Reserving S442 to n=100k (25-100k) for BOINC

Progress update
 

K4 S 467 at 395K

My Sandy Bridge is a bit idle at the moment, so I want to do another test of srbsieve and make sure that nothing in my ini files can cause a problem, so I'm going to reserve following Sierpinski bases as new to n=25K (using srbsieve):

S388
S457
S465
S477

I expect this test to be fast. I did not go all the way with S1023 using srbsieve, so currently the Haswell is LLR testing n>80xx. On a sidenote, R3 13G-21G is currently running for free on a Xeon E5 2670v3 that I got access to, so there is plenty ressources for this test.

[QUOTE=KEP;436091]I expect this test to be fast.[/QUOTE]

I expect this test to not be fast. Regardless of software and machines, that is a lot of k's! :smile: Combined it's likely to take at least 1/3rd as long as S1023...and you've already seen how long S1023 is taking just to get to n=8K.

[QUOTE=gd_barnes;436094]I expect this test to not be fast. Regardless of software and machines, that is a lot of k's! :smile: Combined it's likely to take at least 1/3rd as long as S1023...and you've already seen how long S1023 is taking just to get to n=8K.[/QUOTE]

Guess you are right. However compared to S1023 the conjectures is only ~1/8th the size and the base is only ~42% the size of S1023. Well I'm gonna stick with them till they are complete and then by that time hopefully S1023 will also be complete :smile:

S442 tested to n=100k (25-100k)

44 primes found, 50 remain

22879*442^25023+1
20017*442^27518+1
6036*442^29729+1
31402*442^29923+1
16891*442^31185+1
4798*442^32137+1
2574*442^32139+1
6099*442^33063+1
16506*442^33405+1
13111*442^33899+1
35038*442^36306+1
14304*442^37578+1
2377*442^38222+1
11121*442^38285+1
32751*442^39648+1
23002*442^39732+1
12309*442^40554+1
15691*442^40712+1
33393*442^40896+1
32883*442^41320+1
28527*442^43824+1
9921*442^45751+1
30003*442^47398+1
22132*442^47595+1
2551*442^48144+1
17917*442^51188+1
29332*442^52730+1
6546*442^55013+1
32532*442^58442+1
16524*442^58853+1
34611*442^59499+1
14473*442^62014+1
35589*442^62459+1
16956*442^62863+1
28863*442^66386+1
36346*442^68852+1
10383*442^78161+1
25993*442^79794+1
29460*442^80163+1
33519*442^80649+1
4524*442^80651+1
10096*442^89736+1
35539*442^94242+1
17076*442^96005+1

Results emailed - Base released

Reserving R355 to n=100k (25-100k) for BOINC

Reserving R255 to n=10k (2.5-10k) for BOINC

R287 reserving K25-K100
 

I wanna reserve K25 to K100 from R287. I just downloaded the data from CRUS and its now running in LLR.

As info: "Just" 222 k´s remaining.

[QUOTE=MisterBitcoin;437599]I wanna reserve K25 to K100 from R287. I just downloaded the data from CRUS and its now running in LLR.

As info: "Just" 222 k´s remaining.[/QUOTE]

Before we reserve this for you, can you please give us some info about your background in math and distributed computing as it relates to prime searching? How many cores do you plan to use? If more than one (we hope!), how do you plan to divide up the work amongst your cores? Do you have an approximate amount of time that you think that it will take?

I'm asking because you have not posted this in the correct place (I have now moved it to the correct place) and it is a huge amount of work. It is likely several CPU years on modern machines.

Thanks,
Gary

Not clear if he reserves all remaining k, from n=25k to n=100k. The electrician in me says he only reserves the k between 25 and 100 (which is k=28 only), hehe...

Edit: joking apart, there are ~485000 lines in the sieved file there, the lines at 25k may take few seconds (10 to 20), but the lines at 100k will take anything between 5 and 10 minutes each, depending on your CPU. That is (with an "average" of ~3 minutes per line per core) 485000*180/3600=~24000 core-hours, or 1000 days in a single core. Make it ~250 days for 4 cores, or about 7-8 months with a very good CPU.

@OP, do you have so much patience? (background in DC and math may not be relevant, I myself was "innocent" when I started... but I kept going...)