Sierp base 6 - team drive #3

[Mini-Geek]

Originally Posted by mdettweiler

Not much, but we've had some others helping out on 1400 lately, so things have been moving much more quickly. Particularly, I see one core each of an Ivy Bridge and a Haswell on there - those carry a lot of weight.

I have all 4 cores of my Haswell alternating between three servers in roughly equal proportions, so it's around 1.3 cores of my Haswell on there - maybe 28-32 tests per day if I'm not using the computer too much (yesterday I did 28, exactly 7 per core).

Originally Posted by gd_barnes

Another thought: After Sierp base 6 is done here to n=1M, what would everyone think of crunching Riesel base 6 on port 1400 until our new base 2/4 drive is done sieving? R6 currently has 2 k's remaining and is at n=1.5M with no reservations and is sieved thru n=2M. The tests would be ~2-3 times as long as S6 at n=1M but there are < 1/5th as many candidates so the n-ranges would process more quickly. We would not have to test the entire rest of the range thru n=2M in the highly likely event that the base 2/4 effort is done sieving first, although we could ask people what they think at that point.

I like this plan.

[mdettweiler]

Originally Posted by gd_barnes

Another thought: After Sierp base 6 is done here to n=1M, what would everyone think of crunching Riesel base 6 on port 1400 until our new base 2/4 drive is done sieving? R6 currently has 2 k's remaining and is at n=1.5M with no reservations and is sieved thru n=2M. The tests would be ~2-3 times as long as S6 at n=1M but there are < 1/5th as many candidates so the n-ranges would process more quickly. We would not have to test the entire rest of the range thru n=2M in the highly likely event that the base 2/4 effort is done sieving first, although we could ask people what they think at that point.

Originally Posted by Mini-Geek

I like this plan.

Seconded - I'd be game for this as well. R6 hasn't seen much attention lately, but it's a really interesting base given that there's only two k's left - IIRC they're the two lowest-weight of the original batch of 8, but from the beginning on that base we've had an incredibly better-than-expected run of primes. I keep wondering (in a totally hypothetical unsubstantiated way) if there's something yet to be discovered about this base/range to corroborate that...in any event it should be interesting to see if the run continues with these last two k's.

Regardless, I'll be keeping my cores on port 1400 for the foreseeable future. I don't have too much time these days to micromanage personal efforts, so I'm glad to run whatever we're doing lately as a team effort.

[gd_barnes]

Very good. When we get past n=~995K on S6, I'll load up R6 for n=1.5M-1.6M into port 1400 to start with and will continue loading more later on until we're done with the base 2/4 sieving effort.

[Puzzle-Peter]

Originally Posted by gd_barnes

Another thought: After Sierp base 6 is done here to n=1M, what would everyone think of crunching Riesel base 6 on port 1400 until our new base 2/4 drive is done sieving? R6 currently has 2 k's remaining and is at n=1.5M with no reservations and is sieved thru n=2M. The tests would be ~2-3 times as long as S6 at n=1M but there are < 1/5th as many candidates so the n-ranges would process more quickly. We would not have to test the entire rest of the range thru n=2M in the highly likely event that the base 2/4 effort is done sieving first, although we could ask people what they think at that point.

I thought about doing some more R6 work. What keeps me from doing so ATM is that I want to finish one or two of my longtime reservations first. But there are at least 4 cores I will switch to R6 when it's online.

[gd_barnes]

Originally Posted by Puzzle-Peter

I thought about doing some more R6 work. What keeps me from doing so ATM is that I want to finish one or two of my longtime reservations first. But there are at least 4 cores I will switch to R6 when it's online.

S6 will be done to n=1M in port 1400 within ~9-10 days with current resources. Does that work for you? Or if you want to help with S6, it will be done sooner. Either way, I'll keep a daily eye on the port and will load up R6 when S6 is at n>995K. Once R6 is loaded, if we get more resources hitting it, I may load a bigger range of R6 than just n=1.5M-1.6M. Also in the future if the consensous is that people want to complete R6 to n=2M before we start the base 2/4 drive, we can do that too.

In the mean time, shortly we're going to need to start a sieving effort for both sides of the team efforts for base 16 n=500K-1M. R16 is already complete to n=500K and S16 will be there within a couple of days.

...busy times ahead for the project

[Puzzle-Peter]

Originally Posted by gd_barnes

S6 will be done to n=1M in port 1400 within ~9-10 days with current resources. Does that work for you?

I was thinking weeks or months rather than days before getting a manual reservation LOL. But those 4 cores will be free in about a week, so that fits quite nicely.

[Lennart]

I will add about 20 core to clean up base 6.

Lennart

[gd_barnes]

Originally Posted by Lennart

I will add about 20 core to clean up base 6.

Lennart

Excellent! I'll go ahead and add all of Riesel base 6 for n=1.5M-2M shortly.

[gd_barnes]

Wow. Port 1400 is already handing out R6 work. Although we still have a little over 100 tests left to finish up on S6, that was very fast to get to this point!

[gd_barnes]

I have moved posts about the speed difference between LLR and PFGW to a new thread here.

[gd_barnes]

All pairs have been processed and the drive is now officially complete to n=1M! Thank you everyone for your contributions.

In a manner similar to both sides of base 16, we will keep the drive intact and eventually begin sieving a higher n-range.