Riesel base 16 - team drive #2

[mdettweiler]

Taking 176K-180K.

[gd_barnes]

Give it up dudes. If you'll just let me take the last 20K, I'll find some primes.

I have to own one corner of the prime-search world. Ian owns most of the rest on the 2 projects.

[mdettweiler]

Originally Posted by gd_barnes

Give it up dudes. If you'll just let me take the last 20K, I'll find some primes.

I have to own one corner of the prime-search world. Ian owns most of the rest on the 2 projects.

Go ahead, take the last 20K if you're up to it--I'd be glad to see this drive get seriously moving, regardless of who's searching it.

Or, even better: you could hammer away at the Sierp. side while I do the Riesel side. They've both got the exact same # of k's remaining, so the effort required should be equivalent. One whole quad should be sufficient to keep up with me, since that's exactly what I've got on Riesel base 16 right now.

[gd_barnes]

Originally Posted by mdettweiler

Go ahead, take the last 20K if you're up to it--I'd be glad to see this drive get seriously moving, regardless of who's searching it.

Or, even better: you could hammer away at the Sierp. side while I do the Riesel side. They've both got the exact same # of k's remaining, so the effort required should be equivalent. One whole quad should be sufficient to keep up with me, since that's exactly what I've got on Riesel base 16 right now.

He he. Naw...I'm good. After the upcoming sieving drive at NPLB gets finished and the remaining 2 drives set up, I'm going after some non-power-of-2 top-5000 primes.

We'll have to do some serious sieving on these drives once they hit n=200K. I'd erred when I didn't sieve them to n=400K to start with like we did on base 6. Now that we're at n=200K, we almost may as well sieve them to n=600K or 800K with the next sieving effort. Testing that should keep us entertained for a few years.


Gary

[mdettweiler]

Gary, would you mind posting more files for this drive--preferably everything up to n=200K? I won't be reserving quite that big of a chunk all at once, but unless anyone else grabs any files I'll probably eventually be taking this drive all the way to 200K.

[mdettweiler]

Taking 180K-190K.

[mdettweiler]

Finally, after a long dry stretch, this drive has produced a prime:

19413*16^181080-1 is prime!
(submitted as 19413*2^724320-1)

And this just *had* to be right after I'd loaded an entire 10K range into the LLRnet server I'm using for this, rather than in the smaller 2K or 4K chunks I was doing before--so now I'm going to be stuck with k=19413 left in all the way up until n=190K. Oh well--at least since it's top-5000 range, it won't be entirely a waste since there's still a chance I might pull a bonus top-5000 prime out of it.

Max

[gd_barnes]

Originally Posted by mdettweiler

Finally, after a long dry stretch, this drive has produced a prime:

19413*16^181080-1 is prime!
(submitted as 19413*2^724320-1)

And this just *had* to be right after I'd loaded an entire 10K range into the LLRnet server I'm using for this, rather than in the smaller 2K or 4K chunks I was doing before--so now I'm going to be stuck with k=19413 left in all the way up until n=190K. Oh well--at least since it's top-5000 range, it won't be entirely a waste since there's still a chance I might pull a bonus top-5000 prime out of it.

Max

NO!! Only I can find primes for n>125K. lol

Good one! Hey, you're a server guru. Just pull the range out, use srfile to delete the k's, and put it back in. But, what the heh, if you leave it in, you might get a bonus prime from the same k.

Edit: Is your n<180K range done or were you going to report this whole big range at once?


Gary

[mdettweiler]

Originally Posted by gd_barnes

Good one! Hey, you're a server guru. Just pull the range out, use srfile to delete the k's, and put it back in. But, what the heh, if you leave it in, you might get a bonus prime from the same k.

Well, as we discussed before with the public LLRnet servers, removing a k in the middle of an LLRnet range can cause big problems with sorting and verifying the results later on. That said, however, there is no reason why I couldn't remove the k for, say, the 183K-190K range, which the server hasn't even touched yet...

Of course, yeah, I guess I could very well leave it in to possibly nab a bonus prime. k=19413 is not too terribly high-weight; however, at >720K base 2, these tests are taking about 15 minutes apiece, and even just a handful of unnecessary candidates left in could mean a number of extra CPU hours. Given that, I think I'll remove k=19413 from 183K on. (Of course, it will still be left in up to 183K, so I may very well find another prime below that. )

Max

[gd_barnes]

Originally Posted by mdettweiler

Well, as we discussed before with the public LLRnet servers, removing a k in the middle of an LLRnet range can cause big problems with sorting and verifying the results later on. That said, however, there is no reason why I couldn't remove the k for, say, the 183K-190K range, which the server hasn't even touched yet...

Of course, yeah, I guess I could very well leave it in to possibly nab a bonus prime. k=19413 is not too terribly high-weight; however, at >720K base 2, these tests are taking about 15 minutes apiece, and even just a handful of unnecessary candidates left in could mean a number of extra CPU hours. Given that, I think I'll remove k=19413 from 183K on. (Of course, it will still be left in up to 183K, so I may very well find another prime below that. )

Max

Your choice there. I should have been more specific. Of course I knew it's a problem if you pull the entire file out and remove the k. Doing it above a set nice round point like n=183K is certainly the way to go.

Were you going to wait until you were complete to n=190K to report any n<180K range complete? Just curious.


Gary

[mdettweiler]

Originally Posted by gd_barnes

Were you going to wait until you were complete to n=190K to report any n<180K range complete? Just curious.

Yep, that's the plan. Actually, what I'm going to do is wait until it completes to 183K, *then* report 170K-183K--that seems to be a good breaking off point, since that's the point at which I'll have to switch to verifying things with a sieve file that does not include k=19413.

After that, yes, I'll wait until at least 190K to report the results--in fact, after 190K, unless something else comes up that I need to shift my quad to, I'll probably reserve 190K-200K. In that case I'll just wait until 200K to report the rest of my results (unless of course I find a prime in which case I'll do what I'm doing at 183K).

Max