Thread for posting tiny primes - Page 33

science_man_88 · 2010-09-07, 16:35

Quote:

Originally Posted by CRGreathouse

because I don't know what a general number is.

he still hasn't defined it well has he I'll look back through his lists to see if i can make sense of it.

CRGreathouse · 2010-09-07, 16:38

Pi, I started work on a second number as soon as I finished the first. Your number shouldn't take too long, so you'll probably get a record before I smash it.

Quote:

Originally Posted by science_man_88

he still hasn't defined it well has he I'll look back through his lists to see if i can make sense of it.

Good luck. If you find an answer, post it -- and tell me if mine is OK.

3.14159 · 2010-09-07, 16:47

Quote:

Originally Posted by Charles

Pi, I started work on a second number as soon as I finished the first. Your number shouldn't take too long, so you'll probably get a record before I smash it.

I dunno. I have to ensure PFGW does not have a primelimit to obey for trial factoring.

It is supposed to perform 71-72 billion divisions..

3.14159 · 2010-09-07, 16:54

I don't think 71 to 72 billion divisions is going to finish anytime soon.. And, no, I'm not on the applet, I am using PFGW.

CRGreathouse · 2010-09-07, 16:55

Quote:

Originally Posted by 3.14159

I dunno. I have to ensure PFGW does not have a primelimit to obey for trial factoring.

I was thinking about that, trying to compute how much work you'd have to do and what tools you'd use, since pi(sqrt(3660797218706330586200749)) = 70,239,261,003 > 2³².

Quote:

Originally Posted by 3.14159

It is supposed to perform 71-72 billion divisions.

Something like that. If that's literally true, it must only sieve to some limit, not quite up to the square root, for determining which 'primes' to use. This is safe, of course -- the extra nonprimes won't ever divide the number, just add to the running time. (Of course if testing the extra 1-2 billion numbers is shorter than doing the extra sieving to remove them, it's worthwhile.)

3.14159 · 2010-09-07, 16:57

And, in the fractions, it shoots higher than 100%. Maybe the program is being confused by the factor limit value, due to it being too large?

So far, it's performed 2.9 billion divisions out of the 70.2 billion necessary to prove its primality.

My best guess is that it is dividing by any odd number, regardless of whether or not it is prime.

Here is what I observe:

Code:

F: 3660797218706330586200749 3335000000/2734577474 mro=0

CRGreathouse · 2010-09-07, 17:06

Quote:

Originally Posted by 3.14159

I don't think 71 to 72 billion divisions is going to finish anytime soon.. And, no, I'm not on the applet, I am using PFGW.

On a 3 GHz machine with a 30 cycle (throughput) division it would take 71.5e9*30/3e9 = 715 seconds for the division alone, if your number fit into a word (which it doesn't). Even with a factor of 20 allowance for multiprecision division, that's only 4 hours. If it takes an hour for the sieving the find the primes to use, that's just 5 hours.

Now this assumes a certain amount of optimization, but I'd certainly hope it finishes in a day.

CRGreathouse · 2010-09-07, 17:08

Quote:

Originally Posted by 3.14159

My best guess is that it is dividing by any odd number, regardless of whether or not it is prime.

That would take 956 billion trial divisions.

3.14159 · 2010-09-07, 17:09

Quote:

Originally Posted by Charles

On a 3 GHz machine with a 30 cycle (throughput) division it would take 71.5e9*30/3e9 = 715 seconds for the division alone, if your number fit into a word (which it doesn't). Even with a factor of 20 allowance for multiprecision division, that's only 4 hours. If it takes an hour for the sieving the find the primes to use, that's just 5 hours.

It reaches a maximum point, then cycles over, each time with an increased trial factor bound. Is this going to continue until it reaches 70.2 billion?|

Here is what is being observed now:

Code:

F: 3660797218706330586200749 440000000/3159943140 mro=0

CRGreathouse · 2010-09-07, 17:10

Quote:

Originally Posted by 3.14159

So far, it's performed 2.9 billion divisions out of the 70.2 billion necessary to prove its primality.

Ah, much better -- 70.2 billion is spot-on (see post #357).

CRGreathouse · 2010-09-07, 17:12

So how long do you expect it to take, compared to your earlier efforts?

Mine should take about 24 hours, if I can trust the scaling. If I hit a bad breakpoint it could be as long as 48 hours.

Last time I used a single core of a slow computer. I make no guarantee that I will do either this time.

2010-09-07, 16:54	#356
3.14159 May 2010 Prime hunting commission. 3220₈ Posts	I don't think 71 to 72 billion divisions is going to finish anytime soon.. And, no, I'm not on the applet, I am using PFGW. Last fiddled with by 3.14159 on 2010-09-07 at 16:54

2010-09-07, 16:57	#358
3.14159 May 2010 Prime hunting commission. 1680₁₀ Posts	And, in the fractions, it shoots higher than 100%. Maybe the program is being confused by the factor limit value, due to it being too large? So far, it's performed 2.9 billion divisions out of the 70.2 billion necessary to prove its primality. My best guess is that it is dividing by any odd number, regardless of whether or not it is prime. Here is what I observe: Code: F: 3660797218706330586200749 3335000000/2734577474 mro=0 Last fiddled with by 3.14159 on 2010-09-07 at 17:02

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Prime posting thread, part 2. (With a catch.)	3.14159	Miscellaneous Math	55	2010-11-19 23:55
Tiny range request .... 555.1M	petrw1	LMH > 100M	1	2010-07-13 15:35
Other primes thread	nuggetprime	No Prime Left Behind	32	2009-10-21 21:48
Error: tiny factoring failed	10metreh	Msieve	26	2009-03-08 23:28
Tiny error on nfsnet pages.	antiroach	NFSNET Discussion	1	2003-07-08 00:27

2010-09-07, 17:12	#363
CRGreathouse Aug 2006 1011101011011₂ Posts	So how long do you expect it to take, compared to your earlier efforts? Mine should take about 24 hours, if I can trust the scaling. If I hit a bad breakpoint it could be as long as 48 hours. Last time I used a single core of a slow computer. I make no guarantee that I will do either this time.