Bases 501-1030 reservations/statuses/primes

[MyDogBuster]

Riesel Base 904
Conjectured k = 1266
Covering Set = 5, 181
Trivial Factors k == 1 mod 3(3) and k == 1 mod 7(7) and k == 1 mod 43(43)

Found Primes: 687k's - File attached

Remaining k's: 15k's - File attached - Tested to n=25K

k=9, 144, 324, 729, & 1089 proven composite by partial algebraic factors

Trivial Factor Eliminations: 557 k's

Base Released

[MyDogBuster]

Reserving Riesel 954 and 1009 as new to n=25K

[gd_barnes]

S863, S881, and S902 k=8 conjectures are proven and added to the pages.

With Mark's latest round of them on the Riesel side, this now completes all proven k=8 conjectures on both sides. I have 5 others on the Sierp side that have one k remaining at n=25K. I'll post those over the next few days.

[rogue]

Originally Posted by gd_barnes

S863, S881, and S902 k=8 conjectures are proven and added to the pages.

With Mark's latest round of them on the Riesel side, this now completes all proven k=8 conjectures on both sides. I have 5 others on the Sierp side that have one k remaining at n=25K. I'll post those over the next few days.

Hey, I thought you said only two per day!!! That looks like three.

Isn't that annoying that so many of these small conjectures have a single k remaining at n=25000?

[gd_barnes]

Originally Posted by rogue

Hey, I thought you said only two per day!!! That looks like three.

Isn't that annoying that so many of these small conjectures have a single k remaining at n=25000?

You might also remember me saying that if its 3-4 to complete a grouping of something, then that is fine.

As for 1 k remaining, although annoying, it's definitely expected especially for bases > ~250. I'd be surprised otherwise. I don't know what others do but in this case, I searched 10 bases at once by n-value for n=5K-25K. I was fairly lucky to find a final prime on 5 of them since they were all b>300. 5 ended up with 1 k remaining. It is a pain to have to sieve them all separately but searching them all at once sure saves a lot of human time. It involves a bit of manual manipulation to get it into the proper PFGW formatted sieve file after sieving all of the bases. Srfile can't bring together multiple bases into one PFGW-formatted sieve file. I personally like searching them all at once upwards by n-value since that finds a prime the most quickly for most of them.

One hint if you search several bases at once, be very careful with the stop-on-prime option. Since many of the k=8 conjectures had k=4 remaining, you certainly wouldn't want to stop-on-prime for k=4. But in this case, since they all had only 1 k remaining, I was able to have it stop when a prime was found for the BASE. Note that that wouldn't work if you had more than one k in some of the bases since you'd miss searching the remaining n's for the k('s) without a prime, but it does work well for a bunch of 1 k remaining bases.

PFGW isn't sophisticated enough to be able to differentiate k=4 on one base from k=4 on a different base within the same search.

Mark, can PRPnet handle searching multiple bases at once? If so, can it stop on prime for a specific k / base combo instead of just stopping when a specific k -OR- a specific base finds a prime? That would be very cool.


Gary

[rogue]

Originally Posted by gd_barnes

You might also remember me saying that if its 3-4 to complete a grouping of something, then that is fine.

Hmm... You're giving me ideas...

Originally Posted by gd_barnes

Mark, can PRPnet handle searching multiple bases at once? If so, can it stop on prime for a specific k / base combo instead of just stopping when a specific k -OR- a specific base finds a prime? That would be very cool.

Yes, I use it for that frequently. If configured as a Sierpinski/Riesel server, the PRPNet server will stop sending out tests for a k/b/c combination when a prime is found it. Both Sierpinski and Riesel can be mixed in the same server even if the same k/b combos show up for both forms. It was how I distributed base 928 across multiple clients.

There is no way to stop searching if a prime is found for a base (regardless of k and c). Is this is need? If so, I would like to understand it further. If I didn't know any better, I suspect that you would want this for a GFN type search. PRPNet supports such a search, but does not stop if a prime is found for one of the bases.

[gd_barnes]

Originally Posted by rogue

Hmm... You're giving me ideas...



Yes, I use it for that frequently. If configured as a Sierpinski/Riesel server, the PRPNet server will stop sending out tests for a k/b/c combination when a prime is found it. Both Sierpinski and Riesel can be mixed in the same server even if the same k/b combos show up for both forms. It was how I distributed base 928 across multiple clients.

There is no way to stop searching if a prime is found for a base (regardless of k and c). Is this is need? If so, I would like to understand it further. If I didn't know any better, I suspect that you would want this for a GFN type search. PRPNet supports such a search, but does not stop if a prime is found for one of the bases.

Cool! No, afaik, stopping when a prime is found for a base would not be needed in PRPnet for our needs. It just came in handy for me on a pure PFGW search on many bases with 1k remaining. It would be handy if PFGW itself could stop on a k/base combo.

On the other topic, please don't "finish up" a group of something several days in a row. (lol, it wouldn't be finishing up a group of something then) If you really are finishing up a group of something, then that's fine. Weekends are very busy in my personal/business life but I have plenty of time for the projects on Monday and Tuesday; the opposite of most people. As an example, I skipped late Fri, all Sat, and most of Sun. updating the pages. I then updated them very late Sun./early Mon. There were already 10-12 new bases plus 3 more that I did. I had to follow up on 2 of them and there was one that was involved with 2 different kinds of algebraic factors. Now there's the added task of running srsieve whenever there is 1 k remaining.

If you guys wanna help me out a little, whenever you post a status on a new base with 1 k remaining, please run srsieve to P=511 for n=100001 to 110000 and let me know how many candidates are remaining. That will be the weight shown in the 1k thread.


Gary

[MyDogBuster]

Reserving Sierp 939 and Riesel 789 as new to n=25K

[rogue]

Primes found:

2*935^72-1
4*935^1-1
6*935^3-1
8*935^2-1
10*935^1-1
12*935^2-1

2*983^200-1
4*983^1-1
6*983^1-1
8*983^2-1
10*983^1-1
12*983^12-1

With a conjectured k of 14, both of these are proven.

[rogue]

I have completed this to n=100000 and am releasing it. No primes found. The residues are attached.

[MyDogBuster]

Sierp Base 939
Conjectured k = 46
Covering Set = 5, 47
Trivial Factors k == 1 mod 2(2) and k == 6 mod 7(7) and k == 66 mod 67(67)

Found Primes: 18k's - File attached

Remaining k's: 1k - Tested to n=25K
30*939^n+1

Trivial Factor Eliminations: 3k's

k weight = 1855

Base Released