[QUOTE=jinydu]Is it possible to force the program to write a save file (as in when I'm about to turn off the computer)? I've noticed that it doesn't always write a save file when I push Ctrl-c.[/QUOTE]


The little bit that is not updated is so minor compared to the rest that I have never noticed any extra delay when restarting.

[QUOTE=Uncwilly]BTW, I am taking 3321928307 up to 72 (against my desire to do more to lower levels).[/QUOTE]

Is there some way to keep this fun for you and me too? You have fun doing lots of low tests, but I don't enjoy doing frequent web updates. Perhaps there is a second statistic we could track that benefits from less intense efforts? Maybe a ten-wide level that would require 10 candidates at 67 bits then 20 candidates at 68 bits then 30 at 69 bits? Or just total number of candidates to 69 bits? Or perhaps you'd like to take over maintaining the web page?

William

I downloaded and tried the program abit. I'm curious about this:

[QUOTE]I'm checking factors mod 120 instead of 1 or 7 mod 8[/QUOTE]

So you don't check all factors with 1 or 7 mod 8? Is the sourcecode to the program available? I'm not really a programmer but I would like to look at it and see if I understand how it works.

[QUOTE=ATH]I downloaded and tried the program abit. I'm curious about this:



So you don't check all factors with 1 or 7 mod 8? Is the sourcecode to the program available? I'm not really a programmer but I would like to look at it and see if I understand how it works.[/QUOTE]

I use mod 120 as an enhancement of the quadratic residues condition. ([url]http://www.mersenneforum.org/showthread.php?t=92&highlight=quadratic+residues[/url])

As E.W.Mayer wrote on this forum,

"Among other shortcuts, Prime95 trial factoring considers only the 16 congruence classes mod 120 (+-1, +-7, +-17, +-23, +-31, +-41, +-47, +-49) that might contain prime Mersenne factors. Thus it skips fourteen of the mod 120 congruence classes (+-9, +-15, +-25, +-33, +-39, +-55, and +-57) which satisfy +-1 mod 8 but are never prime."

I simply implemented it before checking if 2kp+1 was prime.

If, after this, you still want to peek into my sourcecode, just drop a PM with your email, but mind that some comments are still written in old good Italian :wink:

Luigi

Your list was only 7 primes short of reaching 3,321,930,000 so I tested the program on them:

M3321929929 no factor from 2^2 to 2^65.
M3321929957 has a factor: 1921749767844329 (up to 62bit)
M3321929959 has a factor: 106494430625623 (up to 62bit)
M3321929969 has a factor: 26575439753 (up to 62bit)
M3321929971 no factor from 2^2 to 2^65.
M3321929987 no factor from 2^2 to 2^65.
M3321929993 has a factor: 245822819482001 (up to 62bit)


Now I will take 3321928319 from 69 to 72 bit if its ok, it looks free.

[QUOTE=ATH]Your list was only 7 primes short of reaching 3,321,930,000 so I tested the program on them:

M3321929929 no factor from 2^2 to 2^65.
M3321929957 has a factor: 1921749767844329 (up to 62bit)
M3321929959 has a factor: 106494430625623 (up to 62bit)
M3321929969 has a factor: 26575439753 (up to 62bit)
M3321929971 no factor from 2^2 to 2^65.
M3321929987 no factor from 2^2 to 2^65.
M3321929993 has a factor: 245822819482001 (up to 62bit)


Now I will take 3321928319 from 69 to 72 bit if its ok, it looks free.[/QUOTE]

ATH, I'm afraid Uncwilly had already worked on that range by himself :unsure:

As Wblipp and Clowns789 said, it would be better to reach 72 bits for all exponents before going farther: this should keep us tidy and compact, and make updates easier.

If Uncwilly agrees, I will credit factors to both, but this will be the last time... :razz:

Luigi

Ok, np. I was really just playing with the program, testing speed on my 2 computers.

[QUOTE=ATH]Ok, np. I was really just playing with the program, testing speed on my 2 computers.[/QUOTE]

I re-read my previous post, and it may appear somewhat misleading...

3321928319 from 69 to 72 was free, and is now assigned to you.

I meant the range up to 3321929999 had already been worked.

Luigi

[QUOTE=ET_]ATH, I'm afraid Uncwilly had already worked on that range by himself :unsure:

If Uncwilly agrees, I will credit factors to both, but this will be the last time...[/QUOTE]

I have worked on the numbers up to 3321999991 (end of the range). in checking up on this I have found an error in my Results.txt massaging program that I must fix (it drops the last expo off.) :furious:

I am not worried about the credit. There is no I in team. :cool:

[QUOTE=Uncwilly]I have worked on the numbers up to 3321999991[/QUOTE]

In hopes or warning away other newcomers, I've updated the status page to reflect the existence of these results. Were these numbers taken to 62 bits? Was Will Edgington informed about the factors?

[url]http://ElevenSmooth.com/Billion.html[/url]

William

[QUOTE=Uncwilly]

I have run every thing from 3321929927 to 3321999999 upto 57 bits (in 4 stages removing exponents with found factors at each stage.)[/QUOTE]

Uncwilly, if you wish to take all these exponents to 60 (the starting range of Operation Billion Digits) I'm here to help!

Let's share the remaining exponents.

[B]William[/B], I know Uncwilly usually sends his batch of new factors to Will Edgington.

Luigi