[quote=Xentar;165450]Done.
Did you receive the files?

Could you please recheck just one of the latest results? I overclocked my CPU a little bit. The prime stresstest says, everything is ok, but I think it is better to be sure :)[/quote]

Yes, I got them. Thanks.

OK, I'll run a few results and check the residues.


Gary

Riesel base 22 at n=106K; no primes
Sierp base 22 at n=103K; no primes
Riesel base 28 at n=110K; no primes
Sierp base 28 at n=114K; 1 prime already reported

Riesel Base 23 at n=166.2K No primes

[QUOTE]Riesel Base 23 at n=166.2K No primes [/QUOTE]

Meant to add - continuing to 200K. Sorry

Base 10 Riesel + Sierpinski tested till n=250000, no primes so far. I will sieve it a little bit more before continuing (current sieve depth ~50T). Attached are results for 195000<n<250000.

Due to memory restriction, I've to suspend the sieving of Sierp. base 63 candidates with approximately 400,000,000 k/n pairs remaining and sieved to ~1.5G... however I need something less RAM intensive to put on my Quad, so I'm reserving Sierpinski base 25 to n=25K. Already at the moment there has been tested to n=50 and sieving is ongoing for the remaining 5,967 k's with 9,482,052 k/n pairs remaining at current sieve depth of ~650M. Currently 1 k/n pair is removed every 0.14 seconds and optimal sieve depth si approximately 72 seconds per k/n pair at most, and likely less. I've removed every k mod 3 = 2 before sieving aswell every even k has been excluded from the beginning. As soon as I complete this task I hope to be able to find a way to resume the sieving of Sierpinski base 63 without consuming the entire amount of availeable RAM :smile:

Take care everyone.

Kenneth!

Ps. Riesel base 3, k=3677878 is at n=510000 and continuing :smile:

Is expanding the Sierpinski base 25 reservation to at least n=131072 (2^17), is expecting to have all k's tested to n=32178 in less than 1 month and after that a combined sieving of all remaining k's for the entire remaining n-range will begin, using sr2sieve :smile:

KEP

Ps. I've removed following k's from further testing since they appear to be at n=~225K through the Sierpinski base 5 testing (please correct me if I'm wrong to exclude these k's):

25570 (Primepages base=5 k=5114)
120160 (Sierpinski base=5 k=24032)
133990 (Sierpinski base=5 k=26798)
158560 (Sierpinski base=5 k=31712)
186460 (Sierpinski base=5 k=37292)
186640 (Sierpinski base=5 k=37328)
208690 (Sierpinski base=5 k=41738)
221560 (Primepages base=5 k=44312)
221740 (Sierpinski base=5 k=44348)
223690 (Sierpinski base=5 k=44738)
255880 (Sierpinski base=5 k=51176)

Riesel base 22 at n=125K; no primes
Sierp base 22 at n=115K; no primes
Riesel base 28 at n=120K; no primes
Sierp base 28 at n=125K; 1 prime previously reported

# of primes is since starting at n=100K.

Stubborn bases!

I'll get everything updated for statuses in the last week here today.

[quote=KEP;167480]Due to memory restriction, I've to suspend the sieving of Sierp. base 63 candidates with approximately 400,000,000 k/n pairs remaining and sieved to ~1.5G... however I need something less RAM intensive to put on my Quad, so I'm reserving Sierpinski base 25 to n=25K. Already at the moment there has been tested to n=50 and sieving is ongoing for the remaining 5,967 k's with 9,482,052 k/n pairs remaining at current sieve depth of ~650M. Currently 1 k/n pair is removed every 0.14 seconds and optimal sieve depth si approximately 72 seconds per k/n pair at most, and likely less. I've removed every k mod 3 = 2 before sieving aswell every even k has been excluded from the beginning. As soon as I complete this task I hope to be able to find a way to resume the sieving of Sierpinski base 63 without consuming the entire amount of availeable RAM :smile:

Take care everyone.

Kenneth!

Ps. Riesel base 3, k=3677878 is at n=510000 and continuing :smile:[/quote]


Thanks for the update.

I must warn you: Base 25 is BY FAR the most difficult base to tackle < 32 except for bases 3, 7, & 15. (In other words, it's the most difficult that is not b=2^q-1. There are going to be far more base 5 primes that you can remove then what you have shown. The base 5 project has an entire list of base 5 primes found up to their current test limit, including very small primes starting from n=1.

Please go back to when Willem was testing Riesel base 25 and read the entire discussion all the way through before starting. After reading that, I can provide you with the link to all of the base 5 project's Sierp base 5 primes.

I'm confused as to why you are sieving starting from n=50. Wouldn't it be much easier to test to n=1000 first and then sieve those k's? A lot of k's could be quickly eliminated by PFGW that wouldn't have to be sieved.

Kenneth, I've asked this before. I would much rather that you choose easier bases. There are many of them out there < 100 (or 200). I was thinking of starting Sierp base 25 myself in the next few months, due to its difficulty in coordinating primes already found with the base 5 project and only searching k's that they are not. Also, as you've noticed, there are many "converted" base 25 k's that they have either found prime for or are, in effect, searching. The balancing on base 25 is extremely tricky.

I'm a little confused. How is there a memory error on base 63 now but there wasn't when you started the sieving? I'm just curious. When the sieving started, I would think that MORE memory would have been required. Have you shut your machine down and then turned it back on to clear out any "memory leaks"? The memory needed should go down as the sieving progresses.

Is there anything that you can send us on base 63?


Gary

[QUOTE=gd_barnes;168192]Thanks for the update.

I must warn you: Base 25 is BY FAR the most difficult base to tackle < 32 except for bases 3, 7, & 15. (In other words, it's the most difficult that is not b=2^q-1. There are going to be far more base 5 primes that you can remove then what you have shown. The base 5 project has an entire list of base 5 primes found up to their current test limit, including very small primes starting from n=1.

Please go back to when Willem was testing Riesel base 25 and read the entire discussion all the way through before starting. After reading that, I can provide you with the link to all of the base 5 project's Sierp base 5 primes.

I'm confused as to why you are sieving starting from n=50. Wouldn't it be much easier to test to n=1000 first and then sieve those k's? A lot of k's could be quickly eliminated by PFGW that wouldn't have to be sieved.

Kenneth, I've asked this before. I would much rather that you choose easier bases. There are many of them out there < 100 (or 200). I was thinking of starting Sierp base 25 myself in the next few months, due to its difficulty in coordinating primes already found with the base 5 project and only searching k's that they are not. Also, as you've noticed, there are many "converted" base 25 k's that they have either found prime for or are, in effect, searching. The balancing on base 25 is extremely tricky.

I'm a little confused. How is there a memory error on base 63 now but there wasn't when you started the sieving? I'm just curious. When the sieving started, I would think that MORE memory would have been required. Have you shut your machine down and then turned it back on to clear out any "memory leaks"? The memory needed should go down as the sieving progresses.

Is there anything that you can send us on base 63?


Gary[/QUOTE]

Well at the moment I'm sieving from n=2049-32168 n for Sierp base 25, with 656 k's remaining.

Regarding base 63, it took up all my memory to sieve all k's for the entire n=1001-25000 range, so it is faster aswell easier to manage by breaking it up in smaller pieces :smile: Also is still working on this base, though it is idle for the next five days, but then the sieving pf the n=1001-5000 for all k's will begin...

In stead of reading through the entire Riesel base 25 discussion I would much more prefer to be shown the way to the sierp base 5 primes, in order for me to remove even further k's that can be multiplied by 5 and not exceed the kmax, if in fact further k's can be removed.

Also I only tested to n=50 at the start, to save testing times, since I was using the base=3 script that Michaf created back then, to start up the base, so a lot of k's with no primes ever, would then have needed to be tested to n=1000, which was several extra days of work, which I didn't bother to put up on the startup phase :smile:

Kenneth!

I found the primelist for sierpinski base=5, now please tell me how to convert a k*5^n+1 prime to a k*25^+1

Is it like I think, that I have to either divide or multiply the k value and then keep the n-value the same?

Kenneth