Stats Milestone - NEW - Page 4

VJS · 2008-07-28, 22:21

Humm...

A very good point, but you see Lars does have control of when and were all primes are found so...

LOL O.k. scratch #1

Joe O · 2008-07-29, 01:59

We still have the goal of 1st pass reaching 5M.

Quote:

Countdown to finish all n<5Mil: 10070 tests to go

This is reachable. We've just done 6959 tests in 115 days.

Sloth · 2008-07-29, 02:27

Quote:

Originally Posted by Joe O

We still have the goal of 1st pass reaching 5M.

This is reachable. We've just done 6959 tests in 115 days.

Who has the reservation for most of the remaining stuff under 5m? The LLRnet server is handing out 5.25-5.5M range. The reservation thread only has a couple ranges listed below 5M and they are already being worked on. I am guessing that the 4.2 to 4.8M range is set aside for PrimeGrid.

S.

themerrill · 2008-07-29, 02:32

This is relevant to our discussion, just a little long.
One thing I've wondered about prime grid's LLR:
They are obviously given a special selection of .txts. Which is great and likely worth it, with all the results they provide and all. When the crunch, do we get the results back little by little (PG confirms that a k/n pair is not prime and sends it in) or do we get big batches when the finish entire ranges of numbers.
Relevant because they just finished crunching 9500 k/n pairs where 4M<n<5M. If we are about to get that back in one lump, we are pretty much there in our n<5M countdown.

Edit - To answer sloth's question - Yes, most of them do appear to be tied up on PrimeGrid, which would confirm my question as mentioned above. They just finished, and should be sending it in soon, if they do one big send off. The remaining 500 or so are reserved by drakkar and should be done in no more than a month

drakkar67 · 2008-07-29, 11:57

I reserved those ranges before primegrid entered PSP project, moreover all our desktops were changed to notebooks (so my crunching power significantly decreased). I'm aware my reservations are creating a "risky gap". In 3-4 days time I'll rearrange my todo pairs and unreserve k=79817 pairs so they an be added to knpairs or may be crunched by other manual testers faster than me.

drakkar

ltd · 2008-07-29, 16:15

About the ranges for primegrid. The residues will be returned as one big chunk. Normaly I would mark the ranges as finished to have more realistic
number about open tests. But there is a limitation in my database design and in my scripts as they were never planned to be used in an environment with two server platform plus manual ranges.

So if i would mark the primegrid ranges out there would be problems to give them the stats credits when the ranges are really returned.

Till now I have no good idea to solve this issue.

Sloth · 2008-07-29, 22:47

LTD: The way to deal with PrimeGrid is to do what ever is easiest. There is little enough free time so no point putting a lot of effort into making the numbers look pretty.

As far as getting to 5M - we are currently at 10030 left. If PG is about to hand us 9500 results then we just have a little cleanup to do to finish up 5M. I think 6M would be the challenge by years end. I believe the tests for the last prime are still in the list too so that will knock things back a little more too.

S.

Joe O · 2008-08-11, 17:54

Quote:

Originally Posted by Joe O

We still have the goal of 1st pass reaching 5M.

Code:

Countdown to finish all n<5Mil: 10070 tests to go

This is reachable. We've just done 6959 tests in 115 days.

Quote:

Originally Posted by Sloth

As far as getting to 5M - we are currently at 10030 left. If PG is about to hand us 9500 results then we just have a little cleanup to do to finish up 5M. I think 6M would be the challenge by years end. I believe the tests for the last prime are still in the list too so that will knock things back a little more too.

S.

Time to bring this thread up-to-date now that the information has been updated:

Code:

Stats from: 2008/08/11 18:30 CEST

We have 900904 untested values left under n=50 Million.

Countdown to reach 1 Million digits (n>=3321929): 0 tests to go

Countdown to finish all n<4Mil: 0 tests to go

Countdown to finish all n<5Mil: 1349 tests to go

Countdown to finish all n<6Mil: 17950 tests to go

Days since last prime: 18

PRP test done since last prime: 10604

Just to have all the relevant information in this one post, so that we can watch our progress to this goal.:

Code:

Open n per block(Shown upto 6Mil): 
upper_bound amount 
4025000 10 
4050000 22 
4125000 24 
4150000 22 
4275000 63 
4300000 73 
4325000 1 
4475000 1 
4500000 19 
4525000 51 
4550000 44 
4575000 57 
4600000 63 
4625000 61 
4650000 57 
4675000 31 
4700000 63 
4750000 1 
4775000 45 
4800000 68 
4825000 53 
4850000 91 
4875000 84 
4900000 98 
4925000 46 
4950000 44 
4975000 86 
5000000 71 
5050000 143 
5125000 525 
5150000 528 
5175000 522 
5200000 484 
5225000 533 
5250000 498 
5275000 501 
5300000 536 
5325000 466 
5350000 498 
5375000 479 
5400000 497 
5425000 24 
5450000 72 
5475000 45 
5500000 434 
5525000 351 
5550000 494 
5575000 470 
5600000 512 
5625000 527 
5650000 518 
5675000 486 
5700000 505 
5725000 531 
5750000 496 
5775000 514 
5800000 493 
5825000 489 
5850000 492 
5875000 475 
5900000 520 
5925000 479 
5950000 449 
5975000 504 
6000000 511

VJS · 2008-08-12, 02:27

I just wanted to bring this one up as one of our previously posted milestones for the year.

Quote:

Originally Posted by Chino112

3. Sieve to 750 trillion

750T...

> 4P WOW!!!

Joe O · 2008-08-12, 12:33

Quote:

Originally Posted by VJS

I just wanted to bring this one up as one of our previously posted milestones for the year.

750T...

> 4P WOW!!!

Two points:
1) Getting to .75P was our goal for the previous year. This year it was getting to 5P.
2) The 4.1P you are referring to from another post, is just the sum of the ranges that PrimeGrid has sieved to date.

In regard to the goal of getting to 5P, we reached 5.102P by July 31, 2008. On that day there were also completed ranges up to 7.466P. Today PrimeGrid has sieved completely through 6.514P and partially through 8.081P. Their goal is to reach 10P by October 13th, their 1 year anniversary of joining PSP.
So, yes, this goal has been well and truly met.

Joe O · 2008-09-25, 12:01

Just updating our progress.

Code:

Stats from: 2008/09/25 13:30 CEST

We have 892331 untested values left under n=50 Million.

Countdown to reach 1 Million digits (n>=3321929): 0 tests to go

Countdown to finish all n<4Mil: 0 tests to go

Countdown to finish all n<5Mil: 516 tests to go

Countdown to finish all n<6Mil: 15197 tests to go

Days since last prime: 63

PRP test done since last prime: 13636

Code:

Open n per block(Shown upto 6Mil): 
upper_bound amount 
4150000 9 
4275000 22 
4300000 19 
4325000 1 
4475000 1 
4500000 2 
4575000 28 
4600000 62 
4625000 61 
4650000 57 
4675000 30 
4700000 59 
4750000 1 
4775000 1 
4800000 2 
4825000 8 
4850000 27 
4875000 25 
4900000 61 
4925000 2 
4975000 20 
5000000 18 
5050000 32 
5125000 522 
5150000 521 
5175000 522 
5200000 480 
5225000 531 
5250000 493 
5275000 496 
5300000 531 
5325000 462 
5350000 495 
5375000 475 
5400000 496 
5425000 22 
5450000 28 
5475000 14 
5500000 304 
5525000 122 
5550000 26 
5575000 53 
5600000 108 
5625000 526 
5650000 515 
5675000 485 
5700000 505 
5725000 529 
5750000 491 
5775000 512 
5800000 488 
5825000 487 
5850000 490 
5875000 473 
5900000 516 
5925000 477 
5950000 445 
5975000 501 
6000000 508

2008-07-29, 02:32	#37
themerrill Jun 2008 2·5 Posts	This is relevant to our discussion, just a little long. One thing I've wondered about prime grid's LLR: They are obviously given a special selection of .txts. Which is great and likely worth it, with all the results they provide and all. When the crunch, do we get the results back little by little (PG confirms that a k/n pair is not prime and sends it in) or do we get big batches when the finish entire ranges of numbers. Relevant because they just finished crunching 9500 k/n pairs where 4M<n<5M. If we are about to get that back in one lump, we are pretty much there in our n<5M countdown. Edit - To answer sloth's question - Yes, most of them do appear to be tied up on PrimeGrid, which would confirm my question as mentioned above. They just finished, and should be sending it in soon, if they do one big send off. The remaining 500 or so are reserved by drakkar and should be done in no more than a month Last fiddled with by themerrill on 2008-07-29 at 02:39

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Another milestone!	tcharron	PrimeNet	3	2013-08-29 06:44
Another milestone	frmky	Msieve	7	2012-04-25 22:12
New Milestone	opyrt	Prime Sierpinski Project	65	2010-10-06 13:18
CSVs for stats available + New combined stats	opyrt	Prime Sierpinski Project	3	2010-05-31 08:13
Stats milestone	Footmaster	Prime Sierpinski Project	64	2006-10-30 18:36

2008-07-28, 22:21	#34
VJS Dec 2004 299₁₀ Posts	Humm... A very good point, but you see Lars does have control of when and were all primes are found so... LOL O.k. scratch #1

2008-07-29, 11:57	#38
drakkar67 Feb 2003 Istanbul 52₁₀ Posts	I reserved those ranges before primegrid entered PSP project, moreover all our desktops were changed to notebooks (so my crunching power significantly decreased). I'm aware my reservations are creating a "risky gap". In 3-4 days time I'll rearrange my todo pairs and unreserve k=79817 pairs so they an be added to knpairs or may be crunched by other manual testers faster than me. drakkar

2008-07-29, 16:15	#39
ltd Apr 2003 1100000100₂ Posts	About the ranges for primegrid. The residues will be returned as one big chunk. Normaly I would mark the ranges as finished to have more realistic number about open tests. But there is a limitation in my database design and in my scripts as they were never planned to be used in an environment with two server platform plus manual ranges. So if i would mark the primegrid ranges out there would be problems to give them the stats credits when the ranges are really returned. Till now I have no good idea to solve this issue.

2008-07-29, 22:47	#40
Sloth Mar 2006 2×47 Posts	LTD: The way to deal with PrimeGrid is to do what ever is easiest. There is little enough free time so no point putting a lot of effort into making the numbers look pretty. As far as getting to 5M - we are currently at 10030 left. If PG is about to hand us 9500 results then we just have a little cleanup to do to finish up 5M. I think 6M would be the challenge by years end. I believe the tests for the last prime are still in the list too so that will knock things back a little more too. S.