S774
 

S774 tested n=25K-100K - Nothing found

Results emailed - Base released

Riesel base 878 completed to n=25000 and released. 45 k remain.

Reserving R837 as new to n=25k

Reserving R927 as new to n=25k

R655 &R687
 

Reserving R655 and R687 as new to n=25K

[QUOTE=Puzzle-Peter;252497]Reserving R927 as new to n=25k[/QUOTE]

FYI:
k=144, 2116, and 4900 are eliminated by a factor of 29 on odd-n and algebraic factors on even-n and do not need to be tested.

There are no such k's on R837.

[QUOTE=gd_barnes;252547]FYI:
k=144, 2116, and 4900 are eliminated by a factor of 29 on odd-n and algebraic factors on even-n and do not need to be tested.

There are no such k's on R837.[/QUOTE]


Thanks!

Reservations
 

Reserving S655 and S687 as new to n=25K

R790
 

R790 tested n=25K-100K

20*790^40772-1 is prime

48*790^n-1 is now a 1ker with a weight of 1343

Results emailed - Base released

Riesel base 938 done to n=25000 and released. 17 k remain.

Edit: The files show 27k's remaining and the total k's with the 270 primes does = 297 which it should.
I'll assume a typo. Ian

Mark,

115*938*2223-1 is composite. It has a factor of 19.

Testing on my part found that 115*938^22223-1 is prime. It was only a lucky guess that a "2" was left out of the exponent.

I don't run primality proofs on everyone's primes, although I probably should. It just takes too much time with the # of primes that I get in on a daily bases. I only found this because I have files of primes to n=5K for many of the remaining bases with CK < 10K. It came out because with the # of primes that you have for n>5K (a total of 8), using my file of k's remaining for n<=5K (a total of 36), I showed that there should have been 28 (not 27) k's remaining. Changing k=115 to the correct prime so that there were 9 primes for n>5K put things in balance.

My question is: Why are primes being manually typed into files? That is the only explanation for a missing "2" in the exponent of a prime. I have griped about this before and have requested that people not do it. Finding this causes me to question the integrity of the k's remaining on many of your bases. How am I to know that you are not taking your file of k's remaining at n=1K and manually removing the k's found prime for n=1K-25K? I don't.

For everyone, based on this, here is a requirement for what I need in the future for testing up to n=25K:
1. A file of scripted primes and k's remaining at your nominal testing limit of n=1K or 2.5K or 5K.
2. A file of primes for n=1K (or 2.5K or 5K) to 25K.
3. Optional: A file of k's remaining at n=25K.

The files need to come directly from PFGW/LLR/PRPnet with one exception: The files can be sorted by k or n using an automated program of your choice but cannot in any other way be manipulated, manually typed in to, etc. The primes from #1 and #2 cannot be combined. There are several people that already send me separate files like the above and I'm much more comfortable with it.

If you don't include the optional file in #3, it's very easy for me to take k's remaining in #1 and subtract primes found in #2. I would actually much prefer doing that over getting manually manipulated files for smaller conjectured bases. I'd only ask for the file in #3 on large CK bases like S63 and if you have an automated way of creating them.

The math is too important and it's too easy to automate tasks to have primes manually typed into the files and potentially mess up the proofs if the k is typed incorrectly or is incorrectly removed from a k's remaining file.


Thank you,
Gary

edit...P.S. The above only applies to bases that are both scripted to a small limit and then sieved/tested to n=25K. This would exclude base 3 if it is fully scripted to n=25K, which is a common way of testing solely that base. All that I need there are primes and k's remaining.