[quote=Kosmaj;119127]I don't know about the drive. Regarding the limit to test I'm enclosing the direct mail I got from Nugget in June, and you decide by yourself (two unrelated parts removed). He hasn't mailed me residues he is referring to.[/quote]

Oh, OK, thanks, Kosmaj. You're right. It shouldn't be included in the drive. If he happens to drop you a note before year end, can you let him know that I'll have it tested to n=600K and will have it sieved to n=800K?

I'll stop LLRing at n=600K (probably 3rd week in Dec. or so) and see if we hear anything from him by the end of the year. If not, I'll try dropping him a line by both a note in a thread here and a PM. If he gets in touch and wants to test it further, I'll be glad to provide him with my sieve file from n=600K-800K at that point. If not, I'll start in late Jan. testing from n=600K-800K.


G

Curtis

I don't think it's the best idea to plan a drive based on hardware that may appear in the future, it may not appear as well, or not at that point in time. And more important, neither individuals nor institutions can afford to update their hardware as soon as something new appears. [Until late this summer we had in my office an old 800MHz P-3 machine bought way back in 2000, and it was still in a good condition.]

So your plan, even with "only" 12 Ks appears to me "a bit" unrealistic but regardless of everything if that's what you want to do, then go ahead, we'll help you sieving and testing (once the 5th drive is over) as much as we can. I don't see any problems up to about 1.05M anyway.

And good luck!

I agree that 12 ks to 1.5M is an aggressive goal, but it shouldn't hurt much
to sieve up to 1.5M. If progress peters out too much after 1M we could
split up the ks and make them available for individuals.

Curtis,

I don't think we should sieve beyond 1.3M as there is an FFT increase just above there. k=17 is currently using the FFT length that we would finish with at 1.3M and is taking over 30 mins per candidate on a 2.4 Ghz Core 2 Duo. My feeling is the drive won't get much beyond 1.1M. I think that it will be much more palatable for individuals then to do 200k with a single k than be faced with 400k with such large FFT's.

I also don't expect CPU speeds to increase significantly enough to change these conclusions over the next 18-24 months.

Dave

My inclination at this time is to split Larry's and Amphoria's opinions. It doesn't hurt much to sieve to 1.4M, and we can release individual k's from 1.1 or 1.2 to 1.4 if interest wanes in a team drive at that depth.

Keep in mind that by 12 months from now, there won't be much work on k<300 below 1M. I doubt this 6th drive will be at 1M by 1/09, so 1.1-1.4M tests in 2009 will be par for any k's under 300.

I'll start the sieve to 1.4M shortly. I'll report when it hits 1T, making the file available for others to sieve if they wish.
-Curtis

The sieve is now running full-time on one Athlon64 1.6Ghz core. It is at 700G, with factors found every 40 seconds or so. With 64 bit sr2, sieve speed is around 1050 kp/sec, almost 100G/day.

If anyone is interested in contributing sieving work, please reply in this thread and/or PM me your email. I'll send the sr2data file, and coordinate sieve ranges. If possible, please reserve ranges in multiples of 1T. I'll get to 3T on my own, so shared sieving will start there if there is any interest.

-Curtis

Curtis,

I will be able to help in early Jan. I'll PM you when I am ready.

Dave

I'll do a 3e12 range too. I'm ready to start anytime.

This drive is now sieved to 11T, with AES and Amphoria sieving ranges to 16T. We should be ready to begin testing near 1 April, if the 5th drive is complete.

Sieve work completed:
AES 6T, 1 more in progress.
Amphoria 2T, 4 more in progress.
VBCurtis 3T, more progress soon.

Thanks for your help, guys.
-Curtis

Amphoria has done nearly all of the sieving from 16T to 28T. At 30T, we are ready for the drive to begin on 600-650k. When the 5th drive is done, the 6th drive will be ready and fully sieved for the early testers!

Thanks for all the sieving work, Dave.
-Curtis

Update please. Thank you.