LLRnet servers needed

[gd_barnes]

I just want to post here where I think we will need LLRnet servers now and in the near future:

NPLB:
2 for 400<k<=1001 for n=333.3K-600K
1 for 400<k<=1001 for n=260K-320K
2 for 300<k<400 for n=333.3K-600K in early March (I would like to do the next rally on this one; after the rally 1 server may be sufficient)
1 for 300<k<400 for n=260K-333.3K in early March (smaller # of k's and n-range; shouldn't take too long)
1 for 300<k<400 for n=600K-1M in late March (will only be the few k's that have already been searched to n=600K)
1 for 300<k<400 for n=100K-260K double check

CRUS (Conjectures 'R Us):
1 for Riesel base 16 (currently @ n=~430K base 2; data available up to n=800K base 2)
1 for Sierp base 16 (currently @ n=~440K base 2; data available up to n=800K base 2)
1 for Sierp base 6
1 for another base to be determined in the near future
1 for Sierp base 4 for one k for n=1.85M-2M base 2, i.e. the mini-drive (temporary; shouldn't take long but is a nuisance for a single person to easily knock out due to testing times)


Total of 8 servers for NPLB and 5 for CRUS. Later, I would like to add possibly 2-3 more servers for CRUS when we get testing up past n=200K on several of the bases, especially the lower ones. The best example would be Sierp base 6 where the # of k's remaining will dictate that we probably need 2 servers.

This is why I am thinking that we need all the servers that we can get.

Note to folks with some or mostly non-connected machines: Both projects will always have manual reservations available for all efforts.

I just pulled all of this off of the cuff, so to speak. Input and opinions are welcome.


Gary

[mdettweiler]

I'm hoping to be able to get an older computer set up as an always-on server sometime soon (i.e. within a month or two), so once I've got that set up, I should be able to run one server. (No more than one, though, since it's only got 128MB of RAM.)

[mdettweiler]

Hmm, let's see how we're doing with each one of those needs:

NPLB:
2 for 400<k<=1001 for n=333.3K-600K

Got one so far, still need one more.... (see below)

1 for 400<k<=1001 for n=260K-320K

We've got two here right now; maybe after this range is done (looks like it won't be too long yet), IronBits's port 500 server that's currently on this range can be moved to the 333.3K-600K range for 400<k<1001?

2 for 300<k<400 for n=333.3K-600K in early March (I would like to do the next rally on this one; after the rally 1 server may be sufficient)

Hey Gary, maybe you could do one of the ones for this range with the new server you're working on getting set up right now?

1 for 300<k<400 for n=260K-333.3K in early March (smaller # of k's and n-range; shouldn't take too long)

Maybe either Gary or IronBits would be interested in doing this one?

1 for 300<k<400 for n=600K-1M in late March (will only be the few k's that have already been searched to n=600K)

Hmm...maybe Carlos will want to do this one, since the tests are bigger and thus it won't use as much RAM and bandwidth?

1 for 300<k<400 for n=100K-260K double check

If and when I can get my server up and running, I'd be glad to do this one...since I've sort of spearheaded this doublecheck effort anyway.

CRUS (Conjectures 'R Us):
1 for Riesel base 16 (currently @ n=~430K base 2; data available up to n=800K base 2)

Check.

1 for Sierp base 16 (currently @ n=~440K base 2; data available up to n=800K base 2)

Check.

1 for Sierp base 6

Hmm...any takers?

1 for another base to be determined in the near future

Any takers for this one, too?

1 for Sierp base 4 for one k for n=1.85M-2M base 2, i.e. the mini-drive (temporary; shouldn't take long but is a nuisance for a single person to easily knock out due to testing times)

This one might be nice for Carlos, too, becuase it will use less RAM and bandwidth due to the huge size of the numbers involved.

Anyway, just suggestions.

[kar_bon]

please keep also in mind that all these servers has to be maintained (in a timely-manner) and evaluated say k/n-pairs statistics for user and primes!
so for NPLB all resultfiles from the servers goes to me.

[mdettweiler]

Originally Posted by kar_bon

please keep also in mind that all these servers has to be maintained (in a timely-manner) and evaluated say k/n-pairs statistics for user and primes!
so for NPLB all resultfiles from the servers goes to me.

As well as to me, too, so I can run the results files through my processing scripts to turn them into regular manual-LLR-format lresults.txt files.

[em99010pepe]

You are really crazy Gary!!!

I'll only host max of three servers.

[mdettweiler]

Originally Posted by em99010pepe

You are really crazy Gary!!!

I'll only host max of three servers.

Then I guess it's a good thing that we've got some more people interested in hosting LLRnet servers, too!

Hey, maybe we can set a world record for most LLRnet servers in one project?

[em99010pepe]

Originally Posted by Anonymous

Hey, maybe we can set a world record for most LLRnet servers in one project?

A think all those server are nonsense but who am I to judge this....

[mdettweiler]

Originally Posted by em99010pepe

A think all those server are nonsense but who am I to judge this....

Well, of course, we must keep in mind that even as we add more servers over time, old ones will be dropping away--such as the ones for 260K<n<320K on 400<k<1001, since that range is not far away from being completed.

[em99010pepe]

Yes but for the moment NPLB is still a small project with only a few members. We still don't have such a visibility to have 8 servers running, you will have the same members jumping from server to server...

[mdettweiler]

Originally Posted by em99010pepe

Yes but for the moment NPLB is still a small project with only a few members. We still don't have such a visibility to have 8 servers running, you will have the same members jumping from server to server...

You do have a point...maybe we should make sure that until we grow significantly, we don't have more than 5 unique servers running at a time (I say "unique servers" since some of the ones listed in Gary's post would be duplicates for load distribution reasons).