Bases 2 & 4 reservations/statuses/primes

[gd_barnes]

Originally Posted by mdettweiler

Ah, but what makes you think I have one coming up?

Also, I intended "profession" to include student status as well.

Easy peasy.

You've been very busy this time of year the last two years and are excessively (more than usual) busy this year indicating study and work for finals with this being your final/senior year. Also you're planning on moving late summer and are in the market for a new laptop for your new place with little room for such at your current place indicative of living with parents/brother/friends at the moment. Also you're available at odd hours like yours truly indicative of a student, having your own business, retirement, or just simply a slacker who doesn't work much (lol). You don't strike me as a slacker or of near retirement age. The only other reasonable possibility besides being a student is having your own business but other things you've said about your situation (that you might not want me to mention here) all but rule that out.

I think we need to set up a poll: What is Max's current life situation? rofl

[mdettweiler]

Originally Posted by gd_barnes

Easy peasy.

You've been very busy this time of year the last two years and are excessively (more than usual) busy this year indicating study and work for finals with this being your final/senior year. Also you're planning on moving late summer and are in the market for a new laptop for your new place with little room for such at your current place indicative of living with parents/brother/friends at the moment. Also you're available at odd hours like yours truly indicative of a student, having your own business, retirement, or just simply a slacker who doesn't work much (lol). You don't strike me as a slacker or of near retirement age. The only other reasonable possibility besides being a student is having your own business but other things you've said about your situation (that you might not want me to mention here) all but rule that out.

I think we need to set up a poll: What is Max's current life situation? rofl

Right...though I could just as easily be a college professor (tied to the same schedule as students) planning to retire after this semester and move to a retirement home with limited space (hence the laptop).

Though you did sort of already cover that contingency in the above analysis...

But hey, don't let me stop you from theorizing--it's funny to see what kinds of guesses you've come up with.

P.S.: should this discussion be split off into another thread? I seem to recall we have a place for these kinds of off-topic discussions somewhere...

[Jean Penné]

Originally Posted by gd_barnes

Jean,

Thanks for the status. As for k=84363, all other k's besides your k=60357 for the Riesel/Sierp odd/even-n conjectures are reserved by our PRPnet mini-drive. We have reached at least n=~1.35M with all of them. It would work best if you set up the PRPnet client on your machine and contributed to the drive. See the PRPnet mini drive for details.

Max,

With the more deeply sieved files for these, I might suggest deleting all candidates n>1.4M from the PRPnet server (excluding the even k's that I sieved) and replacing them with the better sieved files.


Gary

Gary,
I agree to release all my reservations (k = 84363 and, even, k = 60357) and contribute to the drive(s), but I cannot find any info to know how to do that, so, would you help me ?

Thank you by advance,
Jean

[gd_barnes]

I moved all of the recent posts that were more related to PRPnet info. and questions to the PRPnet mini drive thread at http://www.mersenneforum.org/showthread.php?t=12225.

[Puzzle-Peter]

Originally Posted by Jean Penné

I tested k = 9519 up to n = 2110696 base 2 ; no prime... I release it for now!

I'd like to pick up this one. Is there any benefit in sieving / testing speed when converting to base 2? The sieve files are base 4.

[mdettweiler]

Originally Posted by Puzzle-Peter

I'd like to pick up this one. Is there any benefit in sieving / testing speed when converting to base 2? The sieve files are base 4.

Short answer: testing yes, sieving no. LLR and PFGW both automatically convert to base 2 whenever possible for this reason.

Long answer:
There is a small advantage in testing speed (it used to be a huge advantage, but since about two years ago when PFGW and LLR were upgraded to perform more efficiently on non-base-2 numbers; now other bases are still a little slower than base 2, but not hugely so). There is, however, no difference when sieving; the srsieves handle all bases with equal aplomb, such that sieving n=500K-1M base 4 is identical speed-wise to sieving 1M-2M base 2.

Note that LLR automatically converts numbers with power-of-2 bases to base 2 when testing. That is, if you tell it to test 3*4^1000000-1, it will actually test 3*2^2000000-1. All screen and results file output will give the base 2 number, to avoid ambiguity as to whether the residual should be that of a PRP or LLR test (they are not cross-compatible). I'm pretty sure that PFGW is able to do a similar conversion and take advantage of base 2 speeds; though since it does a PRP test for everything, including base 2, it will produce the same (PRP) residual either way.

If you're using a PRPnet server to run such numbers, the server will handle the numbers in the original base; the client knows how to handle base conversions, so it will give LLR the number in base 2 (this is a little redundant since LLR can do the conversion internally, but it helps to avoid confusion on the client's part by ensuring that LLR gives a result for the same number it was given). The server will then be sent back to the server with the original base. Note that this does introduce a slight ambiguity in that there is nothing on the server end to distinguish that the residuals are LLR, not PRP as would normally be the case for non-base-2 numbers; for this reason, I personally prefer to pre-convert all such numbers to base 2 with a script of my own before loading them into a server. That's what I've done, for instance, with the work currently in port 1300. (In this case those are actually base 2 numbers, but were sieved in base 4 because that's an easy way to make srsieve include just even or odd n's. For even n's you use the same k in base 4, for odd n's you use 2*k in base 4.)

[Puzzle-Peter]

Thanks!

Jean's last pair was 9519 1055348 (base 4 as in the sieve file on his page). So - just to make sure - for optimum speed, I discard this pair and all pairs lower than that, multiply all the exponents by 2 and change header to base 2, right?

Reserving to the boredom threshold ;-)

[gd_barnes]

Originally Posted by Puzzle-Peter

I'd like to pick up this one. Is there any benefit in sieving / testing speed when converting to base 2? The sieve files are base 4.

Knocking this one out would be nice. It is the only k that I am aware of at CRUS that is remaining for 4 different bases: 2 even-n, 4, 16, & 256.

[Jean Penné]

Originally Posted by Puzzle-Peter

Thanks!

Jean's last pair was 9519 1055348 (base 4 as in the sieve file on his page). So - just to make sure - for optimum speed, I discard this pair and all pairs lower than that, multiply all the exponents by 2 and change header to base 2, right?

Reserving to the boredom threshold ;-)

Good luck with this k !!
Jean

[gd_barnes]

Originally Posted by Puzzle-Peter

Thanks!

Jean's last pair was 9519 1055348 (base 4 as in the sieve file on his page). So - just to make sure - for optimum speed, I discard this pair and all pairs lower than that, multiply all the exponents by 2 and change header to base 2, right?

Reserving to the boredom threshold ;-)

That is correct.

[Puzzle-Peter]

9519*2^n-1 has reached n=3M, no primes