mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Riesel Prime Search (https://www.mersenneforum.org/forumdisplay.php?f=59)
-   -   k=13236795 (https://www.mersenneforum.org/showthread.php?t=6529)

em99010pepe 2007-05-29 16:16

Mine has 3805, sieve up to 1.2T but a factor is found every 1700 sec.
20T?????? I was thinking to stop at 2T.

Carlos

VBCurtis 2007-05-29 19:40

Stop sieving when the factor removal rate is equal to the time to run an LLR test on a median candidate in the range.

For a small range like 425-450, candidates take roughly the same time to LLR, so run LLR on a single candidate and sieve until removal rate is equal to that time.

It's very inefficient to sieve small ranges; 1.2T is certainly plenty for the 425-450 range.

Carlos, if you have spare sieving resources, I have months of sieving work to do using sr2sieve. I'm sieving 11,13,31,45,99,127 all from 1M to 4M (except 11, which is already LLRed to 1.5M). This sieve will need to run to roughly 80T before LLR should begin, and I'm at 20T. If you'd like to do a T or so as a public service, I'd love it.

-Curtis

em99010pepe 2007-05-29 20:28

Cedric,

I stopped sieving.

Curtis,

You got a private message. Yes, I'll do 1T or 2T or even more....I like to sieve....

Carlos

Flatlander 2007-05-29 21:24

[quote=VBCurtis;107266]
...
Carlos, if you have spare sieving resources, I have months of sieving work to do using sr2sieve. I'm sieving 11,13,31,45,99,127 all from 1M to 4M (except 11, which is already LLRed to 1.5M). This sieve will need to run to roughly 80T before LLR should begin, and I'm at 20T. If you'd like to do a T or so as a public service, I'd love it.

-Curtis[/quote]
How long does it take to sieve a T? On what CPU?

VBCurtis 2007-05-30 02:37

[QUOTE=Flatlander;107278]How long does it take to sieve a T? On what CPU?[/QUOTE]

I have two data points:
TurionX2-1600, 600kp/sec using 32bit sieve.
Conroe-1950, 1500 kp/sec using 64bit sieve.
(per core for each)

I think that means the conroe does 1T/week/core, while the X2 takes 2.5 weeks/T/core.

I am happy to share the files for team LLRing once they are ready, if any of you potential sievers are interested. 13 and 11 are fairly low-weight, but 45 is very high-weight (higher than 15, I think). I want to LLR 99 on my own, and 31 is for Steven (as shown on k<300 page).

Also: if any of you are interested in testing a k from 1M to 4M, we can add it to this sieve and work together. You'd have to sieve it alone up to 20T to catch up to me, but adding another k to the sieve would not slow it down much.

Carlos-- I'll build an updated sr2data file and email it tomorrow.
-Curtis

ValerieVonck 2007-05-30 16:44

Carlos,

Thank you for your sieving ... I will check my gmail account.

Regards
C.

ValerieVonck 2007-06-05 11:53

Search still going on...

ValerieVonck 2007-07-09 16:06

Currently @ 402728 - No primes to report.

ValerieVonck 2007-07-28 09:34

Currently @ 411901 - No new primes to report

ValerieVonck 2007-09-06 16:05

Tested to 424998 -
Results available @ [url]http://users.skynet.be/bk261068/lresults.txt[/url]

Will continue to 450.00 (will decide later on to go further)

ValerieVonck 2007-10-29 10:19

Tested to 438384 still no new prime to report :sad:


All times are UTC. The time now is 06:31.

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