Bases 501-1030 reservations/statuses/primes

[rebirther]

R742 tested to n=25K

39 primes found - 78 remain

Results emailed - Base released

[rebirther]

R810 tested to n=10K

353 primes found - 534 remain

Results emailed - Base released

[rebirther]

Reserving R784 to n=200k (for testing in my BOINC project) with the available sieve file sieve-riesel-base784-100K-200K.txt

[MyDogBuster]

Reserving S757 to n=10k

[rebirther]

Reserving R920 to n=100k (for SRBase project) with the available sieve file http://www.noprimeleftbehind.net/cru...0-50K-100K.txt

[LaurV]

Stimulated by Serge's tremendous success , I put one of my cores to base 972, with the file downloaded from here, 15 k's, was to 25k, going toward 100k, it started with about 20 seconds per line (PFGW, is this the right tool?) and reached now close to 30k, getting slower and slower, I may let it run to 100k or stop at 50k, no idea. No primes till now. I am saving the residues (does it make sense?).

Also, trying to see what srXsieve are doing (wanted to try them for long time, but never found the time) I started "fresh" on base 967 (single k) and sieve it as high as I can (sr1sieve, after a newpgen start; I was afraid to start with srsieve, hehe; so, is sr1sieve the right tool now? is there a file already sieved for this base? or I do need to extract it from the big - 108MB - file containing all bases?). Still sr1sieving, be patient with me if I am taking a wrong path here.

[gd_barnes]

Originally Posted by LaurV

Stimulated by Serge's tremendous success , I put one of my cores to base 972, with the file downloaded from here, 15 k's, was to 25k, going toward 100k, it started with about 20 seconds per line (PFGW, is this the right tool?) and reached now close to 30k, getting slower and slower, I may let it run to 100k or stop at 50k, no idea. No primes till now. I am saving the residues (does it make sense?).

Also, trying to see what srXsieve are doing (wanted to try them for long time, but never found the time) I started "fresh" on base 967 (single k) and sieve it as high as I can (sr1sieve, after a newpgen start; I was afraid to start with srsieve, hehe; so, is sr1sieve the right tool now? is there a file already sieved for this base? or I do need to extract it from the big - 108MB - file containing all bases?). Still sr1sieving, be patient with me if I am taking a wrong path here.

Welcome to the project LaurV! :-)

For future reference, please refer to which side you are reserving; that is Riesel (-1) or Sierpinski (+1). Your link and descriptions made it clear that it was Riesel here but it will be easier to understand if you say R972 or S972 (or R967 or S967). I will show as reserved R972 to n=100k and R967 to ? for you.

You can use PFGW or LLR to do the searches. I have experienced and heard that newer versions of LLR are slightly faster. You might want to experiment for yourself. Be sure and use the "stop-on-prime" option for whichever program that you use to avoid additional searching for already primed k's.

Sr1sieve is the correct sieving program for one k. You should start sieving a range with srsieve. NewPGen is very inefficient for all searches here.

There is not an already sieved file for R967. I do not know which "big 108MB" file that you are referring to. We have no big file that contains all bases. The current team sieving effort is sieving the k's that we are working on for only bases 2 and 4.

[LaurV]

Originally Posted by gd_barnes

Be sure and use the "stop-on-prime" option for whichever program that you use to avoid additional searching for already primed k's.

very good point, thanks a lot! I was desperately watching for pfgw's separate log file, which he does when he finds a prime, hehe. It didn't come. What you say makes more sense.

Sr1sieve is the correct sieving program for one k. You should start sieving a range with srsieve. NewPGen is very inefficient for all searches here.

It does not take long to get to the right limit (over k, n) like few minutes, but I let it for a while to make sure. Then I switched to sr1sieve. I was afraid to start with srsieve from scratch, as I never used the srX sievers before. "Taking over" a file already created by a tool I knew (newpgen) was easier and somehow "error proof" (in case I was giving an idiotic command line to sr1sieve).

There is not an already sieved file for R967. I do not know which "big 108MB" file that you are referring to. We have no big file that contains all bases. The current team sieving effort is sieving the k's that we are working on for only bases 2 and 4.

I was talking about that one Carlos and JimB made, here in a parallel thread, and fortunately didn't download it due to its size, but after your explanation I realized that it was a total confusion in my head, that is different k's for base 2 only, written large at the beginning of the post.

[LaurV]

ok, we arrived home and checked the files. Yesterday we didn't look (sunday, shopping, blahblah) and in the morning we were in hurry to leave for work. It seems in this part of the world that 78*972^27907-1 is prime, and it was found 48 hours ago (exactly). The testing ran all this time on the 15 k's, which means that we wasted ~48/15= 3 hours and 12 minutes testing higher n for this k (assuming they are uniform distributed in the file, we didn't check exactly how many lines have 78 in front in the file). In the next 10 minutes we will make a small script to grab out of the file all lines with "n smaller than 31915", (which we already reached) or k=78 (which we already don't need anymore). (we don't know other method to filter the file). What's left of it, we will face it against pfgw again, to see what and how. If I understood right, this is a 14 k-er now.

[edit: script done, of course, now we learned our lesson and will use "stop if a prime is found", in spite of the fact we don't exactly know how to do it, in this very moment, but we are going to investigate pfgw help]

[gd_barnes]

Originally Posted by LaurV

[edit: script done, of course, now we learned our lesson and will use "stop if a prime is found", in spite of the fact we don't exactly know how to do it, in this very moment, but we are going to investigate pfgw help]

Who is "we"? Are you more than one person? :-)

Make the header of the sieve file look like this:

ABC $a*972^$b-1 // {number_primes,$a,1}

LLR is more clear. Put the following line in the llr.ini file:

StopOnPrimedK=1

Please report all primes at the end of your search and not one at a time. For searches n>25k I'll also need the residues, that is the pfgw.out file in PFGW.

[gd_barnes]

Originally Posted by rebirther

R862 tested to n=10K

185 primes found - 321 remain

Results emailed - Base released

I did not get an Email from you on this one. Can you resend it? Thanks.