Bases 251-500 reservations/statuses/primes

[MyDogBuster]

Sierp 328 the last k, tested n=50K-100K. Nothing found.

Results emailed. Base released

[MyDogBuster]

Riesel 368 the last k, tested n=25K-100K. Nothing found.

Results emailed. Base released

[MyDogBuster]

Sierp 353 the last k, tested n=77.5K-100K. Nothing found.

Results emailed. Base released

[paleseptember]

Taking Sierpinski 446 for a run to n=25e3.
It's looking pretty average at n=3e3 (11 k-values remain.)

[Batalov]

Taking R436 to n=25K.

[MyDogBuster]

Reserving the following 2ker's to n=100K

S264 S289 S298 S304 S311

[paleseptember]

Taking Riesel 378 (conj k=1517) out to n=25e3.
21 k-values remain at n=3e3.

When moving to srsieve, I got the following warning:

WARNING: 9*378^n-1 has algebraic factors.
WARNING: 9*378^n-1 has algebraic factors.
WARNING: 9*378^n-1 has algebraic factors.
WARNING: 9*378^n-1 has algebraic factors.
WARNING: 361*378^n-1 has algebraic factors.
WARNING: 361*378^n-1 has algebraic factors.
WARNING: 361*378^n-1 has algebraic factors.
WARNING: 361*378^n-1 has algebraic factors.

Is there anything that I should be doing?

[Mini-Geek]

I recommend that you run hiddenPowers.pl (give it a sieve file in addition to a file with each k on a line, in the form of 9*378^n-1, like pl_remain.txt or a sequences file you can feed to srsieve). It's the easiest way to remove algebraic factors.

[paleseptember]

Thanks Mini-Geek. r378.txt is my input (a*378^n-1 for 21 a-values) and sr_378.abcd is my output from srsieve after running it up to P=1e9

C:\factor\srsieve>hiddenPowers.pl r378.txt sr_378.abcd
9*378^n-1       n=0 mod 2 factors due to 3^2
removed 0 line(s)
112*378^n-1     n=2 mod 3 factors due to 168^3
removed 0 line(s)
361*378^n-1     n=0 mod 2 factors due to 19^2
removed 0 line(s)

Looks about right?

[Mini-Geek]

Originally Posted by paleseptember

Thanks Mini-Geek. r378.txt is my input (a*378^n-1 for 21 a-values) and sr_378.abcd is my output from srsieve after running it up to P=1e9

C:\factor\srsieve>hiddenPowers.pl r378.txt sr_378.abcd
9*378^n-1       n=0 mod 2 factors due to 3^2
removed 0 line(s)
112*378^n-1     n=2 mod 3 factors due to 168^3
removed 0 line(s)
361*378^n-1     n=0 mod 2 factors due to 19^2
removed 0 line(s)

Looks about right?

Close, one problem: the sieve file wasn't in the format it was expecting, so it never found lines it could remove, even though I'm betting there were plenty (Edit: yep, something like 15% of the candidates for those k's, most likely). Convert it to a NewPGen-like format first, (e.g. "srfile -G sr_378.abcd") then run hiddenPowers.pl with that version of the sieve file. Then you can convert it back into ABCD format (e.g. "srfile -a t17_b378.prp").
Edit: Just as a sanity check that some lines should be removed, I sieved those 3 k's over your n range to P=1e6 and ran it through:

hiddenpowers seqs.txt t17_b378.prp
9*378^n-1       n=0 mod 2 factors due to 3^2
removed 237 line(s)
112*378^n-1     n=2 mod 3 factors due to 168^3
removed 362 line(s)
361*378^n-1     n=0 mod 2 factors due to 19^2
removed 317 line(s)

It removed about 16% of the candidates, though this is only considering those 3 k's. (was 5777 total, now 4861) Might not add up to too much when you consider all the k's, but every bit helps. And the one with the most removals was one srsieve didn't even recognize as having algebraic factors.
Edit 2: If you're curious about how these algebraic factors work, see http://www.mersenneforum.org/showpos...&postcount=814. It breaks down to just simple algebra: difference of squares, or difference or sum of cubes, etc. e.g. "9*378^n-1 n=0 mod 2 factors due to 3^2", when n=0 mod 2 is can be rewritten with 2m=n as 3^2*378^(2m)-1=(3*378^m)^2-1^2=(3*378^m+1)(3*378^m-1)

[paleseptember]

Ah, I didn't realise that .abcd format wasn't acceptable. Ran the script in .prp format, much better. Knocked out about 1.5% of the total candidates, which is nothing to sneeze at :)

Should I make running the hiddenPowers script part of my routine when testing new bases?