mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Riesel Prime Search (https://www.mersenneforum.org/forumdisplay.php?f=59)
-   -   Low weight stats page. (https://www.mersenneforum.org/showthread.php?t=2934)

storm5510 2020-02-09 23:07

[I]k[/I] = 10009. Some lightweight results from the start:

[QUOTE]10009*2^3-1
10009*2^27-1
10009*2^71-1
10009*2^101-1
10009*2^157-1
10009*2^279-1
10009*2^311-1
10009*2^679-1
10009*2^837-1
10009*2^1101-1
10009*2^1569-1
10009*2^2219-1
10009*2^2343-1
10009*2^5527-1
10009*2^7109-1
10009*2^7847-1
10009*2^9797-1
10009*2^25797-1
10009*2^47871-1
10009*2^57271-1[/QUOTE]

storm5510 2020-02-10 03:51

[I]k[/I] = 10009, [I]n[/I] = 192999
Hardware issue forced a stop.

Jean Penné 2020-02-10 07:22

k=138847
 
[QUOTE=Happy5214;537138]To confirm, does that include the range from n=1283793 to n=10M (which is the earliest progress report you posted)?[/QUOTE]
Yes, indeed, and I am also sieving the input file, now up to 1.61P

Regards,
Jean

VBCurtis 2020-02-10 07:34

[QUOTE=storm5510;537176][I]k[/I] = 10009. Some lightweight results from the start:[/QUOTE]

What makes you think k=10009 is a low-weight k value? This is quite a list of primes, if this was indeed a low-weight k.

storm5510 2020-02-10 16:43

Another project running on my HP

[I]6001*2^114775-1 is prime! (34555 decimal digits)
[/I]
[QUOTE=VBCurtis]What makes you think k=10009 is a low-weight k value? This is quite a list of primes, if this was indeed a low-weight k. [/QUOTE]

How do you define "low weight?"

storm5510 2020-02-10 18:14

[I]6001*2^140449-1 is prime! (42284 decimal digits)[/I]
[I]6001*2^156555-1 is prime! (47132 decimal digits)[/I]

paulunderwood 2020-02-10 23:55

[QUOTE=storm5510;537233]
How do you define "low weight?"[/QUOTE]

Given k, there is a threshold for which the k*2^n-1 numbers left in a sieve is below, given that one has sieved up to some maximum value. Less in the sieve means it is "lighter" and as a result one expects fewer primes. An extreme of weight is a [URL="https://en.wikipedia.org/wiki/Riesel_number"]Riesel number[/URL] which has no weight at all. Sometimes Nash weight is used for weight. Heavy means lots of numbers are left in the sieve and consequently one can expect more primes below a maximum n,

There is a program somewhere on this subforum which calculates weights,

storm5510 2020-02-11 14:13

[QUOTE=paulunderwood;537267]Given k, there is a threshold for which the k*2^n-1 numbers left in a sieve is below, given that one has sieved up to some maximum value. Less in the sieve means it is "lighter" and as a result one expects fewer primes. An extreme of weight is a [URL="https://en.wikipedia.org/wiki/Riesel_number"]Riesel number[/URL] which has no weight at all. Sometimes Nash weight is used for weight. Heavy means lots of numbers are left in the sieve and consequently one can expect more primes below a maximum n,

There is a program somewhere on this subforum which calculates weights,[/QUOTE]

Got it. Thank you! :smile:

I sieved a range of 50,000 possibilities on my HP overnight to 1e12. There were 1,100 remaining this morning.

[I]6001*2^315529-1 is prime! (94988 decimal digits)[/I]

I hope someone is recording what folks are turning in here!

VBCurtis 2020-02-11 15:08

You've chosen a k value (10009) from a range that nobody is tracking, noted that nobody is tracking it, and posted a bunch of primes of a size that aren't tracked by the top 5000.

If you expect a project to record your work, maybe pick a k value from a range that project cares about? The forum you're posting in, RPS, focuses on k < 300. For historical reasons, it also cares about multiples of 15 even if they're larger than 300 (the project was once called 15k).

No Prime Left Behind focuses on 300 < k < 2000 (I think?).

It should be obvious that you have an unlimited list of k values to choose from, but that doesn't mean someone cares about every possible k. I imagine the prime pages listings might include 6001, but you haven't reached a size that's all that interesting- a day or three of work isn't exactly a massive effort.

storm5510 2020-02-11 15:38

[QUOTE=VBCurtis;537347]You've chosen a k value (10009) from a range that nobody is tracking, noted that nobody is tracking it, and posted a bunch of primes of a size that aren't tracked by the top 5000.

If you expect a project to record your work, maybe pick a k value from a range that project cares about? The forum you're posting in, RPS, focuses on k < 300. For historical reasons, it also cares about multiples of 15 even if they're larger than 300 (the project was once called 15k).

No Prime Left Behind focuses on 300 < k < 2000 (I think?).

It should be obvious that you have an unlimited list of k values to choose from, but that doesn't mean someone cares about every possible k. I imagine the prime pages listings might include 6001, but you haven't reached a size that's all that interesting- a day or three of work isn't exactly a massive effort.[/QUOTE]

<300 has been worked heavily. 10009 did not appear in the top 5000 list nor the Riesel and Proth database. The latter is no longer being updated, it would appear.

6001 on my HP did not seem to have much effort put into it, so I picked it up. It also does not appear in the top 5000 list.

Make a suggestion and I will give it a go...

Happy5214 2020-02-12 05:09

[QUOTE=VBCurtis;537196]What makes you think k=10009 is a low-weight k value? This is quite a list of primes, if this was indeed a low-weight k.[/QUOTE]

Its Nash weight is 1261, which puts it a bit above what is generally considered low-weight (<1000).

I'll add that all k's between 10k and 100k have been fully tested below n~=1k, and k's between 10k and 15k have been tested to n=20k. The results from those can be found at [URL]https://www.rieselprime.de/Data/10e04a.txt[/URL].

(After reading the next few posts after a page cutoff...)

The old Riesel database is gradually being replaced with [URL]https://www.rieselprime.de/ziki/Riesel_prime_table[/URL], which is in a wiki format. [URL]https://www.mersenneforum.org/forumdisplay.php?f=89[/URL] would be a more appropriate place to post low-priority data, or you could apply for a wiki account at [URL]https://www.mersenneforum.org/showthread.php?t=24141[/URL] to enter the data yourself.

If you're looking for useful work, you may consider filling in gaps in the data for unreserved k's already in the old or new Riesel DB. Check with [URL]https://github.com/happy5214/rps/[/URL] to make sure you don't overlap with my work in that area. I also am working on the RPS 9th and 10th Drive k's below n=400k, all Woodall and near Woodall k's > 10k, and all k's between 10k and 15k with missing data, all on local PRPNet servers.

Edit: [url]https://www.mersenneforum.org/showthread.php?t=7213[/url] has the Nash weight calculators alluded to by Paul.


All times are UTC. The time now is 22:15.

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