Sierp. base 23 is complete to n=175K, no primes; results for 150K-175K are attached.

Continuing to (at least) 185K. :smile:

[quote=mdettweiler;159295]Sierp. base 23 is complete to n=175K, no primes; results for 150K-175K are attached.

Continuing to (at least) 185K. :smile:[/quote]


Oh heck, just take it up to n=200K! :-)

Your file name has base 33 in it. Just thought I'd let you know so you don't possibly misfile other files for base 23 in some base 33 folder on your machine. I've done that a time or two (or 3 or 4). :smile:

On another note: I'm getting ~3100 sec. test time for Sierp base 12 at n=~240K vs. your ~4700 sec. for Sierp base 23 at n=~175K. Even though the base is lower, the tests are still bigger. k*12^240000+1 is ~259K digits in length and k*23^175000-1 is ~238K digits digits. Even before your fftlen jump at n=~164K, your tests were still longer.

I don't know if that's your machine or if Phrot just likes base 12 better than base 23. One of my 64-bit Intel's is running it.


Gary

[quote=gd_barnes;159312]Oh heck, just take it up to n=200K! :-)

Your file name has base 33 in it. Just thought I'd let you know so you don't possibly misfile other files for base 23 in some base 33 folder on your machine. I've done that a time or two (or 3 or 4). :smile:

On another note: I'm getting ~3100 sec. test time for Sierp base 12 at n=~240K vs. your ~4700 sec. for Sierp base 23 at n=~175K. Even though the base is lower, the tests are still bigger. k*12^240000+1 is ~259K digits in length and k*23^175000-1 is ~238K digits digits. Even before your fftlen jump at n=~164K, your tests were still longer.

I don't know if that's your machine or if Phrot just likes base 12 better than base 23. One of my 64-bit Intel's is running it.


Gary[/quote]
Regarding the filename: thanks, I've fixed the attachment. And no, I don't think I have anything misfiled on my end--as soon as I had verified that the attachment uploaded correctly, I deleted the results file on my hard disk so, nothing to misplace. :smile:

As for how big my next reservation chunk will be: well, I see what you mean about just taking it up to 200K rather than in these smaller chunks, but, I must admit, these testing times are getting pretty hairy up here in the stratosphere...I'm not quite sure how much longer I can sustain this effort at a reasonable pace on one core. Especially if I hit a few more FFT jumps--and at the rate I've been hitting them on this base, I'm probably due up for another one sometime soon (probably within the next 5K or 10K). So, long story short, I figured I'd take things in small chunks for now so that I don't bite off more than I can chew. And since I'm using a presieved file anyway, it's not like I'm losing any sieving efficiency by doing things in smaller chunks. :smile:

Of course, all of this talk of testing times brings me to the topic of Phrot's testing times on base 23: yeah, I suspect that it does have to do with the fact that Phrot is definitely "friendlier" to some bases than others. Your 40% speed increase over LLR on base 12 is far, far more than any other previously reported speed increase with Phrot--indicating that base 12 is one of the most Phrot-friendly bases we've discovered yet. Thus, most any other base out there will probably be comparatively slower than base 12 with Phrot.

I haven't tested Phrot vs. LLR for base 23 yet--I'll probably do that tomorrow when I'm planning to do a number of such tests on various bases.

Max :smile:

Base 10 (Riesel + Sierpinski) tested till n=210000. I'm still working on it :flex:

[quote=Cruelty;159413]Base 10 (Riesel + Sierpinski) tested till n=210000. I'm still working on it :flex:[/quote]


If there is anyone who can prove these things with 1-3 k's remaining, it's Cruelty! This is the guy who will keep searching a k to n=2M or more if it refuses to yield a top-5000 prime. :smile:

Hey, Cruelty, I know you were going to take a couple of the k=300-400 k's that you had previously reserved at NPLB above n=1M because they hadn't yielded a prime for n>600K. Did you keep going with them? If so, how high did you get and any BIG primes found yet?


Gary

Time to take on a good-sized CRUS effort for a change:

Reserving all k's for Riesel and Sierp bases 22 and 28 and Riesel base 27 except for Sierp base 22 k=1611 & 5128, which are already at n=200K. 16 k's total.

On bases 22 & 28, I'll take them to n=150K. On the one k remaining for Riesel base 27, I'll take it to n=200K. If I'm feeling motivated, I ma take bases 22 & 28 to n=200K.

Now...to get sieving base 28. 4 of the k's which are only at n=50K.


Gary

Sierp. base 23 is complete to n=185K, no primes. Results for 175K-185K are attached.

Continuing to n=200K. :smile:

Riesel Base 6
 

search completed to n=150k

prime 79815*6^113777-1 not mentioned before!

all results and remaining test-files sent to Gary.

good luck for a new Team Drive!

Sierp base 12 k=404 is now complete to n=250K. No primes. Now unreserved.

If you don't mind long tests (slightly > 1 hour each using Phrot at n=250K), this is a nice low base with one k remaining that could be proven at any time. :smile:


Gary

IMO, someone should set up PRPNet for the conjectures that have a single k remaining. I suggest putting all of them into a single server. If enough people get involved, one or more those k might fall within a few weeks.

[quote=rogue;161823]IMO, someone should set up PRPNet for the conjectures that have a single k remaining. I suggest putting all of them into a single server. If enough people get involved, one or more those k might fall within a few weeks.[/quote]
Yes, we'd discussed that and such a setup is planned. However, at the time, our one existing PRPnet server (Sierp. base 6) is completely idle--come on guys, where'd everyone go? :smile: If we can get at least one or two people who are interested, then we'll go ahead and start a second PRPnet server for the single-k conjectures.