Bases 6-32 reservations/statuses/primes

[Xentar]

Just for your information:
at the moment, I am forcing sierp-17. I set up a PRPNet server, and it is at 580K at the moment. Damn, when it finds a prime, it'll be a big one

Results not sent to Gary. I will do this, when the last prime is found, or the range is finished (n=1M).
Continuing.

[gd_barnes]

In some status reports in Emails from Mathew, he has completed the following k's on Riesel base 30 to n=140K:

225
239
249
659
774
1024
1580
1642
3256

Nothing to report yet.

[paleseptember]

Am going to take Sierpinski base 7 out for a spin. There are 7 k remaining, at 25K each, testing done by Mini-Geek.

I'll say up to 30K, as I don't know how long extending it that far will take. If it's reasonably quick, I'll update the reservation to indicate that I'm taking it further.

[Mini-Geek]

Originally Posted by paleseptember

Am going to take Sierpinski base 7 out for a spin. There are 7 k remaining, at 25K each, testing done by Mini-Geek.

I'll say up to 30K, as I don't know how long extending it that far will take. If it's reasonably quick, I'll update the reservation to indicate that I'm taking it further.

Note that those k's are for k<=1M. The conjectured k is 1112646039348. (like base 3, the conjecture is so high that it's infeasible to prove now or in the foreseeable future; I just did a little work on it because nobody had done any at all, unlike base 3)
While you can certainly extend those k<=1M k's to higher n's, note that there is still plenty of sideways mobility (i.e. higher k values) if you're interested in that sort of Sierp base 7 work.

[paleseptember]

Originally Posted by Mini-Geek

Note that those k's are for k<=1M. The conjectured k is 1112646039348.

I didn't realise that the conjectured k was quite so large! Nevertheless, I will take the 7 k values a little further along. You never know, I may find a prime along the way!

[paleseptember]

Sierpinski base 7 taken to 30K, releasing.

550366*7^28905+1 is prime! (apparently)

(I just searched the pfgw.out file for the string 'prime'.)

Primality testing 550366*7^28905+1 [N-1, Brillhart-Lehmer-Selfridge]
Running N-1 test using base 2
Calling Brillhart-Lehmer-Selfridge with factored part 99.98%
550366*7^28905+1 is prime! (14.6887s+0.0016s)

[Xentar]

Usually, PFGW creates an extra file pfgw.log where all the PRPs are listed. So you don't have to search the big output file :)

[paleseptember]

Originally Posted by Xentar

Usually, PFGW creates an extra file pfgw.log where all the PRPs are listed. So you don't have to search the big output file :)

>.>

Okay, that makes perfect sense. But I could have sworn that that file wasn't there when I first looked this morning!

[gd_barnes]

Originally Posted by paleseptember

>.>

Okay, that makes perfect sense. But I could have sworn that that file wasn't there when I first looked this morning!

If you were proving the primes as you went through the file doing primality testing using the -t switch, the primes would have been in the pfgw-prime.log file and doing a find on "prime" in pfgw.out would work OK. But if you were not using the -t switch, you would have needed to search for "PRP" in pfgw.out.

It appears that you were using the -t switch since you had a proven prime in your pfgw.out file. A hint for future reference: It's quite a bit faster to avoid their proof as you are primality testing to get a list of PRPs that will show up in pfgw.log. Then at the end, use the pfgw.log file (renamed) as input to prove all the PRP's with the -t switch for Sierp tests or -tp switch for Riesel tests.

Do you have the pfgw.out file still available? If so, can you post it here?


Thanks,
Gary

[paleseptember]

Please find attached the pfgw.out file for Sierpinski base 7, 25-30K.

[gd_barnes]

Originally Posted by paleseptember

Please find attached the pfgw.out file for Sierpinski base 7, 25-30K.

OK, thanks. You might consider running PFGW without the -t switch (primality proof) in the future. You'll save quite a bit of CPU time. Without the switch, it will write out unproven PRPs to pfgw.log and then you can do proofs on those.

Although it still needs to be proven after your run, the chance of a PRP being composite for base 7 at n>25K is much less than there being a memory error on your computer that causes invalid tests. Likely 1 in 1 million or less.