mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Riesel Prime Search (https://www.mersenneforum.org/forumdisplay.php?f=59)
-   -   Welcome to "Riesel Prime Search" (https://www.mersenneforum.org/showthread.php?t=4917)

Kosmaj 2005-10-31 14:31

Welcome to "Riesel Prime Search"
 
The new project was established less than 30 minutes ago and we are now ready to go! We already have two pending primes. :cool: I'm sure more will come soon!

When you submit your primes please be sure to create a new prover's code and to include "RPS", the user-ID of the project, in the field next to the sieving program you used.

The first team effort [URL=http://www.mersenneforum.org/showthread.php?t=4918]has already started[/URL]. :geek:

Finally, thanks to everybody who decided to join us and who helped us decide the name of the project quickly!

The project descriptive data (to be included in the Top-5000) page:

[QUOTE]The project is searching for Riesel primes k*2^n-1, k>1.
We have three subprojects as follows.
(1) k's that can produce many primes in the given range of exponent n,
(2) low-weight k's that produce a very small number of primes
[opposite to (1) above],
(3) small k < 300 (with exception of those k's already reserved by
other projects).[/QUOTE]
If you have any comments on this please let me know.

Flatlander 2005-10-31 14:53

[QUOTE=Citrix]Please start a new sub-forum, instead of having new threads in this sub-forum. Talk to the admins for this. This will help keep things seperate in terms of which project is doing what. (What shane's 15K is doing and what your rieselsearch is doing)

Citrix

Also I vote for rieselsearch as the name.[/QUOTE]
So will there be a new sub-forum?

Kosmaj 2005-10-31 14:59

There will be soon a new forum in the "Other Projects" section.

fatphil 2005-10-31 15:02

[QUOTE=Kosmaj]If you have any comments on this please let me know.[/QUOTE]

How is an outsider supposed to know which k are '15k', and which k are 'RPS'?

My back-filling of small primes for dense k's (be they 15*k for some k or not, but of course all the dense ones are) is done purely for the completeness of the record, and it would be nice to know how this split of ownership/responsibility will take place? Potentially, the tables at 15k.org look like the best record, but if they're not updated by dint of the k values being now RPS owned, then confusion will ensue -- at least for me. How is maintainance of those tables now going to be handled -- to whom should I report factors?

OmbooHankvald 2005-10-31 15:04

[QUOTE=Kosmaj](2) [b]low-wight[/b] k's that produce a very small number of primes
[opposite to (1) above],[/QUOTE]

You should fix this typo before submitting the text ;)

Kosmaj 2005-10-31 15:14

[B]Phil[/B], we are now in the transition phase, please be patient for a while.

[B]Omboo[/B], thanks, it's fixed now.

Flatlander 2005-10-31 17:03

Should new prover codes include 'NewPGen'?

Citrix 2005-10-31 18:00

If you guys have decided on using this sub forum please rename it to rps than 15k search.

lsoule 2005-10-31 18:52

There will be a new RPS sub-forum soon (off of "Other Projects")

gribozavr 2005-10-31 20:56

What should old 15k searchers do with their current reservations? I'm testing a low-weight k=177941 n=1 to 1M, now n=710722. What project does this k belong to? What k's belong to the other (with forum on yahoo) project?

Kosmaj 2005-10-31 21:21

1) All low-weight and all k<300 stay with RPS.

2) If you already have a reserved 15k multiplier and find a prime it depends on you do you wish to report it using an RPS or 15k prover's code. In any case it "belongs" to you until you stop working on it.

3) Team work on k=2995125705, k=210885, and k=355424355 all stay with RPS but note that the last one has no new ranges to reserve.

4) If you want to reserve a new 15k multiplier from the "Robinson weight" page there is currently no sharing system with 15k in place. If you just want one or two of them in a coming week or so you can reserve them I think, then (2) above applies. Something general will be worked out soon.

BotXXX 2005-11-02 10:32

New forum?
 
Shouldn't the posts concerning RPS be moved to the newly created forum? [url]http://mersenneforum.org/forumdisplay.php?f=59[/url]

edorajh 2006-03-17 10:56

[quote=Flatlander]Should new prover codes include 'NewPGen'?[/quote]

I can't find answer to this question. Should we include NewPGen in prover code even if we find prime in Team Drive? Not that I find one (but I hope I will :smile: ), I'm just curious.

Kosmaj 2012-08-17 09:47

I'll try to collect answers to some frequent questions, a kind of FAQ, and the first one is:

Q: What does the first line of the LLR input (NewPGen output) file means?
A:[URL="www.mersenneforum.org/showthread.php?t=4903"] this thread[/URL]

Kosmaj 2012-08-17 10:05

Q: Can I sieve myself?
A: Yes, for k*2^n-1 and k*2^n+1, fixed k, the fastest sieving program to use is [URL="http://primes.utm.edu/bios/page.php?id=905"]srsieve[/URL].
Use sr1sieve to sieve a single k, or sr2sieve to sieve multiple k's at once.

However, for k<10000 (=10^4), on both plus and minus side, Psieve already sieved deep enough all n<3M (3E6). Larger exponents up to about n=6M are still being sieved but the latest output files are already well sieved and can be used for prime search.

In short, for k<10000 use the Psieve files and give credit to them, for k>10000 sieve yourself using srsieve.

shanecruise 2012-09-13 05:08

i'm trying to run srseive on linux
can you help me with a sample input

Thomas11 2012-09-13 09:32

[QUOTE=shanecruise;311401]i'm trying to run srseive on linux
can you help me with a sample input[/QUOTE]

Let's assume that you want to sieve k=12345 for nmin=1M and nmax=1.2M:
[CODE]./srsieve 12345*2^n-1 -n 1000000 -N 1200000 -f
[/CODE]

The option "-f" means that the factors are written to a logfile (srfactors.txt). You may omit this.
Of course, you may specify more than one sequence (multiple k).
pmin and pmax can be specified by "-p" and "-P". If not given, the sieve will start at p=2 and will run until manually stopped (by pressing Ctrl-C).

The output will be written to the file "srsieve.out".
This file may be converted by using the tool called "srfile":
[CODE]./srfile -g srsieve.out[/CODE]

This will create individual files for each k in NewPGen format. The option "-G" will create a combined file ("PRP" format). The option "-a" converts to "abcd" format. You will need this file format when you want to use sr2sieve.

Hope this helps...

VBCurtis 2012-09-14 03:33

I believe srsieve accepts the srfile flags for file type directly. So, you can end your srsieve command with a -g or -G to produce LLR-readable output directly.
-Curtis

shanecruise 2012-09-14 04:36

ok so far so good

I figured out srseive and sr1seive, playing with them :)

now how do i give multiple k's?

Thomas11 2012-09-14 08:11

[QUOTE=shanecruise;311542]ok so far so good

I figured out srseive and sr1seive, playing with them :)

now how do i give multiple k's?[/QUOTE]

Just by entering all expressions one after another separated by spaces, e.g. this way:
[CODE]./srsieve 12345*2^n-1 12347*2^n-1 12349*2^n-1 -n 1000000 -N 1200000[/CODE]

Note that sr1sieve works for single k only. If you want to sieve multiple k at once you should use sr2sieve.

[QUOTE=shanecruise]Also is this sr1sieve great for mersenne prime? k=1[/QUOTE]

I never tried. But why should you do this?

Kosmaj 2012-09-14 09:34

Of course you can sieve Mersenne primes with srsieve but it will very unefficient, because srsieve is a general purpose sieving software not employing the fact that factors can be only of form 2*m*n+1 (of 2^n-1). So even PrimeForm which has this feature will be faster than srsieve. But Gimps has many more speed-ups, and can do P-1 test, ecm curves and so on. You can download Gimps and try sieving without joining the project.

Regarding 5*n^2-1, can you tell us which file have you tried to sieve, and what factors have you found? Those test files are not the latest version, but they are fairly new, from a sieving file released a few months ago.

shanecruise 2012-09-14 09:51

told you just playing with things :)


All times are UTC. The time now is 00:56.

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