Riesel base 3 reservations/statuses/primes

[MisterBitcoin]

Reserving >63G up to n=750K.



Also considering to take a bit from n=100K-250K but only a range of 0,1G to see if I am able to quickly sieve and test it. So will only reserve it if such a small range makes sense.

[rebirther]

R3 tested to n=1.07M (1M-1.07M) (0-2.147G)

15 primes found, 561 remain in this range

Results emailed, Base released

[MisterBitcoin]

A bit ago I started to do some little work on an experiment that I intended to only give it 2-3 days; welp I let it run for almost 6 months. A lot of time honestly.

So anyway I took a range of 0,1G because taking an entire 1G range would have taken much longer, it is now at n=240K (must say its poorly sieved tho, because it was just an test) with 95 primes found.

Regardless I think that would suit me in the near future regarding the proposed goals 9 and 10 for 2023. So I would like to reserve the range 62.9G up to 63G up to n=250K.

If you are fine with me working on 100M ranges I could start to make my sieve files now and reserve a few ranges. (Expecting about 3G in total, so quite some amount of work) With an deeper sieved file I think it should be possible to reach 200K in about 3-4 months. I can run up to 8 such small ranges since; so far I noticed that single threated llr is faster than multi-threats.




About above 63G, the range reached 655K. I found 5 primes so far. That makes 28 k´s remain in that range. Still going up to n=750K.

63047009554*3^509964-1 is prime! (243326 decimal digits, P = 3) Time : 545.431 sec.
63063431926*3^515258-1 is prime! (245852 decimal digits, P = 3) Time : 672.682 sec.
63018423542*3^616923-1 is prime! (294358 decimal digits, P = 3) Time : 871.867 sec.
63041827754*3^617337-1 is prime! (294556 decimal digits, P = 4) Time : 884.167 sec.
63000520274*3^653328-1 is prime! (311728 decimal digits, P = 3) Time : 861.871 sec.

[gd_barnes]

Originally Posted by MisterBitcoin

If you are fine with me working on 100M ranges I could start to make my sieve files now and reserve a few ranges. (Expecting about 3G in total, so quite some amount of work) With an deeper sieved file I think it should be possible to reach 200K in about 3-4 months. I can run up to 8 such small ranges since; so far I noticed that single threaded llr is faster than multi-threads.<snip>

Having 5 or 10 different small reservations with different search depths for each would be a mess to show on the pages.

How about you reserve the entire k=62G-63G range, sieve the whole thing at once, and then test most of it at once? After you are done sieving, then you can split it up by k-range among your various machines. When doing interim reporting, you can report the lowest tested n-value that any of your machines has reached and send all of the primes from all of your machines that are below that n-value at that point in time. After finishing that k=1G range, then you can report the remainder of the primes and move on to the next k=1G range.

I realize that takes a long time but it takes the same amount of time as doing 10 different pieces of it at different times and having to report all of those pieces that are at different test depths at different times.

Please no smaller than k=1G ranges for sieving and testing. If we split the whole base in 100M pieces, we'd have 631 different reservations and search depths.

[MisterBitcoin]

Originally Posted by gd_barnes

How about you reserve the entire k=62G-63G range, sieve the whole thing at once, and then test most of it at once? After you are done sieving, then you can split it up by k-range among your various machines. When doing interim reporting, you can report the lowest tested n-value that any of your machines has reached and send all of the primes from all of your machines that are below that n-value at that point in time. After finishing that k=1G range, then you can report the remainder of the primes and move on to the next k=1G range.

I agree on that and reserve the entire k-range 62G-63G up to n=250K since other ranges are already at that n-value.
In total I would be able to take roughly 4G in total across my machines; though depends if this will help with the next years goals.

[gd_barnes]

Originally Posted by MisterBitcoin

I agree on that and reserve the entire k-range 62G-63G up to n=250K since other ranges are already at that n-value.
In total I would be able to take roughly 4G in total across my machines; though depends if this will help with the next years goals.

If we have base 3 goals, it would be for the lower k-ranges on R3 and S3. So no, k>60G for R3 would not be in the goals.

[MisterBitcoin]

Originally Posted by gd_barnes

If we have base 3 goals, it would be for the lower k-ranges on R3 and S3. So no, k>60G for R3 would not be in the goals.

Yeah, I am just taking the range because I already started in 62,9G. Following ranges would be k>10G of course.

[KEP]

I'm going to reserve R3 k=10G-11G from n=100K to n=250K, sieving using srsieve2 version 1.6.2 - 1668 sequences reserved. Sieve will be done to p=210G and complete in 4-5 days. After that LLR testing on 12 cores.