Reply With QuoteWell, I guess it's obvious, but anyways. I mean, for a single factor.
Who will be the first to score 100,000?
Well, this might be any of us. But, I'd vote for one of the P-1 guys.
A more tricky question is; "When will this happen?"
For that, one has to guess both PRP and sieve progress.
My humble guess is December 30th, 2004.
And, here's a quick cheat table that covers the possible n/p combinations in 50,000 n steps. The limits are; current n+200k at one hand (i.e. if PRP stops today), and current 90% sieve point on the other hand (i.e. if sieving stops today).
I'd be happy to hear your guesses.
Last edited by Nuri; 09-19-2004 at 06:07 PM.
Well, I guess it's obvious, but anyways. I mean, for a single factor.
I think vjs has a pretty good chance with the 7.5m, 14.1 bias. It's just over 100k, and not too far away n wise. Plus, the bias is 10.6 right now. Not too too far away.
Thanks for the vote of confidance royanee,
I know you can vote fore yourself in an election but I'm not sure of it in DC.
Truth of the matter is we will need a n>7.45m factor to enter the main active window, so n>7.25m in prp, to get a 100k score. By that time I would think the main sieve effort would be past or around 2^49 so it may just be luck of the draw.
I'm hoping but not optimistic,... yet.
Nuri's table looks like it has some pretty good info but I'm not sure what it represents exactly, to score 100k? Not sure why the 90% sieve point is decreasing or I'd say it's projected 90% sieve?
Yeah, I didn't understand the table either. Nuri?Originally posted by vjs
Nuri's table looks like it has some pretty good info but I'm not sure what it represents exactly, to score 100k? Not sure why the 90% sieve point is decreasing or I'd say it's projected 90% sieve?
Upon futher reflection,
584000-586000 Nuri
606000-608000 Nuri
628000-630000 Nuri
and
Looks like what it will take go get a 100K score...current n+200k at one hand (i.e. if PRP stops today), and current 90% sieve point on the other hand (i.e. if sieving stops today)
Of course to get a 100K score the factor must enter the prp+200k window.
Good example is one of mine
563.834T 21181^7581644 will score 101280.964 with bias of 14.10,
However someone could beat me to 100k with a smaller n (b/c it would score first)and higher T, but they would have to fight the bias from a higher T which is related to the 90% sieve point, which is what nuri sis showing assuming the 90% sieve point stays constant at ~426 which is won't.
I think this is what he is gettting at.
So if this is the case I'd say someone will beat me to a 100K score unless I find a few more factors around 7.4m.
UPDATE:
:shocked:
752.772T 4847 7192527 121696.059 18.82
looks like its going to be nuri if I understand this correctly.
Last edited by vjs; 09-24-2004 at 04:41 PM.
remenber once the 90
% sieve reaches 450 its going to jump intrantly up ot around 485. This is going to raise some scores because of the inreased bias. Some factors may go up as much as 5,000 points or more with this.
Sorry about the confusion regarding the table. I mistakenly wrote G instead of T.
You can read it like, n+200k / p 90% pairs that would lead to 100,000.
Let's assume we're currently 7,000,000 for the PRP effort. Then, the highest possible n within the active window would be 7,200,000. That is the first column. The first column increases in 50,000 n steps in the table. And the second column gives the lowest possible 90% sieve points that lead to a score of 100,000 for the n+200k value corresponding to it.
At the time I prepared the table, PRP effort was at around 6,950,000. Hence, the lowest theoretical limit for n would have been 7,150,000. Let's assume for a moment that PRP suddenly stopped at that date, but sieve continued. Then, the active window would cease to move upwards. But the factors which have a p larger than 90% sieve point and that are within the active window range would see their points increase. I know, this scenario is highly unlikely, but as I said, it's the theoretical limit.
The lowest row of the table corresponds to the other end, i.e. if for some reason sieving stoped suddenly, but PRP continued. And the rows in between are the scenarios corresponing to various degrees of relative progress at PRP and sieve.
Currenly we have 7,179,272 as our n+200k, and 428.859T as our 90% sieve point. So, theoretically, speaking, the highest possible score that someone can get from a factor is 69075.7113 [=(7.179272^2)*125*(428.859/40)] at this moment. If, let's say, PRP suddenly stopped, then, we would have to wait until 90% sieve point reaches at least 620.85T to see a score of 100,000 for a single factor.
Let's say, you are looking at 8,000,000 / 500T pair from the table. If you say we'll reach 100,000 score when n+200k and 90% p reaches those numbers, then you are implicitly saying that, n will move upward by 1,020,718 (=8,000,000-6,979,272), and 90% sieve point will move by 71.141T (500-428.859) from now until that time comes. Is this scenario likely? Again, not very much. 90% sieving point will probably reach 500T within 40 days or so, roughly around the end of October. At that time, PRP would probably be at around 7,250,000, way below 8,000,000.
Ok, too many bla bla bla....
How to calculate the date when a score reaches 100,000?
Here's my method:
Currently we have 7,179,272 as our n+200k, and 428.859T as our 90% sieve point.
Let's assume PRP will continue to move upwards by, say 7,000 per day on the average, and sieve 90% will average 1.15T per day.
Then, 100,000 will be reached 87 days from now. At that date, n+200k will be at 7,788,272 (=7,179,272+87*7000), and 90% sieve point will be at 528.91T (=428.859+87*1.15). And, 7.788272^2*125*528.91/40 =100,257.
Today is September 25th, so 87 days from now would be December 21st.
Of course, the progress on either fronts will not be linear. One has to correct for, say, if there is a FFT boundary close by (or any other factor what would increase/decrease PRP progress), and/or the bunch of completed sieve ranges between 450T and 500T (or any other factor what would increase/decrease sieve progress), etc.
As a result, it seems like we'll see a 100,000 score a few weeks earlier than my first guess of December 30th.
And, by the way,
584000-586000 Nuri
606000-608000 Nuri
628000-630000 Nuri
etc. are just coincidences. A couple of weeks ago, I convinced three of my friends to run sieve client on their PCs, but I was not sure of their reliability and speed. I chose those ranges for their machines. If they progress slowly, they'll have time until sieve catches up. Ranges are a bit apart from each other, because if they did nothing at all, I'll be able to take over at some point, and still catch up with the sieve.
My guess is november 20th by engracio (beginners luck
I'm really going to vote for nuri on this one, the only thing holding nuri back from getting the first 100k is his choices.
So nuri do you play blackjack I'd love for you to sit to my right for a few hands, we could dedicate the money to a new server.
Looks like it may actually be one of the P-1 people, considering they have the new code for p-1 it looks like they will be producing alot more factors.
Wishing for a sse2 machine
As far as the timing of 100,000 score is concerned, it looks like the time will come much earlier than we predicted. Mainly because we have a faster PRP client now, and sieve output (still?) seems strong.
BTW, I will not try to punch "the" 100,000 first.
The new client definitely speeds up things. I vote November 17th - my birthday.
I'll go with Nov 24 and.... what the heck, me!
Largest scores
p (T) k n Score Factor found Score changed Score was Score could be Reqd bias
666.453T 33661 7410744 86615.034 Thu 18-Mar-2004 Sat 23-Oct-2004 6.665 114378.093 16.66 DeadJDona (death)
503.809T 55459 7364938 85399.283 Fri 16-Jul-2004 Mon 18-Oct-2004 5.038 vjs
i'm on a top second time. I got 40k score first... and now 666 range gives me a highest position...
maybe this is offtop but I just want to brag a little
and well, 25 november is my bet
wbr, Me. Dead J. Dona \
Cheers Dude,
Now lets see how long you stay on top.
Death *-( -------
560.540T 10223 7489277 5.605 Tue 12-Oct-2004 98250.887 14.01 <-- coming soon from VJS to a DC project near you
also the sequel a potential 100K score
563.834T 21181 7581644 5.638 Wed 21-Jul-2004 101280.964 14.10
Looks like you and I will be :|punch|: for the first 100K
not sure your
666.585T 21181 7567532 6.666 Fri 26-Mar-2004 119292.848 16.66
May make 100k first
Let's guesstimate what score 563.834T 21181 7581644 5.638 Wed 21-Jul-2004 101280.964 14.10 will end up with.Originally posted by vjs
also the sequel a potential 100K score
563.834T 21181 7581644 5.638 Wed 21-Jul-2004 101280.964 14.10
Looks like you and I will be :|punch|: for the first 100K
not sure your
666.585T 21181 7567532 6.666 Fri 26-Mar-2004 119292.848 16.66
May make 100k first
First-pass is currently at 7239226, and daily change in n is 9480. If that daily change stays constant, min n will pass 7581644 in 36 days, whereafter the score will not increase anymore.
Where will 90% sieve point be in 36 days? Currently it's at 507.7T. I have not calculated cumulative sieve speed recently, so a couple of alternatives:
1000G/day: It'll not increase any further than 97664.
1200G/day: It'll not increase any further than 98958
1361G/day: It'll not increase any further than 99999.
1559.27G/day: It'll get all it's potential and score 101280.964. The score will not increase anymore for any faster G per day average.
I am not sure if the increase in 90% sieve point will be higher than 1361G/day.
Still, it's a good candidate. It'll be interesting to see at what score it'll end up with.
With the current relative speeds, any factor where p>560T and 7560000 < n < 7810000 looks like a good candidate to be the first to score 100,000+ at around the first week of December.
As you both pointed out, currently, the factors that fits the above criteria are:
666.585.485.406.991 21181 7567532 3419 44 (death)
563.833.698.024.767 21181 7581644 5202 81 (vjs)
I'm pretty sure stronger candidates will start to emerge as we come closer to the "day".
Other current, though less likely, 100,000 candidates are:
790.587.618.245.767 28433 7260817 1608 0
666.452.508.787.073 33661 7410744 3419 44
628.849.406.072.687 28433 7524673 1577 0
1.126.101.624.584.750 55459 7880458 3259 0
628.023.875.662.639 33661 7900440 1577 0
629.496.225.340.819 27653 7931553 1577 0
562.880.130.980.129 33661 7987320 5202 81
Looks like Mike is going for the 100K score with his p-1 range.
750000-7550000 since any p-1 score below my 7581644 will beat me.
Have to say my gut feel is that one of the P-1 guys will be the first to the 100k, so I figued I'd give it a try.Looks like Mike is going for the 100K score with his p-1 range.
Good luck Mike.
And all others too, of course.
I think I've already lost it, pxl97 just got in there with a
This is a very possible candidate.Code:12153.158T 21181 7555268 35.000 Tue 02-Nov-2004 2167898.401 303.83 pixl97
Sweet, I've watched this thread for a long time, and never though I be in the running... many thanks to Priime95 and the SSE2 enabled P-1 client.
That would require a 14.01 bias for pixl97 to have a 100k. I think however that it is more likely that the bias won't increase that much unless dmbrubac finishes up that low range soon (I'm not complaining, please don't be mad. I thank you so much for dedicating so much to the sieve!
Sure, enjoy... Why would anyone be mad? And even if so, simply ignore him.Originally posted by royanee
(please don't be mad if I'm wrong, just guessing for fun!).
In the end, it's the mutual fun we get that makes us come together here (at least, that's the case for me...
0-560T has to be sieved except 56T - so basically everyone else is working on this "goal" as well.Originally posted by royanee
That would require a 14.01 bias for pixl97 to have a 100k. I think however that it is more likely that the bias won't increase that much unless dmbrubac finishes up that low range soon
But:
Right now, we're maybe at 83% (it's nearer to 85% than to 80% according to MikeH's page), which means that we need 7% = 39.2T
I'm not sure about the current progress, but I guess we're near 1T/day, which means that more than a month is still missing.
I couldn't get mad at that royanee! Maybe I'm trying to control the bias so I'm the first to 100k.... No, I'm not. I didn't even realize incomplete sieve ranges influenced the bias until I read your post. I gave a chunk of the range to keroberts to help get it out the door. Once its done I will probably go back to P-1 and PRP exclusively.Originally posted by royanee
(I'm not complaining, please don't be mad. I thank you so much for dedicating so much to the sieve!
Well I just got my first shot in
But as you've pointed out for pixl97's, I need a bias of 14.03, which just isn't going to happen for at least another month, which is way too late.Code:794.526T 4847 7550823 7.945 Wed 03-Nov-2004 141561.964 19.86 MikeH
Maybe I should have sat down and worked out the required bias before selecting the P-1 range.
I'm, very close <7days from finishing a 10T range at 490T so that should if I understand the bias correctly increase the 90% bias by 10T. Let hope mike updates my range with due dillegence
Also to Death you were stuck at the top with that 666T for quite sometime, the view is pretty good....
( 268.31G) : 480500-490000 VJS (ETA: Mid Nov)
It counts factors as you submit them. See here.
Yeah, I like the way it looks.Originally posted by vjs
Also to Death you were stuck at the top with that 666T for quite sometime, the view is pretty good....
wbr, Me. Dead J. Dona \
As ceselb points out, the 90% calculation is automatic - it's done from the gap analysis, so it doesn't matter whether I update my copy of the reservations or not. Which is lucky because I'll be out of town next week - my first trip to the southern hemisphere in about 15 yearsLet hope mike updates my range with due diligence when that happens.
Mike,
This is perfect... the way things are going lately, you will be out of town when we need a major dat file update, b/c we found a new prime
Hummmm.....
Hold on a second, this would be a good thing
Isn't this the way life would work, anyway if it does happen I'll inform people on how to reduce the current dat to 10K or upload one as a message it if it's O.K. with you.
Last prime that was found I was out of town, so who knows.Hold on a second, this would be a good thing
<spanishInquisition>Hm, interesting...Originally posted by MikeH
Last prime that was found I was out of town, so who knows.
So you say that you're out of town when a prime is found. Can you prove that it's not YOU who has put it there in the first place?</spanishInquisition>
NOBODY expects the Spanish Inquisition!
Yes but the logic of Pinky and the Brain is flawless
Brain: Perhaps, he is finding them all of the time, and just not telling us
Pinky: why would he do that Brain
Brain: Because Pinky, he is TRYING TO TAKE OVER THE WORLD
WEll what do you guys think, has someone taken it???
Just to let all know I submitted a nice 7.55m a couple days ago, around 560T it would probably make a very low 100K score.
whell, whenn scoress comm backk we'll see
wbr, Me. Dead J. Dona \
Mike,
I just trying to figure out a way of keeping the 100K score for himself.
I'm going to cry foul if he wins
for my first factor.. this is pretty darn good.Code:p k n Score Factor found Score changed Score was Score could be Reqd bias 605.507T 28433 7412233 90070.131 Sat 13-Nov-2004 103960.191 15.14
Posting Permissions
