| Author |
Message |
|---|
John Honorary cruncher
Send message
Joined: 21 Feb 06
Posts: 2875
ID: 2449
Credit: 2,681,934
RAC: 0
                 
|
|
Welcome to the Full Moon Challenge
Come howl at the moon with us and "hunt" for Top 5000 primes. It's a GREAT opportunity for you to find one for yourself! Who knows, there may even be the rare Fermat number divisor. :) We hope to establish a single day record for finding Top 5000 primes!
A 24 hour Challenge is being offered on PrimeGrid's Proth Prime Search (LLR) application. NOTE: Only PPSE work will be released during the Challenge. However, PPS resends will occur. All completed work, PPS or PPSE, will be credited in the Challenge score.
To participate in the Challenge, please select only the Proth Prime Search (LLR) project in your PrimeGrid preferences section. The challenge will begin 25 July 2010 18:00 UTC and end 26 July 2010 18:00 UTC. Application builds are available for Linux 32 bit, Windows 32 bit and MacIntel. These applications will be sent to 64 bit clients. As with all LLR application projects, there is no advantage of 64 bit over 32 bit.
ATTENTION: The primality program LLR is CPU intensive; so, it is vital to have a stable system with good cooling. It does not tolerate "even the slightest of errors." Please see this post for more details on how you can "stress test" your computer. WU's will take ~5 minutes on fast/newer computers and 15+ minutes on slower/older computers. If your computer is highly overclocked, please consider "stress testing" it. Sieving is an excellent alternative for computers that are not able to LLR. :)
Please, please, please make sure your machines are up to the task.
Time zone converter:
The World Clock - Time Zone Converter
NOTE: The countdown clock on the front page uses the host computer time. Therefore, if your computer time is off, so will the countdown clock. For precise timing, use the UTC Time in the data section to the left of the countdown clock.
Scoring Information
Scores will be kept for individuals and teams. Only work units issued AFTER 25 July 2010 18:00 UTC and received BEFORE 26 July 2010 18:00 UTC will be considered for credit. We will use the "prime score" method which is based on the n value (k*b^n-1) to score the challenge. The only difference is that the primary and double checker of a WU will receive the same score.
Therefore, each completed WU will earn a unique score based on its n value. The higher the n, the higher the score. This is different than BOINC cobblestones! A quorum of 2 is NOT needed to award Challenge score - i.e. no double checker. Therefore, each returned result will earn a Challenge score. Please note that if the result is eventually declared invalid, the score will be removed.
For details on how the score is calculated, please see this thread.
At the Conclusion of the Challenge
We kindly ask users "moving on" to ABORT their WU's instead of DETACHING, RESETTING, or PAUSING.
ABORTING WU's allows them to be recycled immediately; thus a much faster "clean up" to the end of an LLR Challenge. DETACHING, RESETTING, and PAUSING WU's causes them to remain in limbo until they EXPIRE. Therefore, we must wait until WU's expire to send them out to be completed.
Please consider either completing what's in the queue or ABORTING them. Thank you. :)
About the Proth Prime Search
The Proth Prime Search is done in collaboration with the Proth Search project. This search looks for primes in the form of k*2^n+1. With the condition 2^n > k, these are often called Proth primes. This project also has the added bonus of possibly finding factors of "classical" Fermat numbers or Generalized Fermat numbers. As this requires PrimeFormGW (PFGW) (a primality-testing program), once PrimeGrid finds a prime, it is then tested on PrimeGrid's servers for divisibility.
Proth Search only searches for k<1200. PrimeGrid created an extension to that which includes all candidates 1200<k<10000 for n<5M. It is this extension which we call PPSE that the Challenge will be on.
Initially, PrimeGrid's PPS project's goal was to double check all previous work up to n=500K for odd k<1200 and to fill in any gaps that were missed. We have accomplished that now and have increased it to n=900K. PG's LLRNet searched up to n=200,000 and found several missed primes in previously searched ranges. Although primes that small did not make it into the Top 5000 Primes database, the work was still important as it may have led to new factors for "classical" Fermat numbers or Generalized Fermat numbers. While there are many GFN factors, currently there are only about 275 "classical" Fermat number factors known. Current primes found in PPS definitely make it into the Top 5000 Primes database.
Once the 900K goal is reached, we will head to 1M before turning our focus to smaller k values and higher n values. For example, k<300 complete to n=2M, k<600 complete to n=1.5M and so on.
Additional Proth prime testing can be found in PrimeGrid's PRPNet's. Current ports available are:
PRPNet Ports
Detailed daily user stats courtesy of Sysadm@Nbg
Today's Prime Finders courtesy of Sysadm@Nbg
port 10000: PPSE n<500K Server : User : Pending Tests
port 11000: PPSE n<500K Server : User : Pending Tests
port 12007: PPSE n>500K Server : User : Pending Tests
For more information about "Proth" primes, please visit these links:
About Proth Search
The Proth Search project was established in 1998 by Ray Ballinger and Wilfrid Keller to coordinate a distributed effort to find Proth primes (primes of the form k*2^n+1) for k < 300. Ray was interested in finding primes while Wilfrid was interested in finding divisors of Fermat number. Since that time it has expanded to include k < 1200. Mark Rodenkirch (aka rogue) has been helping Ray keep the website up to date for the past few years.
Early in 2008, PrimeGrid and Proth Search teamed up to provide a software managed distributed effort to the search. Although it might appear that PrimeGrid is duplicating some of the Proth Search effort by re-doing some ranges, few ranges on Proth Search were ever double-checked. This has resulted in PrimeGrid finding primes that were missed by previous searchers. By the end of 2008, all new primes found by PrimeGrid were eligible for inclusion in Chris Caldwell's Prime Pages Top 5000. Sometime in 2009, over 90% of the tests handed out by PrimeGrid were numbers that have never been tested.
PrimeGrid intends to continue the search indefinitely for Proth primes.
What is LLR?
The Lucas-Lehmer-Riesel (LLR) test is a primality test for numbers of the form N = k*2^n − 1, with 2^n > k. Also, LLR is a program developed by Jean Penne that can run the LLR-tests. It includes the Proth test to perform +1 tests and PRP to test non base 2 numbers. See also:
(Edouard Lucas: 1842-1891, Derrick H. Lehmer: 1905-1991, Hans Riesel: born 1929).
____________
|
|
|
|
|
|
Team Aggie the Pew hope to be there.
Hester is spooling up the rat-wheel* as I type. Socks is building up her energy by scoffing rat treats and Iris, who is blind, is blundering into objects in her cage trying to find the rat-wheel. They are confident they'll be ready for the event.
Good luck to all taking part.
* Actually grown-up rats rarely run in a wheel; they know they are not going to get anywhere. That and they are lazy little creatures** unless food is involved and they don't like the heat either. But hey!
** OK not Hester but then she's a strange rat. |
|
|
|
|
|
could we please avoid trouble like too short deadlines this time? |
|
|
JohnHonorary cruncher
Send message
Joined: 21 Feb 06
Posts: 2875
ID: 2449
Credit: 2,681,934
RAC: 0
|
could we please avoid trouble like too short deadlines this time?
Deadline set at 2.5 days now.
____________
|
|
|
JohnHonorary cruncher
Send message
Joined: 21 Feb 06
Posts: 2875
ID: 2449
Credit: 2,681,934
RAC: 0
|
|
How to Join the Challenge
To participate in the Challenge, please select only the Proth Prime Search (LLR) project in your PrimeGrid preferences section. Also, set "Send work from any subproject if selected projects have no work" to no to guarantee that no other work will be sent. The challenge will begin 25 July 2010 18:00 UTC and end 26 July 2010 18:00 UTC.
Application builds are available for the following:
- Microsoft Windows (98 or later) running on an Intel x86-compatible CPU
- Linux running on an Intel x86-compatible CPU
- Mac OS 10.4 or later running on Intel
Note: As with all LLR application projects, there is no advantage of 64 bit over 32 bit.
Time zone converter:
The World Clock - Time Zone Converter
NOTE: The countdown clock on the front page uses the host computer time. Therefore, if your computer time is off, so will the countdown clock. For precise timing, use the UTC Time in the data section to the left of the countdown clock.
Scoring Information
Scores will be kept for individuals and teams. Only work units issued AFTER 25 July 2010 18:00 UTC and received BEFORE 26 July 2010 18:00 UTC will be considered for credit. We will use the "prime score" method which is based on the n value (k*b^n-1) to score the challenge. The only difference is that the primary and double checker of a WU will receive the same score.
Therefore, each completed WU will earn a unique score based on its n value. The higher the n, the higher the score. This is different than BOINC cobblestones! A quorum of 2 is NOT needed to award Challenge score. Therefore, each returned result will earn a Challenge score. Please note that if the result is eventually declared invalid, the score will be removed.
For details on how the score is calculated, please see this thread.
Stats will be available on the front page next to the clock within 30 minutes after the start.
What to expect at Challenge start
The server always gets hit hard at the beginning of the Challenge. However, we will use the same procedure that we have been for over a year now for all the Challenges which has proved very successful. A few hours before the start (25 July 2010 18:00 UTC) "max cache" will be dropped to 16 tasks per core. Additionally, "max to send" will be dropped to 8 tasks per core. Attempting to request more will be met with "max cache" messages.
We'll raise the buffer to at least 200K. At the start of the Challenge, these settings will be adjusted as necessary. However, if the past year is any indication, we'll make the first adjustment around 30 minutes post start by raising "max cache" to 32. We'll continue to increase "max cache" when the server can handle it until a full 24 hour max cache is possible.
This method allows the greatest opportunity for the most clients to get work. Even if it's just 1 task for the first 5 minutes, at least the client is crunching. Past Challenges have shown that clients have a good chance to reach "max cache" before their first task is complete. We expect the same this time.
Strategies for starting the Challenge
Depending on a variety of factors, different strategies work for different users. Here are just a few to consider:
- large farm, user can be present at start
-Set Computer is connected to the Internet about every to 0 days
-Set Maintain enough work for an additional to 0 days
-Before Challenge start, update PrimeGrid preferences and only select Proth Prime Search (LLR) project
-At worse, you'll only be 1 PPS (LLR) WU late on each core starting the Challenge.
-After all machines have work, increase Maintain enough work for an additional
- large farm, user NOT able to be present at start
-Same settings as above.
-At worse, you'll only be 1 WU late on each core starting the Challenge.
- a few computers, user can be present at start
-Change PrimeGrid preferences and only select Proth Prime Search (LLR) project
-Set computers to "No New Tasks" for PrimeGrid
-At Challenge Start, update computers to "Allow New Tasks"
- a few computers, user NOT able to be present at start
-Set Computer is connected to the Internet about every to 0 days
-Set Maintain enough work for an additional to 0 days
-Before Challenge start, update PrimeGrid preferences and only select Proth Prime Search (LLR) project
-At worse, you'll only be 1 WU late on each core starting the Challenge.
NOTE: This presumes you have all other BOINC projects suspended.
Best of Luck to everyone!
____________
|
|
|
JohnHonorary cruncher
Send message
Joined: 21 Feb 06
Posts: 2875
ID: 2449
Credit: 2,681,934
RAC: 0
|
|
Automatic Prime Submissions
We are expecting an unprecedented discovery of primes during the Challenge. The best way to make sure your primes are submitted in a timely manner is to give PrimeGrid permission to submit the primes on your behalf to the Prime Pages.
To do this, please edit your PrimeGrid preferences page to give PrimeGrid permission to submit. Name and last name (surname) are required by the Prime Pages to report primes.
If you do not wish to submit primes using your name, but wish to give PrimeGrid permission to submit, please enter "Anonymous" in the name field. As with all prime discoveries, the double checker will be given an opportunity to report before the prime, as a last resort, is submitted anonymously.
For those who already have a Prime Pages account from previous prime efforts can also participate in this auto-reporting. No access to your Prime Pages prover account is needed...just permission (and name) to report on the PrimeGrid preferences page.
Thank you and best of luck to everyone!
____________
|
|
|
|
|
|
What is the recent discovery rate (eg primes per 10 000 w/u) on this project please? |
|
|
|
|
|
~5000 wu/prime
Lennart
|
|
|
JohnHonorary cruncher
Send message
Joined: 21 Feb 06
Posts: 2875
ID: 2449
Credit: 2,681,934
RAC: 0
|
|
A Friendly Reminder
ATTENTION: The primality program LLR is CPU intensive; so, it is vital to have a stable system with good cooling. It does not tolerate "even the slightest of errors." Please see this post for more details on how you can "stress test" your computer. WU's will take ~5 minutes on fast/newer computers and 15+ minutes on slower/older computers. Although these are relatively short WU's, it is still important to monitor your machines for heat. If your computer is highly overclocked, please consider "stress testing" it.
Restricted airflow is one of the primary reasons for overheating. Take the time to monitor the fans and review the dust buildup. If needed, give your computer a good cleaning. It will thank you by running cooler and producing valid results. :) Please, please, please make sure your machines are up to the task. Sieving is an excellent alternative for computers that are not able to LLR. :)
____________
|
|
|
|
|
|
About PRPNet. !!!!!!!!!!
I will not load more work in any port before challenge are done. Maybe need to wait a day more when the cleanup is running.
This is to minimize load on DB.
I shall try keep them running, but if load get's high I may have to stop them.
Lennart
|
|
|
JohnHonorary cruncher
Send message
Joined: 21 Feb 06
Posts: 2875
ID: 2449
Credit: 2,681,934
RAC: 0
|
|
24 hours to go!!!
There is coming soon a gathering of Warewolves and Prime Hunters. Be sure not to mix the two groups up or you might find yourself surrounded by blood thirsty animals!! ;) However, some might say that the two groups are NOT mutually exclusive. :D Come join us this Full Moon for a sporting "hunt" of Top 5000 primes in the Proth Prime Search (LLR) project.
Please once again review the "Welcome" post as well as the "How to Join" post so you can be informed and make any last minute adjustments that are necessary.
A few hours before the Challenge, we begin to monitor the Meebo chat room at the bottom of the forum page. If at any time PrimeGrid becomes unresponsive, you can still access the Meebo chat room externally here: http://www.meebo.com/room/primegrid/
Before the start, the Subproject stats on the Server Status page will be disabled as will the 24 hour stats on the bottom of the front page. Also, the Top Primes Finder page will be disabled. This is all done to free up as much resources as possible to allow the server to send and receive work. With these short WU's and the increased competition, we expect a very high volume at start.
What to expect at Challenge start
The server always gets hit hard at the beginning of the Challenge. However, we will use the same procedure that we have been for over a year now for all the Challenges which has proved very successful. A few hours before the start (25 July 2010 18:00 UTC) "max cache" will be dropped to 16 tasks per core. Additionally, "max to send" will be dropped to 8 tasks per core. Attempting to request more will be met with "max cache" messages.
We'll raise the buffer to at least 200K. At the start of the Challenge, these settings will be adjusted as necessary. However, if the past year is any indication, we'll make the first adjustment around 30 minutes post start by raising "max cache" to 32. We'll continue to increase "max cache" when the server can handle it until a full 24 hour max cache is possible.
This method allows the greatest opportunity for the most clients to get work. Even if it's just 1 task for the first 5 minutes, at least the client is crunching. Past Challenges have shown that clients have a good chance to reach "max cache" before their first task is complete. We expect the same this time.
Stats will be available on the front page next to the clock within 30 minutes after the start.
BOINCstats provides hourly updates of Project Challenges. There you can view user and team credit and ranking trends.
____________
|
|
|
|
|
|
All my clients are on and warming up... |
|
|
JohnHonorary cruncher
Send message
Joined: 21 Feb 06
Posts: 2875
ID: 2449
Credit: 2,681,934
RAC: 0
|
|
It appears that the Prime Pages has already been stricken with fright about the upcoming Full Moon. :) The site has been down for the last 24 hours. Should the site remain down, we'll report primes as soon as it's back online.
____________
|
|
|
|
|
|
What is up with this WU:
Min remaining number 516329 (155434 digits long)
It is lurking around there for weeks in the subprojects overwied but the workunit can't be found. |
|
|
|
|
|
Let the games begin.. I look very much forward to this race.
DD,
____________
|
|
|
|
|
|
Hi all, have same problem?
25.7.2010 20:03:23 PrimeGrid Scheduler request failed: HTTP internal server error
maybe heavy load??
JHAPA
Yes it was really Heavy load :-)
____________
 |
|
|
|
|
|
Can't get work!!!! The servers aren't keeping up!!!! :-( |
|
|
|
|
|
Same here:
Dom 25 Jul 19:11:46 2010 PrimeGrid Requesting new tasks
Dom 25 Jul 19:11:56 2010 PrimeGrid Scheduler request completed: got 0 new tasks
Dom 25 Jul 19:11:56 2010 PrimeGrid Message from server: Server can't open database
|
|
|
|
|
|
Update: got new tasks after updating boinc manager to the latest version. |
|
|
|
|
|
Oooohhh! Aggie The Pew (may her shadow never grow less) currently in 36th place!
It'll never last. |
|
|
|
|
|
One of my machines is getting
"Message from server: No work sent"
"Message from server: (reached daily quota of 2400)"
http://www.primegrid.com/results.php?hostid=72476
This is not correct. That machine has been sent only 11 tasks per that page. But in reality, it has received 0 tasks. Those 11 are phantom tasks.
Can someone please reset the quota for that machine?
Edit: Of course, as soon as I finish posting, I try another manual update, and now it is working. No mention the quota from the server any more.
____________
Reno, NV
 |
|
|
|
|
|
On the stats page...... Points
looks like they are for a 5 & 8 Day Challenge not for 24 hrs
Steve
GO SETI.USA
____________
From the High Desert in New Mexico
 |
|
|
 Sysadm@NbgVolunteer moderator
Volunteer tester
Project scientist
Send message
Joined: 5 Feb 08
Posts: 1216
ID: 18646
Credit: 859,053,517
RAC: 193,643
                      
|
|
Did you mean Challenge Scoring as described in first post
on challenge stats are NOT cobblestones
or do I missunderstand you ?
____________
Sysadm@Nbg
my current lucky number: 113856050^65536 + 1
PSA-PRPNet-Stats-URL: http://u-g-f.de/PRPNet/
 |
|
|
|
|
|
This isn't going to work. The server can't keep up with all these short tasks.
The machines connect to the server to report and request tasks. They get "HTTP internal server error", and then BOINC forces a **1 hour** back off. Because of the short due dates, the queues can't fill up enough to cover that amount of back off time. So the machines sit idle.
____________
Reno, NV
|
|
|
|
|
Did you mean Challenge Scoring as described in first post
on challenge stats are NOT cobblestones
or do I missunderstand you ?
Well its a not a problem now been fixed. Points are now missing
But on the (old) current stats page:
Rank Name Score Points
But on the current stats page:
Rank Name Score
____________
From the High Desert in New Mexico
|
|
|
ktoSend message
Joined: 13 Jan 09
Posts: 27
ID: 34164
Credit: 41,469,365
RAC: 0
             
|
|
Just got this message:
"Message from server: Completed result pps_llr_extended_56054418_1 refused: result already reported as success"
Also, it reported 56054419_1 refused.
What does that mean? |
|
|
|
|
Just got this message:
"Message from server: Completed result pps_llr_extended_56054418_1 refused: result already reported as success"
Also, it reported 56054419_1 refused.
What does that mean?
It means your client already reported the tasks to the server. The server should have confirmed receipt when they were originally returned. But due to the server load, that transaction never happened. So when your client re-reported them, the server is just saying that you already returned them. You can ignore these messages.
____________
Reno, NV
|
|
|
|
|
|
When the challenge is over i will let my clients on for an additional 48 hours.
That should help a bit with the cleanup.
Afterwards i will shut down most if not all of my clients as i want to cut down my effort and my spending on energy - i should have reached my 10th million by then. |
|
|
|
|
|
reached daily quota of 2400 task
is that per box or per core
it's on a 16 core box
if it is per core => 38400 per day => I don't think i have done that many
or is it per box => 150 / core / day => @ 9.5 min / WU comes close
When does the clock reset for each day
Steve
____________
From the High Desert in New Mexico
|
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13954
ID: 53948
Credit: 392,395,963
RAC: 168,999
                          
|
reached daily quota of 2400 task
is that per box or per core
it's on a 16 core box
if it is per core => 38400 per day => I don't think i have done that many
or is it per box => 150 / core / day => @ 9.5 min / WU comes close
When does the clock reset for each day
Steve
I got that message earlier today on a machine that had done ZERO WUs since the last challenge, so it likely has nothing to do with how many WUs you already have. It's some kind of server bug; probably has something to do with the overloading.
Most importantly, the bug went away shortly afterwards and WUs started flowing.
____________
My lucky number is 75898524288+1 |
|
|
|
|
Most importantly, the bug went away shortly afterwards and WUs started flowing.
+1 It went away for me too. Getting work just fine after a few more tries.
____________
Reno, NV
|
|
|
|
|
|
I cannot get any work as I am getting an internal error fromm the server
26/07/2010 10:51:29 AM PrimeGrid Sending scheduler request: To fetch work.
26/07/2010 10:51:29 AM PrimeGrid Requesting new tasks for CPU and GPU
26/07/2010 10:52:31 AM PrimeGrid Scheduler request failed: HTTP internal server error
26/07/2010 10:53:32 AM PrimeGrid Sending scheduler request: To fetch work.
26/07/2010 10:53:32 AM PrimeGrid Requesting new tasks for CPU and GPU
26/07/2010 10:54:36 AM PrimeGrid Scheduler request failed: HTTP internal server error
26/07/2010 10:56:28 AM PrimeGrid Sending scheduler request: To fetch work.
26/07/2010 10:56:28 AM PrimeGrid Requesting new tasks for CPU and GPU
26/07/2010 10:57:32 AM PrimeGrid Scheduler request failed: HTTP internal server error
____________
Proud Founder of 
Have a look at my WebCam |
|
|
|
|
|
I just discovered that my machine has been down due to power outage. I got my client up and tried a manual update, but I get a message that says "Not reporting or requesting tasks." I know I'm way behind, but I'd like to take part in the fun. Can someone help me get WUs? |
|
|
|
|
I just discovered that my machine has been down due to power outage. I got my client up and tried a manual update, but I get a message that says "Not reporting or requesting tasks." I know I'm way behind, but I'd like to take part in the fun. Can someone help me get WUs?
Resetting the project and setting the additional work buffer to 1 day let me start getting WU's. |
|
|
|
|
I just discovered that my machine has been down due to power outage. I got my client up and tried a manual update, but I get a message that says "Not reporting or requesting tasks." I know I'm way behind, but I'd like to take part in the fun. Can someone help me get WUs?
Resetting the project and setting the additional work buffer to 1 day let me start getting WU's.
That worked. Thx!
____________
-c
 |
|
|
|
|
I just discovered that my machine has been down due to power outage. I got my client up and tried a manual update, but I get a message that says "Not reporting or requesting tasks." I know I'm way behind, but I'd like to take part in the fun. Can someone help me get WUs?
What other WU do you have running?
I have 2 PSP_LLR WU's crunching away with 25+ hr to go
If I suspend therm I can not get any PPS_LLR WU for the Challenge. I have too much time in the to abort so a few boxes are out of the dash for primes.
PSP & SOB WU's are killers on fast boxes and slow one are just a slow grind in the mud
Crunch On
Steve
____________
From the High Desert in New Mexico
|
|
|
|
|
|
I get mysql problems and no upload to completed work they are just sitting there completed at 100 % and not uploading.
Looks like we have over ran the database with too much return work.
DD,
____________
|
|
|
|
|
|
I can upload, but am getting db error. This was an on/off problem since the beginning of the race. But now it has been like for over an hour solid. I think it is really broken this time. I wonder if anyone in europe is awake yet?
____________
Reno, NV
|
|
|
|
|
|
It just seemed to un-logjam results returned and new work downloaded....someone must have kicked it.
____________
 |
|
|
|
|
It just seemed to un-logjam results returned and new work downloaded....someone must have kicked it.
Yep. I just now was able to report and get new work.
____________
Reno, NV
|
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13954
ID: 53948
Credit: 392,395,963
RAC: 168,999
|
I just discovered that my machine has been down due to power outage. I got my client up and tried a manual update, but I get a message that says "Not reporting or requesting tasks." I know I'm way behind, but I'd like to take part in the fun. Can someone help me get WUs?
What other WU do you have running?
I have 2 PSP_LLR WU's crunching away with 25+ hr to go
If I suspend therm I can not get any PPS_LLR WU for the Challenge. I have too much time in the to abort so a few boxes are out of the dash for primes.
PSP & SOB WU's are killers on fast boxes and slow one are just a slow grind in the mud
I have one box that crunches long-duration WUs such as the SoB and PSP LLRs. It's almost unmanageable trying to work the challenges into the run schedule for those WUs, especially since 4 of them run concurrently. With the length of SoB, for example, I could usually get only 1 set of SoB WUs done between challenges before having to shift to shorter WUs so the machine would be ready for the next challenge. Totally and completely unworkable, so...
What I do on that box is suspend computation when the challenge starts, and then open a second copy of BOINC inside a VMware virtual machine. The VM crunches for the challenge. Once the challenge is over, I let the BOINC in the VM run dry, then re-enable computation on the real machine.
VMware is free (and there's other free virtual machines out there if you don't like that one), and depending on what sub-project is running on the challenge, I can run the VM in as little as 500 meg of memory (for 4 LLRs). For some of the larger sieves, I can get away with running 4 tasks in a 1200 meg VM. That's running 64 bit Ubuntu.
____________
My lucky number is 75898524288+1 |
|
|
JohnHonorary cruncher
Send message
Joined: 21 Feb 06
Posts: 2875
ID: 2449
Credit: 2,681,934
RAC: 0
|
|
OK, we are still under the deluge of the start and are only just beginning to recover. Go ahead and let us have it. All complaints are well justified. :(
Playing around with settings did not show much improvement. However, as a last resort, stats have been halted. For the last 60 minutes now, we have seen markedly better performance. We hope you are experiencing the same on your end.
Stats will now be run every few hours and only under supervision. We know this is not ideal and apologize for the inconvenience.
Please keep posting your observations.
____________
|
|
|
|
|
|
db down again.
____________
Reno, NV
|
|
|
|
|
|
Was down 20 min when i run stats. Is working now.
Lennart |
|
|
JohnHonorary cruncher
Send message
Joined: 21 Feb 06
Posts: 2875
ID: 2449
Credit: 2,681,934
RAC: 0
|
|
Please expect DB to go down during stats updates...thus the change to every few hours. However, expect the next stats update to be a little longer as sleep has now come.
Current settings are:
Max Cache: 28 task
Max to send: 16 tasks
Buffer: 150K tasks
Validator is struggling to keep up...more realistically, it's slowly falling behind. We'll continue to work on this. All work returned before the deadline will be counted. It will just take time for the final results as we wait for the validator to catch up.
____________
|
|
|
STE\/EVolunteer tester
Send message
Joined: 10 Aug 05
Posts: 573
ID: 103
Credit: 3,631,506,192
RAC: 3
      
|
|
WOW Thank's PG, I didn't know how much Electricity I could save by running your Challenges. 15 Box's sitting Idle for Hours on end, Wonderful ...
____________
 |
|
|
HAmstyVolunteer tester
Send message
Joined: 26 Dec 08
Posts: 132
ID: 33421
Credit: 12,510,712
RAC: 0
       
|
WOW Thank's PG, I didn't know how much Electricity I could save by running your Challenges. 15 Box's sitting Idle for Hours on end, Wonderful ...
what a complaint - applause -.-
____________
 |
|
|
|
|
|
Wonderful! Finally, a box for which I was trying to get some work since the morning (since 9am), got some work (it's 1:30pm my time now). Quite a challenging challenge, I'd say!
I am not sure that credit results of such a challenge should be considered fair and regular, comparable to those of previous challenges (I've never experienced this magnitude of not getting work, before).
Maybe the 6 minute work units are simply not suited for a challenge of this type. It must have been obvious that the servers will get completely overwhelmed. |
|
|
|
|
Maybe the 6 minute work units are simply not suited for a challenge of this type. It must have been obvious that the servers will get completely overwhelmed.
That is the reason why the admins asked for help to crunch PPS (LLR) tasks before the challenge to reach higher n values (exponents) in the 550K range.
____________
|
|
|
|
|
|
When will statistics be brought up to date?
____________
|
|
|
|
|
|
I don't think I can start stats again before the end off the challenge.
so we stop at 18:00 UTC and I think it will take about 2-3 hr to get it done.
but that will not show all at once. We need to get all validated first.
Lennart
|
|
|
|
|
Maybe the 6 minute work units are simply not suited for a challenge of this type. It must have been obvious that the servers will get completely overwhelmed.
That is the reason why the admins asked for help to crunch PPS (LLR) tasks before the challenge to reach higher n values (exponents) in the 550K range.
That's why I've been pointed at PPS (LLR) since John mentioned it. On the upside, it looks like I reported "1433*2^531949+1 is not prime" 6.5 minutes before the challenge. 517k (June 21st) -> 532k (July 25th) is pretty good considering.
Also, THANK YOU PG for offering these great challenges and doing your best to make them both exciting and successful. We clearly like them so much that we can't help but overload the servers. Thank you for all of your hard work keeping up with our enthusiasm. :) |
|
|
|
|
|
I know Solar Extreme had some computers down for a little while last night because they could not connect. I am surprised that PrimeGrid has done this well with the short work units. I think everyone at PrimeGrid has done a great job. I wonder what it would take for PrimeGrid to be able to run a challenge like this without server problems? Is it even possible?
____________
May your addiction to Boinc be greater than mine!
End Transmission!
 |
|
|
|
|
|
I think that it'll make it all rather exciting to not the see the results update until after the end. All the work done and watching to see how places change as the validation proceeds will be a bit like election night.
And the rat girls will be able to lie in their hammock and watch the results come in rather than frantically pedalling!
Oh and I'm with royanee and Danial. L. Puckett and thank those who gave their time for nothing to make the event happen. I think we should bear that in mind before any stones get thrown about any glitches.
Cheers!
T
I don't think I can start stats again before the end off the challenge.
so we stop at 18:00 UTC and I think it will take about 2-3 hr to get it done.
but that will not show all at once. We need to get all validated first.
Lennart
|
|
|
|
|
Workunits waiting for validation: 300,084
I'm asking myself how many primes are hidden in this pile of WUs...
____________
|
|
|
JohnHonorary cruncher
Send message
Joined: 21 Feb 06
Posts: 2875
ID: 2449
Credit: 2,681,934
RAC: 0
|
Workunits waiting for validation: 300,084
I'm asking myself how many primes are hidden in this pile of WUs...
Over 60 primes have been found so far before the Warewolves took down the server and halted the validator.
Thanks for the posts of encouragement. It really does help with these challenging times. And yes, the criticism is also fully noted and deserved. Once again, we let our optimism outpace our reality. The Full Moon had more of a gravitation pull than we predicted.
In a sick twist of fate, at least the Prime Pages are up again. :-|
____________
|
|
|
|
|
|
All's working good at the moment. I woke up to 8 cores without work. No biggie, I added two more . Just keep doing what you are doing and I'll do the same. Good luck all. Jack
P.S. It's fun putting a pounding on the server !!! |
|
|
|
|
Workunits waiting for validation: 300,084
I'm asking myself how many primes are hidden in this pile of WUs...
My guess is 60. |
|
|
JohnHonorary cruncher
Send message
Joined: 21 Feb 06
Posts: 2875
ID: 2449
Credit: 2,681,934
RAC: 0
|
|
3 hours remain!
Final preparations are ongoing now for the conclusion of the Challenge. For those "moving on" after the Challenge AND have a full cache of work now (3 hours worth), it might be a good time to set "no new work".
Please update your clients early and often to get those completed WU's reported and counted. Don't wait till the end with a large cache of results to submit.
-----------------------------------------------------
A quorum of 2 is NOT needed to award Challenge score - i.e. no double checker. Each returned result will earn a Challenge score. Please note that if the result is eventually declared invalid, the score will be removed. Therefore, the Challenge will not be considered final until all quorum's affecting scoring positions are completed. Because of this, we have a "clean up" period where the goal is to complete these quorums.
We hope for a fast "clean-up". If everyone can remember to either complete their cache or ABORT their WU's, then we'll be able to finalize earlier rather than later. Expect up to a week or more before results become final.
Normally we'd ask for help with the clean-up. However, in this case, we need to lower the server load as soon as the Challenge is over. Please consider heading over to TRP (LLR) for 24 hours to give the validator time to catch. Once load is down and validator is caught up, we'll post the "all clear" to return for the clean up.
Good Luck to everyone as we finish these last remaining hours...
At the Conclusion of the Challenge
We kindly ask users "moving on" to ABORT their WU's instead of DETACHING, RESETTING, or PAUSING.
ABORTING WU's allows them to be recycled immediately; thus a much faster "clean up" to the end of an LLR Challenge. DETACHING, RESETTING, and PAUSING WU's causes them to remain in limbo until they EXPIRE. Therefore, we must wait until WU's expire to send them out to be completed.
Please consider either COMPLETING what's in the queue or ABORTING them. Thank you. :)
____________
|
|
|
|
|
|
John, Iam not happy with the situation total, 7h idle cores in a challenge is hard but in normal run without challenge I can only have 120 tasks total. Some times ago I had over 1000 for cache.
____________
|
|
|
|
|
|
Well, it's been exciting. I think SETI ran into a similar problem some months ago when they issued a bunch of short WUs.
I will join the chorus of "thank you's". These challenges are fun and so what if we run into a bump now and then. Thanks to all involved in running this site and making this fun.
Jim |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13954
ID: 53948
Credit: 392,395,963
RAC: 168,999
|
|
I'll add my thanks, appreciation, and encouragement to everyone helping to keep PG running.
Thanks in particular for quickly recognizing that the stats pages were a big part of the problem and pulling them. It made a HUGE difference -- from my perspective, server availability jumped from around 10% to 90% when you did that.
Moving forward, perhaps the stats pages should be preemptively disabled for short challenges. For the short challenges, it's not that big a deal to have no stats, and since the short challenges also have the short WUs, that's when you'll have the most trouble, too.
For the longer challenges, the server load will be less.
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
This was my first challenge. It was fun. Thanks to all who made it possible. |
|
|
|
|
perhaps the stats pages should be preemptively disabled for short challenges.
I think the stats are a good motivator. I would not mind if they were less often during a short WU Challenge. I would rather see the validator shut down to start the challenge and then run only if the server is doing well.
I also wonder is this a problem that could be fixed by putting a little, or a lot, of money into hardware?
____________
May your addiction to Boinc be greater than mine!
End Transmission!
|
|
|
|
|
I think the stats are a good motivator.
Stats may be a good motivator, but it doesn't do us any good to have stats at the expense of work because the server has locked up. I had to drop out of challenge mode and restart other projects last night after my main machine spent over 3 hours unable to get WUs and all cores went idle. :(
____________
~It only takes one bottle cap moving at 23,000 mph to ruin your whole day~
 |
|
|
rogueVolunteer developer
Send message
Joined: 8 Sep 07
Posts: 1255
ID: 12001
Credit: 18,565,548
RAC: 0
|
|
Is the problem with stats generation or with the size of the HTML response? If the former, is stats generation done upon request (via a browser) or done via a timer (once per x number of minutes)? |
|
|
|
|
|
We Are DONE
yea
____________
From the High Desert in New Mexico
|
|
|
|
|
|
I get "project backoff: 90 minutes" seems like the server has put the white flag up and isn't accepting any more uploads... :) |
|
|
|
|
I also wonder is this a problem that could be fixed by putting a little, or a lot, of money into hardware?
most likely a fast SSD for the DB would help a lot.
otherwise we will face the same chaos in november.. |
|
|
|
|
|
Only the server-admins have insight into the server internals and could break it down to the bottleneck.
At work with our old Sun T2000 webservers the bottleneck might be more the controller than the HDDs (and the slow CPU-cores of course, but two get more parallelization out of the CPU the controller is apparently to slow). |
|
|
|
|
|
Time to start the validation ;)
____________
|
|
|
|
|
|
It already has begun, we are down to 354,714 unvalidated WUs right now. |
|
|
|
|
|
some kind of stats would be nice... |
|
|
|
|
|
Very nice :D |
|
|
|
|
|
I think once the unvalidated WUs-count is down to 10k it would make sense to create stats. ;) |
|
|
|
|
I think the stats are a good motivator.
Stats may be a good motivator, but it doesn't do us any good to have stats at the expense of work because the server has locked up. I had to drop out of challenge mode and restart other projects last night after my main machine spent over 3 hours unable to get WUs and all cores went idle. :(
No work means no stats! :-)
This challenge should be about actually doing work than actually getting it. Luckily this challenge was not important to me and had other BOINC work to do. I am glad that I had made that decision before the challenge started. |
|
|
|
|
|
Thanks for the challenge folks! But FYI - I have four WUs still showing "in progress" on the website, but no more WUs left in the computer. All four were sent to one of my PCs within about a minute. But like I said, no more PG WUs left on the computer. May need to manually clear these off. Here's the WUs:
126007251
126007542
126007662
126006205
____________
All your shrubberies are beholden to us!
The Knights Who Say Ni! |
|
|
|
|
Only the server-admins have insight into the server internals and could break it down to the bottleneck.
At work with our old Sun T2000 webservers the bottleneck might be more the controller than the HDDs (and the slow CPU-cores of course, but two get more parallelization out of the CPU the controller is apparently to slow).
right - but if i remeber correctly, the server is not that old and got quad-cores in two sockets and a decent controller.
it's pretty clear we are talking about IOPS. which leaves us with two choices - throw in a lot of more ram or improve speed of the drive-system.
|
|
|
|
|
I think once the unvalidated WUs-count is down to 10k it would make sense to create stats. ;)
wich would be - next weekend?
that's crazy!
as i stated during several challenges before: no stats, no fun..
this feels like beeing a one legged guy in a n ass-kicking contest who got the extra fun of a blindfold.. :(
|
|
|
|
|
Only the server-admins have insight into the server internals and could break it down to the bottleneck.
At work with our old Sun T2000 webservers the bottleneck might be more the controller than the HDDs (and the slow CPU-cores of course, but two get more parallelization out of the CPU the controller is apparently to slow).
right - but if i remeber correctly, the server is not that old and got quad-cores in two sockets and a decent controller.
it's pretty clear we are talking about IOPS. which leaves us with two choices - throw in a lot of more ram or improve speed of the drive-system.
No if the controller or the driver is shitty and can't handle the IO even a SSD would not help.
And yes the server got 2 Xeon E5410, that is somewhat old but not that slow, as well as 2 x 4 GB FBDIMM 667 RAM (if it wasn't updated since octobre 2008) and Cheetah 15k.5 for the database. 15k.7 is the current generation. Again, the admins have insight on overall load, we can't help it. (at least not without access ;) )
And by the way, the validator is down to 340k, 14k in 40 minutes - if this would be linear we would be down in 12 hours. Keep in mind that the final stats apply when all WUs crunched in the challenge that count against the point-range have double-checks. |
|
|
|
|
No if the controller or the driver is shitty and can't handle the IO even a SSD would not help.
of course not.
And yes the server got 2 Xeon E5410, that is somewhat old but not that slow, as well as 2 x 4 GB FBDIMM 667 RAM (if it wasn't updated since octobre 2008) and Cheetah 15k.5 for the database. 15k.7 is the current generation.
can't tell which chipset it has, but only 2 channels of probably 4 available would be a bad idea and only 8GB for a DB of that size??. go as slicker how much he has thrown in to keep up with the speed of all those fast ATI's..
those cheetahs were pretty much high end 2 years ago, but the outside world has got much faster since then.
|
|
|
|
|
|
I don't know a great deal about computing (or indeed much else) but if there is a problem due to w/us not being validated should we stop working on short w/us and move on to longer ones for a bit so that the server can get its breath back and process the completed short w/us? |
|
|
Scott BrownVolunteer moderator
Project administrator
Volunteer tester
Project scientist
Send message
Joined: 17 Oct 05
Posts: 2380
ID: 1178
Credit: 17,922,113,080
RAC: 10,564,143
                                              
|
I don't know a great deal about computing (or indeed much else) but if there is a problem due to w/us not being validated should we stop working on short w/us and move on to longer ones for a bit so that the server can get its breath back and process the completed short w/us?
Yes...John suggested TRP LLR above, but I think any of the longer work will ease DB issues.
____________
141941*2^4299438-1 is prime!
 |
|
|
STE\/EVolunteer tester
Send message
Joined: 10 Aug 05
Posts: 573
ID: 103
Credit: 3,631,506,192
RAC: 3
|
I also wonder is this a problem that could be fixed by putting a little, or a lot, of money into hardware?
most likely a fast SSD for the DB would help a lot.
otherwise we will face the same chaos in november..
Wasn't there a Drive for Money last year ??? to Upgrade the Servers, not sure about that but if there was apparently it didn't do any good ...
____________
|
|
|
|
|
|
No it was octobre 2008, not last but last last year. ;)
...as times goes by...
And no the mainboard has an Intel 5100 which only has 2 channels and no FBDIMM, only DDR2. I remembered wrong but looked into this matter again. |
|
|
STE\/EVolunteer tester
Send message
Joined: 10 Aug 05
Posts: 573
ID: 103
Credit: 3,631,506,192
RAC: 3
|
No it was octobre 2008, not last but last last year. ;)
...as times goes by...
Whenever it was it didn't do any good as every Challenge since then has had Server Problems ...
____________
|
|
|
HAmstyVolunteer tester
Send message
Joined: 26 Dec 08
Posts: 132
ID: 33421
Credit: 12,510,712
RAC: 0
|
No it was octobre 2008, not last but last last year. ;)
...as times goes by...
Whenever it was it didn't do any good as every Challenge since then has had Server Problems ...
than you havnt participated in every. except March 2010, the challenge series ran really smooth since the new hardware is online
____________
|
|
|
STE\/EVolunteer tester
Send message
Joined: 10 Aug 05
Posts: 573
ID: 103
Credit: 3,631,506,192
RAC: 3
|
No it was octobre 2008, not last but last last year. ;)
...as times goes by...
Whenever it was it didn't do any good as every Challenge since then has had Server Problems ...
than you havnt participated in every. except March 2010, the challenge series ran really smooth since the new hardware is online
I have Participated in just about every Challenge PG has had, and every one of them has had Problems with some Down Time. I'm not Complaining as I expect Down Time when the PG Challenges run so I put up with as it's my choice to run or not.
But what I don't like though is some of you Sugar Coating everything like everything is just oky doky with the Challenges. You run a 1 Day Challenge and the Projects down for 12 Hr's of it, that's not oky doky to me by any stretch of the Imagination.
____________
|
|
|
|
|
(...)
those cheetahs were pretty much high end 2 years ago, but the outside world has got much faster since then.
The manual of the disk tops out at 350 IO/s with random 8-block-reads.
(Our T2000 at work has the Savvio 10k.2 which tops out at 310 IO/s according to the manual - the box provides us with 125 at maximum (dd from device to /dev/null).)
Could be worth investigating if a SSD would enhance database-performance but keep in mind: this is a private effort and not sponsored by a company or university. |
|
|
|
|
|
Similar to other comments, not especially happy with the efforts from yesterday. Projects appear not to know the force of the CPU/GPU community when they turn their magnifying glass onto a challenge (with a blue dwarf on the other side).
Still, I'll keep crunching PG and others - but maybe I'll actively avoid challenge projects in the future. |
|
|
rogueVolunteer developer
Send message
Joined: 8 Sep 07
Posts: 1255
ID: 12001
Credit: 18,565,548
RAC: 0
|
|
I don't think that participants should abandon challenges because problems occur. I think that it is important that a port-mortem is done and that the results of that post-mortem are posted along with potential solutions to any problems to those problems.
|
|
|
|
|
|
I am wondering what was the difference between the last PPS LLR wth year of the Tiger vs this one..
Were the WUs longer, were there more hosts, more cores?
If more hosts and cores, that would explain a lot.
____________
|
|
|
Ken_g6Volunteer developer
Send message
Joined: 4 Jul 06
Posts: 938
ID: 3110
Credit: 258,656,195
RAC: 91,886
        
|
|
I was thinking, I wonder if it would be possible to bundle a few tests together in one WU? Ten 6-minute tests == one 60-minute WU.
On the other hand, I really like the short WUs on my slower machines.
____________
|
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13954
ID: 53948
Credit: 392,395,963
RAC: 168,999
|
I am wondering what was the difference between the last PPS LLR wth year of the Tiger vs this one..
Were the WUs longer, were there more hosts, more cores?
If more hosts and cores, that would explain a lot.
Assume for a second that the number of *people* doing each challenge remains constant.
Over time, people 1) upgrade computers, and 2) buy additional computers. Those upgraded and new computers are not only faster, but increasingly they have more and more cores.
The end result is that even if the number of people doesn't increase, the computational power that each person is bringing to the table increases. Between January and today, I wonder how many people upgraded from a single core Pentium (say, 20 minutes per PPSE LLR) to a brand new i7 (EIGHT PPSE LLRs every 15 minutes)? That's almost an elevenfold increase in the number of WUs done each hour, and is that much more work for the server to do.
It adds up to a lot more work for the server.
In the last challenge, PG shattered their record for sieving. It stands to reason that more computing power was thrown at this PG challenge than any challenge before, and, likewise, the next challenge will have even more resources.
The problems this time were exacerbated by the very short work units.
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
Sorry all. He did and I didn't see it. Apologies.
I don't know a great deal about computing (or indeed much else) but if there is a problem due to w/us not being validated should we stop working on short w/us and move on to longer ones for a bit so that the server can get its breath back and process the completed short w/us?
Yes...John suggested TRP LLR above, but I think any of the longer work will ease DB issues.
|
|
|
STE\/EVolunteer tester
Send message
Joined: 10 Aug 05
Posts: 573
ID: 103
Credit: 3,631,506,192
RAC: 3
|
It stands to reason that more computing power was thrown at this PG challenge than any challenge before, and, likewise, the next challenge will have even more resources
I don't think so because I felt that the Participation was lighter than the last few Challenges so I did some Fact Checking. In the 2010 Earth Day Challenge there were 1631 Participant's. In the 2010 PrimeGrid's Birthday/Summer Solstice Challenge there were 1298 Participant's, and in the last Challenge the 2010 Full Moon Challenge there were 706 Participant's.
With those figures I doubt that the Technology has advanced that much, or the Participants have Upgraded that much in a 2-3 Month Time Period that 706 Participants threw more computing power at the Project than did 1298 & 1631 Participants did in the two previous PG Challenges.
Traditionally the LLR Challenges draw fewer Participant's and the same held true for this LLR Challenge. Shanky & a few other Big Pharm Participants didn't Participate in this Challenge and that alone should have made things run smoother. I think it all came down to the length of the Wu's, simply to short for today's CPU's & the Server couldn't keep up with it even with fewer Participant's than were in the last 2 Challenges.
____________
|
|
|
|
|
|
Most user have upgraded computers since last PPS challenge.They are faster.
We have more BIG user with many hosts.
The big issue this challenge was not validator. It was the stats script taking to much load on DB.
Lennart
|
|
|
STE\/EVolunteer tester
Send message
Joined: 10 Aug 05
Posts: 573
ID: 103
Credit: 3,631,506,192
RAC: 3
|
|
I have the same Computers with the same CPU's in them & the same Amount of Computers that I had 2-3 Years ago. I'm still beating the same Participants I was beating 2-3 Years ago & getting beat by the same Participants too. So if most People have Upgraded & they are faster I think I'm good to go for quit awhile if that's all their bringing to the table ... ;)
____________
|
|
|
|
|
|
Thanks all for the info.
Even though there were fewer F.M> participants, there is more horsepower in the non-challenge sub-projects that would impact challenges. The big crunchers may not have upgraded, but get enough non-challenge users adding a core, a quad or more, it adds to more load on the server..
____________
|
|
|
|
|
I have Participated in just about every Challenge PG has had, and every one of them has had Problems with some Down Time. I'm not Complaining as I expect Down Time when the PG Challenges run so I put up with as it's my choice to run or not.
But what I don't like though is some of you Sugar Coating everything like everything is just oky doky with the Challenges. You run a 1 Day Challenge and the Projects down for 12 Hr's of it, that's not oky doky to me by any stretch of the Imagination.
I have been participating in the Challenge Series ever since I joined PrimeGrid over a year ago now. To my experience, the server has done plenty well on just about every Challenge till now. Despite any hickups, they've been fun.
As Michael and others said, there's a lot of load on the servers, and every Challenge people are upgrading and buying new computers. And this has been going on for two years now since the 2008 server update. Come to think of it, I have switched to a more powerful computer twice now since I joined, each successive computer being roughly twice as fast as the previous one. In addition, the computers that I upgraded from are still crunching.
Then I looked at the PG Milestones & Significant Primes thread, and I saw that the PrimeGrid members with credit has efectively doubled in the last two years as well as all of this. Even if twice the people didn't participate in the Challenge as in 2008, there has to be more.
And then if the Challenge had been run different, say the validator turned off from the start, stats updates less often (for the short WUs)...
So now I am thinking, its a miracle that this two year old server is holding up so well, and only now, after all this time, running the shortest work unit available at PrimeGrid, did the server have serious, Challenge-threatening problems. Even so, it pulled through.
So my opinion is, with as much problems as the server had this challenge, everything can still be "oky doky", under the two folowing circumstances:
1) If the main purpose of the Challenge, to get some serious crunching done and find some primes, is successful
2) People are relatively happy with the outcome, and can look back and say "that challenge had it's problems but it wasn't a waste of my time"
Now I know that the first one was fulfilled, as over 60 primes were found. The second, I am afraid, is up to everyone participating, to decide.
I for one say this Challeng was not a waste of my time, and for me, everything is "oky doky."
____________
May the Force be with you always.
 |
|
|
Scott BrownVolunteer moderator
Project administrator
Volunteer tester
Project scientist
Send message
Joined: 17 Oct 05
Posts: 2380
ID: 1178
Credit: 17,922,113,080
RAC: 10,564,143
|
|
I don't know about others, but I was forced to run on only about half my possible boxes (brilliant office building manager decided to turn off the AC on the weekends to save energy costs and didn't notify most occupants...lost a motherboard and two hard drives last week and a lot of work time from people who couldn't work weekend hours, so I shutdown several to keep them working). Hope the issue with the stats update can be fixed easily. :)
As always, thanks to PG for hosting the challenge!
____________
141941*2^4299438-1 is prime!
|
|
|
|
|
All's working good at the moment. I woke up to 8 cores without work. No biggie, I added two more . Just keep doing what you are doing and I'll do the same. Good luck all. Jack
P.S. It's fun putting a pounding on the server !!!
Ha! Well said, except for me it was only 4 cores. I appreciate everyone's work on this--it was fun, and I look forward to the next challenge. -c
PS- I haven't been able to find the area where I create a signature. Any help would be much appreciated. |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13954
ID: 53948
Credit: 392,395,963
RAC: 168,999
|
PS- I haven't been able to find the area where I create a signature. Any help would be much appreciated.
Click on Community Preferences on your account page. The link is right above the PrimeGrid Preferences link that lets you select the subprojects you want to crunch.
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
Pending credit: 20,507.08 |
|
|
|
|
|
Thx!
____________
-c
|
|
|
|
|
some kind of stats would be nice...
Oh, yea... :)
One day after the challenge, we do not know anything. :(
____________
 |
|
|
|
|
|
looks like the validator is not working again.
one strange thing: i found a lot of WU's which were generated with quorum 1, are reported but will not get validated. |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13954
ID: 53948
Credit: 392,395,963
RAC: 168,999
|
looks like the validator is not working again.
one strange thing: i found a lot of WU's which were generated with quorum 1, are reported but will not get validated.
Th validator is either not working at all, or is still falling behind. Besides the quorum 1 WUs, I have several quoum two WUs with two results which aren't validated, including this SoB WU that I wouldn't mind getting credit for.
In any event, the number of unvalidated WUs on the server stats page is increasing.
____________
My lucky number is 75898524288+1 |
|
|
|
|
In any event, the number of unvalidated WUs on the server stats page is increasing.
no longer - someone kicked it.
|
|
|
STE\/EVolunteer tester
Send message
Joined: 10 Aug 05
Posts: 573
ID: 103
Credit: 3,631,506,192
RAC: 3
|
some kind of stats would be nice...
Oh, yea... :)
One day after the challenge, we do not know anything. :(
Stat's for what ??? Oh that's right we had a Challenge, I forgot already ... LOL ... Looks like the Wu's waiting for Validation are slowly coming down, around 160,000+ still left to go now
____________
|
|
|
gruenyVolunteer tester
Send message
Joined: 12 Mar 08
Posts: 22
ID: 20129
Credit: 596,251
RAC: 0
 
|
|
a task need time.
what a surprise! |
|
|
|
|
|
Hehe - Maybe one should try to feed the
server whats advertised for
to the left here:
" Try it Free-
The perfect marriage between time,
tracking and project management."
;-))
____________
  
|
|
|
|
|
|
got stuck again..
kind regards to murphy.. |
|
|
|
|
|
finally the fog starts to fade away and stats have shown up. :) |
|
|
|
|
|
Wow. 24 hours and 3 primes.. don't know if I'm the double check or not but that is still great. How long till the server will quit smoking?
Jack |
|
|
|
|
|
"Over time, people 1) upgrade computers, and 2) buy additional computers. Those upgraded and new computers are not only faster, but increasingly they have more and more cores."
Hi
I think the above statment is possible. I am running 4 980x 12 cores @4.00mhz 12 WUs in 9mins . All my boxes ran out during the "problem" all 336 WUs.
So next challenge try using some longer WUs , or doubling or tripling the size for the fast Boxes.
I note we have a 48 core AMD box doing some crunching now and I plan to got to 2 chip 6 core xeons 24 threads by christmas.
Ross*
____________
|
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13954
ID: 53948
Credit: 392,395,963
RAC: 168,999
|
|
FYI, I've still got many WUs that haven't validated yet.
They include all of these TRP Sieves, and this SoB.
The sieves are all from after the challenge, so maybe the validator simply hasn't caught up to those, but the SoB's wingman finished near the beginning of the challenge. Since most of the challenge WUs have already been validated, this one seems a bit overdue.
____________
My lucky number is 75898524288+1 |
|
|
|
|
FYI, I've still got many WUs that haven't validated yet.
They include all of these TRP Sieves, and this SoB.
The sieves are all from after the challenge, so maybe the validator simply hasn't caught up to those, but the SoB's wingman finished near the beginning of the challenge. Since most of the challenge WUs have already been validated, this one seems a bit overdue.
i got about 120 out of the challenge most of which have filled their quorum. on the other hand some sieve-WU's reported this morning were validated.
this does look like anything else but FIFO-mode and with a rising number of WU's in the validator queue...
|
|
|
STE\/EVolunteer tester
Send message
Joined: 10 Aug 05
Posts: 573
ID: 103
Credit: 3,631,506,192
RAC: 3
|
"Over time, people 1) upgrade computers, and 2) buy additional computers. Those upgraded and new computers are not only faster, but increasingly they have more and more cores."
Hi
I think the above statment is possible. I am running 4 980x 12 cores @4.00mhz 12 WUs in 9mins . All my boxes ran out during the "problem" all 336 WUs.
So next challenge try using some longer WUs , or doubling or tripling the size for the fast Boxes.
I note we have a 48 core AMD box doing some crunching now and I plan to got to 2 chip 6 core xeons 24 threads by christmas.
Ross*
I have no doubt people are Upgrading their CPU's to Faster & more Core ones even if I haven't because I've gone off in a different direction. Sooooooooo given the Performance of the Server this Challenge then we better hope no more than 100 People Participate in the next PPS (LLR) or the Server will be down 23 Hr's out of the 24 the Challenge runs ... LOL LOL ...
Everybody take anything I say that I say it in Jest, you have to have some fun with these Challenges because it seems the Stats get FUBARed most of the time so there's no fun to have there ... ;)
____________
|
|
|
|
|
FYI, I've still got many WUs that haven't validated yet.
They include all of these TRP Sieves, and this SoB.
The sieves are all from after the challenge, so maybe the validator simply hasn't caught up to those, but the SoB's wingman finished near the beginning of the challenge. Since most of the challenge WUs have already been validated, this one seems a bit overdue.
I have to agree with Michael that the validators seem to have lost track of some pending tasks. Nearly all my (>500) Challenge WUs have validated, only 5 left. But I have a number of sieves completed since the Challenge ended that don't seem to be going anywhere. For example I've completed 6 321 sieves, none of which have validated. The oldest 124762482 has been queued up over 30 hours, while the youngest 124762627 has over 12 hours in queue. Meanwhile I've had other sieves (not 321) completed since the Challenge that validated fairly quickly.
I'm not complaining: I know the servers are struggling. But obviously something other than FIFO is going on here.
____________
  |
|
|
|
|
But obviously something other than FIFO is going on here.
that would have been a way too simple concept which has proved to work in oh so many situations - even in real life.
we are talking about Space Sciences Laboratory, U.C. Berkeley <BEG>
this can only by design - or why do you think you'll easily find
"Every Tuesday morning (Pacific time) we have a 3-4 hour outage for database maintenance."
when searching for YETI and outage?
btw.: 39k and rising... |
|
|
|
|
|
Yes all 321 Sieve tasks I have completed since the end of the Challenge are awaiting validation. Just need 26 more and will get a bronze badge when the validator starts working properly again.
I once asked how long I should wait before enquiring about fully completed work units that remained unvalidated and was told tasks are automatically rechecked once the original deadline expires. So no worries, still 5 days to go before the first ones I completed reach that. |
|
|
|
|
"Every Tuesday morning (Pacific time) we have a 3-4 hour outage for database maintenance."
Actually that's now a 3 day outage, every week ;) |
|
|
|
|
"Every Tuesday morning (Pacific time) we have a 3-4 hour outage for database maintenance."
Actually that's now a 3 day outage, every week ;)
oh - they improved.. ;)
btw.: +3k in the queue in 45 minutes - does anyone think this will work out? |
|
|
|
|
|
This would only start to normalize if the WU-count in the database is not rising anymore, just look at the server-stats-page, we are one step below the summit - i think. |
|
|
|
|
|
What is the ETA on closing out the results of the challenge? Do we know how many more challenge WUs need to be validated?
____________
|
|
|
JohnHonorary cruncher
Send message
Joined: 21 Feb 06
Posts: 2875
ID: 2449
Credit: 2,681,934
RAC: 0
|
What is the ETA on closing out the results of the challenge? Do we know how many more challenge WUs need to be validated?
The status is unknown at this time. We are having validation issues and we suspect that, ironically, it's related more to erroring tasks with PPS (Sieve) than with PPS (LLR)...or simply the combination of the two. We need Rytis to diagnose the issue and to resolve it. Unfortunately, we don't expect him back until after the weekend. :(
Usually a short LLR Challenge takes 10-14 days to finalize and we expect the same with this one. Ideally we'd like to offer the normal "every 15 minutes" stats generation during the clean up but currently it's just too much for the DB to handle.
We certainly are not pleased with the situation and fully understand y'alls displeasure as well. We apologize for the delay.
____________
|
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13954
ID: 53948
Credit: 392,395,963
RAC: 168,999
|
|
We appreciate the update and I'm sure nearly all of us completely understand that you're doing the best you can.
I guess this isn't a terrible time to mention how grateful I am that you're running one of the most fun projects on the Internet.
Thanks!
Mike
____________
My lucky number is 75898524288+1 |
|
|
|
|
[quote]The status is unknown at this time. We are having validation issues and we suspect that, ironically, it's related more to erroring tasks with PPS (Sieve) than with PPS (LLR)...or simply the combination of the two. We need Rytis to diagnose the issue and to resolve it. Unfortunately, we don't expect him back until after the weekend. :(
you are talking about the MAC-CUDA app ?
i know there are still some hosts out there frying them in a matter of seconds. no way to block them?
|
|
|
|
|
[
Usually a short LLR Challenge takes 10-14 days to finalize and we expect the same with this one. Ideally we'd like to offer the normal "every 15 minutes" stats generation during the clean up but currently it's just too much for the DB to handle.
We certainly are not pleased with the situation and fully understand y'alls displeasure as well. We apologize for the delay.
Team Aggie The Pew are just fine about it; every time there is an update the team gets lower down the list. So if you wanted to make the current order final I'm sure the vermin would support you. :)
|
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13954
ID: 53948
Credit: 392,395,963
RAC: 168,999
|
|
Seems like progress is being made, albeit slowly. My SoB and several sieves have been validated.
At least things aren't at a complete standstill. Slow is significantly better than stopped!
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
Lucky me.. Go on vacation for a week, and my ISP cuts contact on day two. So no challenge for me this time. Will be back in force for the next one :) Can't let my cores go to waste too much. Hope to be able to upgrade with more cores and maybe som GPU's for the next challenge.
--
shanky |
|
|
|
|
Lucky me.. Go on vacation for a week, and my ISP cuts contact on day two. So no challenge for me this time. Will be back in force for the next one :) Can't let my cores go to waste too much. Hope to be able to upgrade with more cores and maybe som GPU's for the next challenge.
--
shanky
That's what you get for being a teracruncher--karma ;D But I hope you enjoyed your vacation.
73 |
|
|
|
|
Seems like progress is being made, albeit slowly. My SoB and several sieves have been validated.
Definitely good news. The 321 Sieve validator finally woke up. All my outstanding sieves cleared over(my)night.
____________
|
|
|
Ken_g6Volunteer developer
Send message
Joined: 4 Jul 06
Posts: 938
ID: 3110
Credit: 258,656,195
RAC: 91,886
|
|
You know you can't use GPUs in any challenge until the December challenge, right? (Unless someone's been getting really creative with sr2sieve.)
____________
|
|
|
|
|
|
I had a great vacation, but had a lot of work waiting for me when I got home. Missing network-access was the least of my problems. Had to reinstall all of my compute-nodes. But everything is up and running again finally.
--
shanky |
|
|
|
|
You know you can't use GPUs in any challenge until the December challenge, right? (Unless someone's been getting really creative with sr2sieve.)
Will be at least December before I get the funds and time to upgrade anything unfortunately. Wish I had a HW-supplier who would deliver the latest HW for free in large containers on a regular basis ;)
--
shanky |
|
|
|
|
You know you can't use GPUs in any challenge until the December challenge, right? (Unless someone's been getting really creative with sr2sieve.)
Will be at least December before I get the funds and time to upgrade anything unfortunately. Wish I had a HW-supplier who would deliver the latest HW for free in large containers on a regular basis ;)
--
shanky
If you find one tell my your contact - please! |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13954
ID: 53948
Credit: 392,395,963
RAC: 168,999
|
You know you can't use GPUs in any challenge until the December challenge, right? (Unless someone's been getting really creative with sr2sieve.)
Will be at least December before I get the funds and time to upgrade anything unfortunately. Wish I had a HW-supplier who would deliver the latest HW for free in large containers on a regular basis ;)
--
shanky
If you find one tell my your contact - please!
Before you get all that free computing equipment, you might want to consider what the electric bill for 170,000 credits/day looks like. :)
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
I connect them in the basement, i would only pay 1/25th of the bill with the running-costs of my flat. ;) |
|
|
|
|
Before you get all that free computing equipment, you might want to consider what the electric bill for 170,000 credits/day looks like. :)
You don't need so many computers if they are state of the art at all times, to get high numbers.
If I take a look at my collection of computers I need a lot to get high numbers, because a big part of my equipment is 4+ years old.
--
shanky |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13954
ID: 53948
Credit: 392,395,963
RAC: 168,999
|
You don't need so many computers if they are state of the art at all times, to get high numbers.
Actually, you're right, with the latest equipment it isn't really all that much power.
For laughs, I figured out what it would take to build a small machine that could crank out one million credits per day. That machine being a rack filled with servers, of course.
It turned out to not be difficult if you don't care what it costs. Dell has a nice high density 2U server called the Poweredge C6100 which contains 4 nodes, each of which has 2 Xeon 56xx CPUs. Six of those should do the trick, and occupy less than 2 feet of rack space.
One of those C6100's should be able to more or less match your RAC, and consumes no more than 1700 watts. That's more or less the same as running eight i7-980X's, but more power efficient because of the shared infrastructure.
It's kind of mind-boggling how much computing a single i7-980X or Xeon X5680 can do.
____________
My lucky number is 75898524288+1 |
|
|
|
|
Lucky me.. Go on vacation for a week, and my ISP cuts contact on day two. So no challenge for me this time. Will be back in force for the next one :) Can't let my cores go to waste too much. Hope to be able to upgrade with more cores and maybe som GPU's for the next challenge.
--
shanky
Well, unofficially speaking, we at SETI.USA hope you had a wonderful vacation. It sure was good for us. :-) |
|
|
JohnHonorary cruncher
Send message
Joined: 21 Feb 06
Posts: 2875
ID: 2449
Credit: 2,681,934
RAC: 0
|
|
No real improvement has been made. However, WU's waiting for validation have been slowing going down. It's being micro managed now until Rytis' return.
FYI...the unclaimed primes thread has been updated.
____________
|
|
|
|
|
Lucky me.. Go on vacation for a week, and my ISP cuts contact on day two. So no challenge for me this time. Will be back in force for the next one :) Can't let my cores go to waste too much. Hope to be able to upgrade with more cores and maybe som GPU's for the next challenge.
--
shanky
You would have destroyed the primegrid servers... they would be in a smoking pile of ash along with the database...
____________
|
|
|
|
|
|
Maybe there should be a post-challenge mini-challenge to complete the stats. :) |
|
|
STE\/EVolunteer tester
Send message
Joined: 10 Aug 05
Posts: 573
ID: 103
Credit: 3,631,506,192
RAC: 3
|
|
Looks like they Awarded Challenge Points although the Individual Credits are still the same as they were a week ago when I still had 500 Pending Wu's and I haven't run any since the Challenge ended. But now I don't have any Pending Wu's ...
Weird, I guess those 500 Pending didn't count towards the Challenge even though they were turned in before the Challenge ended ...
____________
|
|
|
RytisVolunteer moderator
Project administrator
Send message
Joined: 22 Jun 05
Posts: 2653
ID: 1
Credit: 95,133,785
RAC: 122,798
         
|
|
Score is granted as soon as a task is returned and is deducted in case the WU is deemed invalid. So your 500 pendings still counted into the final score.
____________
|
|
|
STE\/EVolunteer tester
Send message
Joined: 10 Aug 05
Posts: 573
ID: 103
Credit: 3,631,506,192
RAC: 3
|
|
Okay, thanks for the Clarification ...
____________
|
|
|
|
|
Lucky me.. Go on vacation for a week, and my ISP cuts contact on day two. So no challenge for me this time. Will be back in force for the next one :) Can't let my cores go to waste too much. Hope to be able to upgrade with more cores and maybe som GPU's for the next challenge.
--
shanky
You would have destroyed the primegrid servers... they would be in a smoking pile of ash along with the database...
Probably waiting for the Leonids challenge for that to happen since someone could always blame a meteorite :-)
Hopefully some 'shielding' will be present or change in challenge by then.
|
|
|
JohnHonorary cruncher
Send message
Joined: 21 Feb 06
Posts: 2875
ID: 2449
Credit: 2,681,934
RAC: 0
|
|
The results are Final!
While the results have been final for a while, the Challenge is now officially over. Sorry for the delay in the summary write-up.
On a good note, Proth Primes were abundant with 150+ added to the Top 5000 Primes List!!! Over 813K tasks were completed resulting in almost 3.5M cobblestones awarded. However, the success was tainted by another poor performance on our part.
The lure of the Full Moon was simply too great for the server to handle. For the second time this year, the servers were overwhelmed by the participation. We had not seen problems like this since the Fall of 2008. It appears that when conditions are less than perfect, the server becomes severely stressed.
To give a little perspective about the load the server experienced, here are the top 4 Challenges of tasks completed in a 24 hour period:
550K tasks 15 March 2009 PPS LLR Ides Of March
745K tasks 2-3 October 2009 PPS LLR Harvest Festival
800K tasks 14-17 February 2010 PPS LLR Year of the Tiger
813K tasks 25 July 2010 PPSE LLR Full Moon
The 813K would have been much more if the server could have kept up. While the short tasks stressed the server, a couple of other factors played into the complications we experienced.
Stats
The stats were reworked during the Year of the Tiger Challenge and have been successful for the last two challenges. However, there was an oversight allowing multiple copies of the script to be run. This occurred if the previous run hadn't completed within the 15 minute interval that stats were run. With the stats issue resolved, we saw marginal improvement. However, by that time, the backlog of requests made it difficult for the server to catch back up. Greatly complicating matters were two non-Challenge related issues which brings us to the next factor.
PPSieve and the Validator
PPSieve is a unique sieve application in regards that we're running it without a sieve file. While this saves us the trouble of downloading a 1.2GB sieve file, it does present us with many more superfluous factors to process. This took up much more of the server resources than we anticipated. Additionally, a few of the builds that were released prior to the Challenge were faulty and trashed tasks rather quickly. This obviously produced further strain on the server with the validator attempting to process all of them.
We believe that it was the combination of short tasks, stats oversight, factor processing, and many trashed tasks that caused the server to be overwhelmed. If the latter three issues were resolved, it is unknown if we would have had any issues with just the short tasks.
What has been done to address the issues? 1) Stats have been fixed so that no more that one instance can run at a time. 2) New versions of PPSieve have been released which return fewer superfluous factors. Additionally, the new versions will correct the trashed task issue.
Finally, we'll review hardware and software to see what improvements can occur there.
Since the Fall of 2008, the track record for Challenge conditions has been excellent...with the exception of the 2010 Year of the Tiger Challenge and this current one. In a repeat of the "2010 Year of the Tiger Challenge": The old adages of "Success is short lived" and "you're only as good as your last performance" certainly apply here.
Thank you to everyone for your encouragement and for your criticism. Both are duly noted.
Looking forward, we can foresee only one other Challenge that may have similar issues: 17 November, Leonids Challenge, SGS (LLR). With the stats issue resolved but more importantly, the PPS (Sieve) application more stable and not trashing WU's, we expect a better outcome. We'll definitely do a dry run before the Challenge to see if it's feasible to run it. On the promising side, SGS (LLR) WU's are longer.
144 teams and 918 individuals participated in the Challenge.
Congratulations to SETI.USA as the top team and Mr. Hankey as the top individual. Remember, the top 100 individual and top 50 team places earn points for this Challenge.
Thank you again to everyone. We appreciate your participation and hope y'all are having wonderful summer. Soon The Dog Days of Summer will be here. Come join us 20-22 August for 48 hours on the 321 (Sieve) application!
Stats: Participants | Teams
Top 10 Teams
1 SETI.USA 611210.67
2 SETI.Germany 433546.27
3 Sicituradastra. 285670.48
4 PBToyz 219763.93
5 Major Polish Teams Alliance 179298.10
6 Team 2ch 158958.44
7 BOINC@Poland 130089.09
8 Team Starfire World BOINC 113938.01
9 BOINCstats 104723.35
10 PrimeSearchTeam 97135.11
Top 10 Individuals
1 Mr. Hankey SETI.USA 188704.53
2 Jeff17 PBToyz 136193.50
3 Mumps SETI.USA 94944.87
4 Grzegorz Granowski SaveTheWorld 77847.02
5 Snf Sicituradastra. 71821.61
6 whizbang Ars Technica 64727.52
7 [GPU Vision] Vision GPU Vision 53768.60
8 Siegfried Niklas Crunching Family 53410.27
9 lennart SM5YMT PrimeSearchTeam 51214.13
10 nyabe Team 2ch 47422.52
____________
|
|
|
