mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Riesel Prime Search (https://www.mersenneforum.org/forumdisplay.php?f=59)
-   -   5*2^n-1 Reservation Thread (https://www.mersenneforum.org/showthread.php?t=9472)

VBCurtis 2007-10-10 05:00

I put one of my Core2 cores on sieving this, since there is such demand. In about a week, I'll be done to 38T, and will release 550-575 for Patrick and 575-600 for Battlemaxx (via kosmaj); as Gary pointed out, I'll take 600-630 myself, and continue sieving before I release 630+.

Gary, I'll just assume you want 630-660 by the time it's ready. Consider the gap a chance to push your own searches quickly. :) You did not start a bad trend- my sieving is slow enough that you and patrick (and anyone else) are welcome to plan ahead, decide you'll be ready to process the next range by the time it's ready, and reserve it ahead of time. In principle, reserving any range-file you are ready to work on upon release makes sense to me.

-Curtis

gd_barnes 2007-10-11 05:54

[quote=VBCurtis;116033]I put one of my Core2 cores on sieving this, since there is such demand. In about a week, I'll be done to 38T, and will release 550-575 for Patrick and 575-600 for Battlemaxx (via kosmaj); as Gary pointed out, I'll take 600-630 myself, and continue sieving before I release 630+.

Gary, I'll just assume you want 630-660 by the time it's ready. Consider the gap a chance to push your own searches quickly. :) You did not start a bad trend- my sieving is slow enough that you and patrick (and anyone else) are welcome to plan ahead, decide you'll be ready to process the next range by the time it's ready, and reserve it ahead of time. In principle, reserving any range-file you are ready to work on upon release makes sense to me.

-Curtis[/quote]

Curtis,

OK, works for me. I'll reserve 630-660 now. IMHO, this k is high priority until we get close to n=1M so I'm right on it with two 3.2 Ghz P4 cores within a day of you sending me the sieve files. Later on, I may temporarily stop part of my twin prime search to put a 3rd core on some of the higher ranges if we maintain the n=25K-30K pieces.

n=525K-550K is now already up to n=534K. No primes yet. ETA is early-evening Friday. So if you want to send me 630-660 this coming Saturday regardless of how far sieved, that works for me.


Gary

VBCurtis 2007-10-11 20:21

I am sieving 3 different places at the moment, and only bother to combine them weekly. I'll combine ~Tuesday at 38T, sending everything up to 660 out at that point. That should keep everyone working a couple weeks, allowing me to get up to 50T before the next batch is ready. "proper" depth at 660k is around 55T, but we all know sieving within 1 bit of proper depth is good enough- the difference between 40T and 55T is perhaps 2 tests' worth of "free" sieving in that 630-660 range.

I plan to LLR 660-700 myself, with no further reservation for me until >850k.

-Curtis

gd_barnes 2007-10-12 23:09

k=5
525-550 complete
no primes found

Next week works on 630-660.


Gary

Kosmaj 2007-10-14 07:53

This is a new thread to coordinate efforts testing k=5. Please post your reservations here.
Previous posts related to k=5 can be found [URL="http://www.mersenneforum.org/showthread.php?t=9136"]here[/URL].

Curtis, if you want to add something to the first post in the thread please let me know and I'll do it.

Thanks.

gd_barnes 2007-10-15 16:48

[quote=Kosmaj;116312]This is a new thread to coordinate efforts testing k=5. Please post your reservations here.
Previous posts related to k=5 can be found [URL="http://www.mersenneforum.org/showthread.php?t=9136"]here[/URL].

Curtis, if you want to add something to the first post in the thread please let me know and I'll do it.

Thanks.[/quote]

Kosmaj,

Thanks for separating this! It looks much better now.

One more thing to add...I think Curtis reserved n=660K-700K per his last message here.


Gary

Kosmaj 2007-10-15 17:24

I just added 660-700.

BTW, I'd like to suggest that we collect results (LLR residues) of tested k=5 numbers, and I'd like to suggest that Curtis collects them, so that people mail their results to him. I hope Curtis agrees?

em99010pepe 2007-10-15 17:54

Can I help? I have a spare machine...

gd_barnes 2007-10-15 20:03

results files and freeing up sieving resources...
 
[quote=Kosmaj;116401]I just added 660-700.

BTW, I'd like to suggest that we collect results (LLR residues) of tested k=5 numbers, and I'd like to suggest that Curtis collects them, so that people mail their results to him. I hope Curtis agrees?[/quote]

Kosmaj,

I completely agree. It's funny that you mention that. Curtis is already collecting them. I've sent him both of mine. He's requesting them from us as he sends us new sieved ranges.

This brings up one question...should we report the ranges LLR'd to Prof. Caldwell's site as we do them or is it better to wait until later when we have one very large range done?


Curtis,

With 5 of us (I think) now ready to do or already doing LLR work on k=5, I'll suspend another sieving effort that I'm working on and free up my dual-core Athlon for sieving on both cores. Send me whatever P-ranges you think are best and if you have time, the latest sieved files with as many factors removed as possible for maximum sieving speed.

At the rate we're going, we may push n=1M by the end of 2007. :surprised That would be cool! :cool:


Gary

VBCurtis 2007-10-16 05:20

Thanks, Gary. I'll send everything tomorrow- new sieve file, suggested p-ranges, and the range you reserved. Thanks for the sieve offer- it'll really help. Get sr1 v. 1.1.12 if you haven't- I think it's faster on some architectures.

Carlos- I'll have work available for you in 10-12 days. We're a little behind on sieving, but with Gary's help we'll catch up enough to have work available near 30 Oct. It will be 700-720k (or a larger block, if you want it).

With so much interest, we'll definitely be at 1M by the end of the year.

Kosmaj is quite correct- Gary suggested to me via PM to collect ranges, and I am doing so. When a range is complete, please email me the results file to the email address you receive the work file from (this name at gmail).

-Curtis

VBCurtis 2007-10-17 06:51

ranges have been sent out:
550-575 Patrick
575-600 Battlemaxx
630-665 Gary (edited from 660, so that he has more work)

665k-4m still sieving, at 38T currently, on 5 cores thanks to Gary offering two Athlon64 cores to get us caught up with demand.
Sieve speed on my Conroe is ~11Mp/sec, 625 sec/factor at 38T. The new motherboard allows a healthy overclock. :grin:

The next batch from 665-800 should be ready at the end of the month, sieved to 58T or so. I have reserved 665-700, but 700-800 will be available in 20k chunks.
Happy Hunting!
-Curtis


All times are UTC. The time now is 16:51.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2022, Jelsoft Enterprises Ltd.