it says Intel Core i7-620M @ 2.66GHz 16GB DDR3 800

why is mine so much slower its not fair you are hiding something from me[/QUOTE]


Jeez, it's an antique. No AVX instructions. I haven't used memory that slow in 15 years.

My result is due on May 4th.

Oh heck, I'll just run them quickly
 

Usually I'm a patient person, buy not in this case. That's why I have these toys.
I just started the 2 smallest exponents; one on a Radeon Vii and one on my Titan V.
Expect results Monday afternoon.

I apologize if I stepped on anyone's toes who was planning to run these.

[QUOTE=tServo;515010]Usually I'm a patient person, buy not in this case. That's why I have these toys.
I just started the 2 smallest exponents; one on a Radeon Vii and one on my Titan V.
Expect results Monday afternoon.

I apologize if I stepped on anyone's toes who was planning to run these.[/QUOTE]

On post 52 George offered
73 dcheuk started

[QUOTE=samuel;514980]so when you starting im anxious to see if i am right or the world is right =))))))[/QUOTE]

Just started, 3 days we shall see if your claim is legitimate. If it turns out to be composite just don't blame it on my hardware claiming they're faulty.

[CODE]Starting M88680457 fft length = 5184K
| Date Time | Test Num Iter Residue | FFT Error ms/It Time | ETA Done |
| Apr 28 09:58:00 | M88680457 5000 0x56bab93a2c17a044 | 5184K 0.03320 2.9901 14.95s | 3:01:39:07 0.00% |
[/CODE]

[QUOTE=samuel;514980]i havent encountered any problems in math that i cant solve[/QUOTE]

Pick one and solve :grin:
[url]https://en.wikipedia.org/wiki/List_of_unsolved_problems_in_mathematics[/url]

[QUOTE=tServo;515010]Usually I'm a patient person, buy not in this case. That's why I have these toys.
I just started the 2 smallest exponents; one on a Radeon Vii and one on my Titan V.
Expect results Monday afternoon.

I apologize if I stepped on anyone's toes who was planning to run these.[/QUOTE]

I just saw your post, it's fine, mine will be a dc then no worries.
That way if he claims your result is bogus then we can counterclaim with a matching residues.

[QUOTE=samuel;514573]the formula produced result as follows, composite = composite number, either the expooents is not prime, or it is divisible by something.



m11-m20
composite
composite
m21-25
composite
composite
m26-m29
composite
m30-m32
composite
m33 (misses m34)
composite
m35
composite x 2
m36
composite x 3
m37
composite
m38
composite
m39-m41
composite
m42
m43
composite
m44
composite x 4
m45
composite x 2
m46
composite x 2
m47
composite x 3
m48
composite
m49
m50
composite x 8
m51
...

there are 4 obvious composite numbers here 3 of them already factored[/QUOTE]

samuel, I have some exercises for you while you are waiting for your results.

1. If you view each of the above as a (Bernoulli) trial, with "composite" as a failure and a Mersenne prime as a success, what is your crude success rate so far? You may wish to compute two figures, one including and one excluding the "obvious" composite numbers.
2. Suppose the figures from #1 are representative of your future odds (and not merely overfit). What is the probability that two further numbers tested (like those tested by Prime95) would both result in Mersenne primes? That neither would? That one would, but not the other?
3. Suppose two numbers were tested and neither was a Mersenne prime. Given your analysis in #2, how confident would be in your method? Suppose three were tested instead without finding a Mersenne prime. (You may need to do some calculations like in #2.) Suppose ten, or a hundred were tested. How confident would you be?

As these are very simple, high-school probability problems I'm sure you'll have no issues with them. You should complete them before Prime95 releases his results.

[QUOTE=dcheuk;515012]Pick one and solve :grin:
[url]https://en.wikipedia.org/wiki/List_of_unsolved_problems_in_mathematics[/url][/QUOTE]

Pillai's conjecture is nice. I'd also love to see someone pick off Landau's fourth problem: are there infinitely many primes of the form n^2 + 1?

[QUOTE=dcheuk;515012]Just started, 3 days we shall see if your claim is legitimate. If it turns out to be composite just don't blame it on my hardware claiming they're faulty.

I just saw your post, it's fine, mine will be a dc then no worries.
That way if he claims your result is bogus then we can counterclaim with a matching residues.[/QUOTE]

Sounds good. I am not poaching so you and George should continue to completion and take the Primenet credits. I will not report my results to PrimeNet, just post them here.
I'll only report thumbs or down and pm the residues.

[QUOTE=CRGreathouse;515020]samuel, I have some exercises for you while you are waiting for your results.

1. If you view each of the above as a (Bernoulli) trial, with "composite" as a failure and a Mersenne prime as a success, what is your crude success rate so far? You may wish to compute two figures, one including and one excluding the "obvious" composite numbers.
2. Suppose the figures from #1 are representative of your future odds (and not merely overfit). What is the probability that two further numbers tested (like those tested by Prime95) would both result in Mersenne primes? That neither would? That one would, but not the other?
3. Suppose two numbers were tested and neither was a Mersenne prime. Given your analysis in #2, how confident would be in your method? Suppose three were tested instead without finding a Mersenne prime. (You may need to do some calculations like in #2.) Suppose ten, or a hundred were tested. How confident would you be?

As these are very simple, high-school probability problems I'm sure you'll have no issues with them. You should complete them before Prime95 releases his results.[/QUOTE]


do you people have to put me down and label me as someone that knows nothing about math? i shall prove it to u once and for all i can do math


1, i count 40 out of 76 are prime, so 40/76 SUCCESS and 30/76 FAILURE


2, the prob that the next 2 being tested is both prime is (40/76)^2
the prob neither would is (36/76)^2.
the prob one is prime and the other is not, is 1 minus both of those numbers


3, then i screwed up yes i get it but can we get to that point after the results are go. stop judging me base on probabilties


anyways the prob of x consecutive failures based on 40/76 success rate is (36/76)^x.

[QUOTE=samuel;515025]2, the prob that the next 2 being tested is both prime is (40/76)^2
the prob neither would is 1- (40/76)^2.
the prob one is prime and the other is not, is 1 minus both of those numbers[/QUOTE]Cool. Let's do the numbers.

"the prob that the next 2 being tested is both prime is (40/76)^2" = [b]0.27700831...[/b]

"the prob neither would is 1- (40/76)^2" = [b]0.72299169...[/b]

"the prob one is prime and the other is not, is 1 minus both of those numbers" = 1 - (0.27700831... + 0.72299169...) = [b]zero.[/b]

So you are saying that there is no chance only one of them is prime. It can only be either none or both.

[QUOTE=retina;515027]Cool. Let's do the numbers.

"the prob that the next 2 being tested is both prime is (40/76)^2" = [B]0.27700831...[/B]

"the prob neither would is 1- (40/76)^2" = [B]0.72299169...[/B]

"the prob one is prime and the other is not, is 1 minus both of those numbers" = 1 - (0.27700831... + 0.72299169...) = [B]zero.[/B]

So you are saying that there is no chance only one of them is prime. It can only be either none or both.[/QUOTE]


i realized the mistake so i corrected it, if you look at timestamp i did that before you post so you can't call me stupid. =)))




[QUOTE=Uncwilly;514758]
blah

How about this qualifying exam from Harvard?
[URL]http://www.math.harvard.edu/graduate/quals/qf18.pdf[/URL]
blahblah blah[/QUOTE]



i was trying to understand the first problem what the heck is lesbesgian nonmeasurable, the wikipedia article is making terminally no sense, is this math for real?