mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Conjectures 'R Us (https://www.mersenneforum.org/forumdisplay.php?f=81)
-   -   Report top-5000 primes here (https://www.mersenneforum.org/showthread.php?t=9782)

Siemelink 2008-12-15 19:29

[QUOTE=mdettweiler;153392]Well, LLRnet doesn't work with multiple different bases/types at once, so we'd still have to use separate servers for each base. Probably the best way would be to hold team drives on each base in turn, and then when one is proven, move on to the next.[/QUOTE]

And how many drives are open at the moment?

Willem.

mdettweiler 2008-12-15 21:17

[quote=Siemelink;153485]And how many drives are open at the moment?

Willem.[/quote]
Currently 3 "regular" drives (for Riesel base 16, Sierp. base 16, and Sierp. base 6, respectively), and two "mini" drives (Sierp. base 3 and Riesel base 3).

P.S.: We've currently got an LLRnet server running for Drive 3 (Sierp. base 6), at crus.ironbits.net port 6. Recently that server has been completely idle; we could always use some help there. :smile:

MyDogBuster 2008-12-23 12:10

Riesel Base 45

1264*45^64666-1 is prime

Primality testing 1264*45^64666-1 [N+1, Brillhart-Lehmer-Selfridge]
Running N+1 test using discriminant 7, base 1+sqrt(7)
1264*45^64666-1 is prime! (3830.0576s+0.0099s)

106910 digits and enters Top5000 at 4686 (just barely)

gd_barnes 2008-12-24 04:19

[quote=MyDogBuster;154713]Riesel Base 45

1264*45^64666-1 is prime

Primality testing 1264*45^64666-1 [N+1, Brillhart-Lehmer-Selfridge]
Running N+1 test using discriminant 7, base 1+sqrt(7)
1264*45^64666-1 is prime! (3830.0576s+0.0099s)

106910 digits and enters Top5000 at 4686 (just barely)[/quote]


Unbelievable! OK, how'd you do that? You're starting to scare me a little. Have you solved the mystery of where primes will be? lol

Well...actually, I was wondering how you were searching base 45. Unless you put a whole bunch of firepower on it, you couldn't be at n=65K on all of the k's yet starting from just n=10K.

Anyway, congrats on something that I don't have yet...a non-power-of-2 top-5000 prime! :smile:


Gary

gd_barnes 2008-12-24 04:56

[quote=MyDogBuster;154713]Riesel Base 45

1264*45^64666-1 is prime

Primality testing 1264*45^64666-1 [N+1, Brillhart-Lehmer-Selfridge]
Running N+1 test using discriminant 7, base 1+sqrt(7)
1264*45^64666-1 is prime! (3830.0576s+0.0099s)

106910 digits and enters Top5000 at 4686 (just barely)[/quote]


A couple of things on your prover code:

1. LLR cannot prove a non-power-of-2 base prime. If you tested it using LLR, then, yes, you should have LLR in the prover code but you also need to add "OpenPFGW". (I'm assuming you used PFGW to prove it.)

2. I thought you were going to search with Phrot since it is much faster. If you did, then LLR would need to be changed to "3Ps" in your prover code. Once again, it cannot prove a non-power-of-2 base prime so like in #1, you'd need to add "OpenPFGW".

For an example where Max found a base 23 top-5000 prime using Phrot, see his prover code p238.

If this was an isolated prime that you just happened to find, I'd say don't worry about it. But since you'll likely find several top-5000 primes for all of the base 45 k's (or base 6 or some other base), I'd suggest Emailing Prof. Caldwell and having him make one or both of the changes above.

If you searched using Phrot, my guess is that he will ask you to create a new prover code and then will move your prime to it since LLR prover codes start with an "L" and Phrot codes start apparently start with a "p". Eventually your code L669 would be deleted by their system after it doesn't have any primes for a period of time. Alternatively, if you searched with LLR, then I'm sure he'd just add OpenPFGW to L669.


Gary

mdettweiler 2008-12-24 05:17

[quote=gd_barnes;154831]A couple of things on your prover code:

1. LLR cannot prove a non-power-of-2 base prime. If you tested it using LLR, then, yes, you should have LLR in the prover code but you also need to add "OpenPFGW". (I'm assuming you used PFGW to prove it.)

2. I thought you were going to search with Phrot since it is much faster. If you did, then LLR would need to be changed to "3Ps" in your prover code. Once again, it cannot prove a non-power-of-2 base prime so like in #1, you'd need to add "OpenPFGW".

For an example where Max found a base 23 top-5000 prime using Phrot, see his prover code p238.

If this was an isolated prime that you just happened to find, I'd say don't worry about it. But since you'll likely find several top-5000 primes for all of the base 45 k's (or base 6 or some other base), I'd suggest Emailing Prof. Caldwell and having him make one or both of the changes above.

If you searched using Phrot, my guess is that he will ask you to create a new prover code and then will move your prime to it since LLR prover codes start with an "L" and Phrot codes start apparently start with a "p". Eventually your code L669 would be deleted by their system after it doesn't have any primes for a period of time. Alternatively, if you searched with LLR, then I'm sure he'd just add OpenPFGW to L669.


Gary[/quote]
Actually, the whole thing with the codes starting with "L" or "p" doesn't have anything to do with LLR vs. Phrot; instead, it has to do entirely with the program that *proved* the prime. That's why you have to select the proof program separately from the "additional credits" when creating a prover code. For example: if you select LLR as the proof program (i.e. for base 2 stuff or power-of-2 bases), the code will begin with L. However, if you select OpenPFGW as the proof program, the code will begin with "p".

Max :smile:

MyDogBuster 2008-12-24 07:05

[quote]Unbelievable! OK, how'd you do that? You're starting to scare me a little. Have you solved the mystery of where primes will be? lol

Well...actually, I was wondering how you were searching base 45. Unless you put a whole bunch of firepower on it, you couldn't be at n=65K on all of the k's yet starting from just n=10K.
[/quote]

I'm really using ESP. LMAO.

I'm searching individual k's to 100K. I sieved all the k's from n=10K-500K up to 10T max. Now I'm using the individual k files to phrot to 100K. I have 4 cores banging away. I didn't like the idea of doing all the k's at once to 100K.

[quote]2. I thought you were going to search with Phrot since it is much faster. If you did, then LLR would need to be changed to "3Ps" in your prover code. Once again, it cannot prove a non-power-of-2 base prime so like in #1, you'd need to add "OpenPFGW".

[/quote]

I'll get the prover code fixed. I couldn't find phrot anywhere in the list.
At least I got the srsieve, CRUS and my name right. 3 outta 4 ain't bad for an older guy. lol

mdettweiler 2008-12-24 07:17

[quote=MyDogBuster;154841]I'm really using ESP. LMAO.

I'm searching individual k's to 100K. I sieved all the k's from n=10K-500K up to 10T max. Now I'm using the individual k files to phrot to 100K. I have 4 cores banging away. I didn't like the idea of doing all the k's at once to 100K. [/quote]
Yeah, I generally like to do the same thing with non-team drive stuff here at CRUS--that way I can make use of Phrot's -s command line flag, which tells it to stop searching its input file as soon as it finds a PRP. That way, I can queue up additional work with a shell script (actually, I do it all on the command line, but it's essentially like a shell script, I won't go into it all here), and thus ensure that I never have an idle or wasted moment. :grin:
[quote]I'll get the prover code fixed. I couldn't find phrot anywhere in the list.
At least I got the srsieve, CRUS and my name right. 3 outta 4 ain't bad for an older guy. lol[/quote]
The reason why you couldn't find Phrot in the list is because it's not the proof program, just a PRP-finding program; PFGW (listed as OpenPFGW on the Prime Pages website) would be the correct choice in this case since you're using that to prove the primes after finding them PRP with Phrot. Instead, you enter Phrot in the "additional credits" farther down the page, like you do with sieving programs. Essentially, the create-prover-code page should look like this:

Proof software: OpenPFGW
Additional credits: CRUS, Srsieve, 3Ps

(Note: 3Ps is the code for Phrot. The reason why it's called this instead of just Phrot is because Phil, the guy who made Phrot, figured it would be simpler to combine it into one prover code along with two other lesser-known prime-finding programs that he's made. I think it stands for "Phil's Prime Pack" or something like that. :smile:)

MyDogBuster 2008-12-24 07:23

[QUOTE]The reason why you couldn't find Phrot in the list is because it's not the proof program, just a PRP-finding program; PFGW (listed as OpenPFGW on the Prime Pages website) would be the correct choice in this case since you're using that to prove the primes after finding them PRP with Phrot. Instead, you enter Phrot in the "additional credits" farther down the page, like you do with sieving programs. Essentially, the create-prover-code page should look like this:

Proof software: OpenPFGW
Additional credits: CRUS, Srsieve, 3Ps

[/QUOTE]

I've emailed Dr. Caldwell to see about getting it fixed. I guess next time I should ask first.

mdettweiler 2008-12-24 07:25

[quote=MyDogBuster;154844]I've emailed Dr. Caldwell to see about getting it fixed. I guess next time I should ask first.[/quote]
Okay, glad to hear that it's getting fixed. And, don't feel bad about it--messing up in some way when reporting primes is a very, very common mistake and has probably happened to at least 75% of us here at some point. :smile:

MyDogBuster 2008-12-24 07:31

[QUOTE] but it's essentially like a shell script, I won't go into it all here), and thus ensure that I never have an idle or wasted moment. :grin:
[/QUOTE]

I'd like to see you shell script if I could. I did use the -s option but haven't quite figured out how to que up work behind it yet.


All times are UTC. The time now is 09:07.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.