| Author |
Message |
|---|
JimB Honorary cruncher
Send message
Joined: 4 Aug 11
Posts: 920
ID: 107307
Credit: 989,207,162
RAC: 300,698
                     
|
|
I created a spreadsheet to calculate the optimal sieving level for each GFN n based on how many candidates we're projected to test over the next five years. I have good data on what b level started testing when and comparing that to the current leading edge gives a good approximation of how many candidates we'll test in the future.
As of 2016-12-27:
n Current Optimal Percentage
Level Level Done
15 26000P 10300P 252%
16 8582P 65500P 13.1%
17 29161P 157900P 18.5%
18 20130P 61300P 32.8%
19 19214P 77700P 24.7%
20 37540P 79400P 47.3%
21 87000P 146400P 59.4%
22 172171P 154900P 111%
The immediate takeaway is that we're oversieved on n=15 and n=22. Sieving for n=15 is already stopping at p=26000P while n=22 is now set to stop at p=173000P. Optimal sieving for n=17 is much higher because with two b ranges being tested, much more of sieving is useful to eliminate candidates.
Sieving priorities are n=20 and n=21. They take the longest to test with genefer but are the fastest to sieve. Next would be n=18 and n=19 because the percentage of optimal is so low. n=16 and n=17 come last: while they have a long way to get to optimal, testing those candidates with genefer is pretty fast. When any n hits the optimal level, we'll suspend it so the other n's get more work.
Feel free to sieve any open n, the above is for those people who are asking which n to prioritize. |
|
|
|
|
|
Great Help, Answers a lot of un asked questions. :)
Happy Holiday's |
|
|
|
|
|
Am I correct?
We are approaching fast the sieving limit of n21:
n21 87000P 146400P 59.42622951% (Posted: 27 Dec 2016
n21 143767P 146400P 98.20150273% (23 Jan 2017)
Will n21 sieve end at the end of January 2017?
|
|
|
JimBHonorary cruncher
Send message
Joined: 4 Aug 11
Posts: 920
ID: 107307
Credit: 989,207,162
RAC: 300,698
|
|
Yes, you are correct. We've had three or four users who seem to be competing to see how much sieving they can do on n=21. As the optimal point is already programmed into the system, it will stop accepting reservations when optimal is hit (and will limit the size of reservations before that so as not to go past optimal). I don't anticipate recalculating the optimal levels until some of the other n's have been sieved a lot further.
All of the sieving status tables now show the optimal sieving level at the bottom and in the case of the graphs they end there. The unreserved portion of n=21 is about half the width of my screen on the graph now. |
|
|
JimBHonorary cruncher
Send message
Joined: 4 Aug 11
Posts: 920
ID: 107307
Credit: 989,207,162
RAC: 300,698
|
|
I've recomputed optimal sieving levels for all GFN n values. GFN20, 21 and 22 are now reopened with new higher optimal sieving levels based on how much work we've done, time to test with the most recent Genefer and time to sieve. |
|
|
|
|
|
It looks like n=16 crossed 65500P but is still issuing new work.
____________
 |
|
|
JimBHonorary cruncher
Send message
Joined: 4 Aug 11
Posts: 920
ID: 107307
Credit: 989,207,162
RAC: 300,698
|
It looks like n=16 crossed 65500P but is still issuing new work.
It'll stop at p=66000P (I like to stop at multiples of 1000P for sieving this high). |
|
|
 Dave
Send message
Joined: 13 Feb 12
Posts: 3204
ID: 130544
Credit: 2,283,569,067
RAC: 979,091
                        
|
|
Bingowings. |
|
|
|
|
|
Why did GFN16 sieving stop? |
|
|
RafaelVolunteer tester
Send message
Joined: 22 Oct 14
Posts: 911
ID: 370496
Credit: 548,293,027
RAC: 432,011
     
|
Why did GFN16 sieving stop?
It has hit optimal depth already. Or rather, there are more important things to sieve right now. |
|
|
|
|
|
OK so does that mean the next one, GFN17 is the priority? It is next shown in the list. |
|
|
RafaelVolunteer tester
Send message
Joined: 22 Oct 14
Posts: 911
ID: 370496
Credit: 548,293,027
RAC: 432,011
|
OK so does that mean the next one, GFN17 is the priority? It is next shown in the list.
n=17 always was the priority, the list just shows stuff from lowest n to highest. |
|
|
JimBHonorary cruncher
Send message
Joined: 4 Aug 11
Posts: 920
ID: 107307
Credit: 989,207,162
RAC: 300,698
|
|
Every six months I reevaluate the sieving levels, GFN16 will almost certainly reopen at that time. GFN19, GFN20 and GFN22 certainly will. For purposes of the computation, I look at what we'll check with genefer over the next five years. I think that makes a lot of sense, especially now that we have to appeal to users for money to keep the servers hosted. |
|
|
GDBSend message
Joined: 15 Nov 11
Posts: 298
ID: 119185
Credit: 4,052,581,703
RAC: 1,879,596
         
|
|
Sieving stats 02/15/18
n
22 58.4% entries sieved out
21 61.4
20 64.5
19 68.0
18 68.1 |
|
|
|
|
|
It seems like GFN n=22 isn‘t accepting any new reservations. Did we achieve optimal sieving depth? |
|
|
JimBHonorary cruncher
Send message
Joined: 4 Aug 11
Posts: 920
ID: 107307
Credit: 989,207,162
RAC: 300,698
|
It seems like GFN n=22 isn‘t accepting any new reservations. Did we achieve optimal sieving depth?
Look two posts before yours. |
|
|
|
|
Sieving stats 02/15/18
n
22 58.4% entries sieved out
21 61.4
20 64.5
19 68.0
18 68.1
what about now?
all GFNs seem to be out
____________
My lucky number is 6219*2^3374198+1
 |
|
|
JimBHonorary cruncher
Send message
Joined: 4 Aug 11
Posts: 920
ID: 107307
Credit: 989,207,162
RAC: 300,698
|
|
GFNs are not 'out". Twice a year I calculate what siving we need to run to be optimal for the next five years' worth of genefer work. If GFN19 and GAFN22 run out before June 30th, I'll simply do my calculations early. |
|
|
|
|
|
so new work will come out before 6.30?
____________
My lucky number is 6219*2^3374198+1
|
|
|
dthononVolunteer tester
Send message
Joined: 6 Dec 17
Posts: 435
ID: 957147
Credit: 1,748,562,214
RAC: 370,038
                             
|
so new work will come out before 6.30?
GFN19 and GFN22 are open for reservation.
|
|
|
|
|
|
Will the sieve credit be counted in the GFN subproject?
____________
My lucky number is 6219*2^3374198+1
|
|
|
HonzaVolunteer moderator
Volunteer tester
Project scientist
Send message
Joined: 15 Aug 05
Posts: 1957
ID: 352
Credit: 6,128,462,973
RAC: 2,260,686
                            
|
|
See Welcome to PrimeGrid Manual Sieving
____________
My stats |
|
|
|
|
|
wait, didn't it run out again?
____________
My lucky number is 6219*2^3374198+1
|
|
|
|
|
|
Oh nice, there is new work available. Therefore, the current status of the manual sieving is
current new optimal percentage
n depth depth done
15 26000P 26000P 100%
16 66382P 120000P 55%
17 175466P 250000P 70%
18 300010P 420000P 71%
19 429395P 460000P 93%
20 242000P 300000P 80%
21 900000P 900000P 100%
22 753200P 950000P 79%
|
|
|
|
|
|
Out again???
I didn't even get a chance to reserve anything!
____________
My lucky number is 6219*2^3374198+1
|
|
|
darkclownVolunteer tester
Send message
Joined: 3 Oct 06
Posts: 331
ID: 3605
Credit: 1,443,342,948
RAC: 650,658
    
|
Out again???
I didn't even get a chance to reserve anything!
There's plenty. Did you look at the Manual Sieving page?
____________
My lucky #: 60133106^131072+1 (GFN 17-mega) |
|
|
|
|
|
On my computer it says that 'No New Reservations' for all GFN sieving.
____________
My lucky number is 6219*2^3374198+1
|
|
|
JimBHonorary cruncher
Send message
Joined: 4 Aug 11
Posts: 920
ID: 107307
Credit: 989,207,162
RAC: 300,698
|
On my computer it says that 'No New Reservations' for all GFN sieving.
All the other tasks at PrimeGrid are controlled by the server, but manual sieving is controlled by me, personally, by hand. It's only for experienced users, users who do not need a large amount of hand-holding to get things done. People who have a track record of getting things done on their own. When people are just following the instructions that someone else has given them without fully understanding what they're doing, tasks get done incorrectly. If you mess up a BOINC configuration or a PRPNet configuration, the server handles that automatically. When you mess up a manual sieving reservation, I need to personally clean up the mess.
Because of this, I can choose who does manual sieving, and who can't. I do not feel that you are a good candidate for manual sieving. We have lots of other projects that are more automated, which you can participate in if you wish. |
|
|
|
|
|
I get the same message Jim but I have successfully completed 37,339,217 points in the Project Staging Area previously. Perhaps you can allow some more users again? |
|
|
JimBHonorary cruncher
Send message
Joined: 4 Aug 11
Posts: 920
ID: 107307
Credit: 989,207,162
RAC: 300,698
|
I get the same message Jim but I have successfully completed 37,339,217 points in the Project Staging Area previously. Perhaps you can allow some more users again?
I just logged in to your account and all manual sieving is available to you. You can have up to 10 simultaneous open reservations with a maximum size of 342P each at the moment. I have no idea what you were looking at as I didn't change anything. |
|
|
