View Single Post
Old 2020-09-20, 08:55   #1412
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

240018 Posts
Default

Quote:
Originally Posted by MisterBitcoin View Post
I know Gary doenst like taking single k´s from larger Bases, but i do have an question about two k´s from S25.

101958*25^n+1 and 207544*25^n+1 have both an very low number of tests; each around 4,5k tests between n=350-1M; with an higher sieve limit even a few hundred less.

I could get those down within a few months / weeks.


But of course my aim is not to bring them up to 1M, only prime counts. If you agree i would take both til i have a prime.
There are many bases with 2 k's remaining at n=300K or 400K. Why do you choose to take two k's only out of a base with so many k's remaining?

You say you will search them until you find a prime. What makes you think you would find a prime on them before you get bored with them? It is very likely that would happen. The chances of you finding a prime on them for n=350K-1M are very remote because of how low-weight they are.

I'll tell you what. Run a single test at n=770K (~65% of the n-range). See how long it takes. Multiply that by the number of tests that you have to get the total CPU time. Then divide that by your number of cores that you will be running on it 24 hours a day 7 days a week. Then report that time back here.

If you can honestly say that you will run all the cores that you reported here in the above calculation full time and you can provide a status report once a month I will reserve them for you.

The biggest problem that CRUS has is people reserving way more work than they realize. We've also had people reserve things "until they find a prime" not realizing how remote that chance is and how long the tests begin to take in rapid fashion.
gd_barnes is offline   Reply With Quote