Releasing S35 due to lack of activity and response over 7 months.

Does R51, with k between 534 and 24308 up to n=50k sound good for a core 2 duo?

[QUOTE=firejuggler;273101]does R51, with k between 534 and 48920 up to n=50k sound good for a core 2 duo?[/QUOTE]

It seems borderline so it depends on your patience more than anything. The base is only searched to n=10K and it would be searching 83 k's from n=10K-50K. To determine for sure, I would suggest sieving them to some smallish limit like P=1G and then run a test at n=35K (~60% of the n-range between n=10K & 50K). Multiply the test time by the # of tests in the sieve file. Then subtract about 25% to allow for candidates removed from additional sieving and from k's found prime throughout the search. That will be your total CPU time. Then divide that by the # of cores and you should have a rough estimate as to how long it will take.

My recommendation for new searchers: Don't do anything that will take you longer than one month in calendar time to begin with regardless of your # of cores. Obviously if you have more cores, you can reserve and do more in a month. Even preferable is 2-3 weeks, which is why I suggested the bases with 3-4 k's remaining at n=25K. In other words, get your feet wet with small tasks to see if it is something that you'll be interested in keeping your machines on for an extended period of time.

The boredom factor is big here. It's very easy to get bored with your machines just crunching away while you check them from time to time for several months.

'cap' k divided by 2, up to 24308

[QUOTE=firejuggler;273101]Does R51, with k between 534 and 24308 up to n=50k sound good for a core 2 duo?[/QUOTE]

I would suggest taking a larger range to n=25000. You should be able to take the first 10000 k from n=10000 to n=25000 in about two months with two cores.

WARNING: 144400*51^n-1 has algebraic factors.
WARNING: 409600*51^n-1 has algebraic factors.
WARNING: 656100*51^n-1 has algebraic factors.
WARNING: 1416100*51^n-1 has algebraic factors.
WARNING: 5253264*51^n-1 has algebraic factors.
that's 5 k less

edit :i keep my reservation from 5xx <k< 24xxx with 10k<50k

firejuggler,

These k's cannot be removed, what you are seeing is the fact that these k's are squares.

380[SUP]2[/SUP]=144400
640[SUP]2[/SUP]=409600
etc.

This means that some of the n's for those k's can be removed, in this case where n==0 mod 2.

There are two methods to go about removing occurrences like these.

1) rogue modified srsieve and sresieve to remove these, this can be [URL="http://www.mersenneforum.org/showpost.php?p=267288&postcount=4"]seen here[/URL] (note Win64 builds already available).

2) The other method is by using Mini-Geek's and Batalov's hiddenpowers with removal script which can be found [URL="http://www.mersenneforum.org/showpost.php?p=209045&postcount=29"]here[/URL].

[QUOTE=firejuggler;273109]'cap' k divided by 2, up to 24308[/QUOTE]

How long will this take you?

What Rogue suggested is a better idea than what you are planning on but I will reiterate again, I would much prefer that you choose a large base with 3 or 4 k's remaining at n=25K and search it to n=50K until you understand the programs and processes a little more. Bases < 256 have mostly been extensively searched already.

Why do you want to do so much work to begin with? Get your feet wet with a small reservation that will take one or two weeks. It will also make for much less work for the admins. Managing bases in pieces is a lot more effort.


Gary

considering 71 second for a medium k@35000, 49 days for the full run. Meaning an average of 36 day if the sequence removal rate is as excepted.

[QUOTE=rogue;273110]I would suggest taking a larger range to n=25000. You should be able to take the first 10000 k from n=10000 to n=25000 in about two months with two cores.[/QUOTE]

I meant the first 1000 k.

[QUOTE=Mathew;273118]There are two methods to go about removing occurrences like these.

1) rogue modified srsieve and sresieve to remove these, this can be [URL="http://www.mersenneforum.org/showpost.php?p=267288&postcount=4"]seen here[/URL] (note Win64 builds already available).

2) The other method is by using Mini-Geek's and Batalov's hiddenpowers with removal script which can be found [URL="http://www.mersenneforum.org/showpost.php?p=209045&postcount=29"]here[/URL].[/QUOTE]

The modified srsieve will find algebraic factors that the hiddenpowers script misses. There are multiple categories of algebraic factorizations. IIRC, hiddenpowers only finds two of them. The nice thing about srsieve is that you won't need to run the hiddenpowers script at all and it will tell you which algebraic factorizations have been removed.