Message boards :
Number crunching :
So, IS seti actually recovering from the outages or what?
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 · 8 · Next
Author | Message |
---|---|
[B@H] Ray Send message Joined: 1 Sep 00 Posts: 485 Credit: 45,275 RAC: 0 |
Crystallize Write:
Thay are quite aware of the problem as you would have noticed reading the message board and the Tech. news. Pizza@Home Rays Place Rays place Forums |
ML1 Send message Joined: 25 Nov 01 Posts: 20323 Credit: 7,508,002 RAC: 20 |
... I wonder if the file deleters recheck the directory after every file delete... If so, then your method would give an immediate speedup. Thanks, good idea. I guess I could risk a little more mail and take a look over there... Cheers, Martin See new freedom: Mageia Linux Take a look for yourself: Linux Format The Future is what We all make IT (GPLv3) |
Matt Lebofsky Send message Joined: 1 Mar 99 Posts: 1444 Credit: 957,058 RAC: 0 |
More various notes before I run out for the day: The large directory isn't a problem because we are actually perusing through large directories scanning for files: We know the exact directory and exact filename at every point in the process. It's strictly the filesystem overhead going through its own internal directory scanning to come up with the inode number - nothing much we can do about that unless we change filesystem types, which isn't an option at this point. Of course competition is encouraged in BOINC. I think the issue is that BOINC credits are much more fair and equitable. Classic credit was insanely arbitrary (one credit per WU even if it only took 5 minutes to do), and doled out easily to people who were tricking the system. Apparently some people like that better. Some people like Monopoly because it is all random luck - unexpected events are fun. Others like Chess because it is all skill - you have to earn your win. Anyway let's not get into the credit argument because there is no argument. BOINC credits have to work this way since multiple projects are involved. Another thing: just because a queue is growing doesn't mean we are completely unaware of the problem. Priorities, dependencies, and possibilities are all part of the equation. We apologize for the annoyance, but the validation queue is one tiny part of a really big picture. And it's being dealt with. - Matt -- BOINC/SETI@home network/web/science/development person -- "Any idiot can have a good idea. What is hard is to do it." - Jeanne-Claude |
ML1 Send message Joined: 25 Nov 01 Posts: 20323 Credit: 7,508,002 RAC: 20 |
More various notes before I run out for the day: Ouch. Hence the present story. BOINC credits are much more fair and equitable. Classic credit was insanely arbitrary ... BOINC credits have to work this way since multiple projects are involved. Agreed here at least. Hopefully the fiddle-factor fix will give a quick fix (even if a little bit of a dirty fix) for getting the time/credits estimates more accurate. The other new API calls for benchmarking/credits hopefully will get usefully used for a full fix. just because a queue is growing doesn't mean we are completely unaware of the problem... As described, it might not even directly be a problem let alone "the" problem. :) I guess you all have a very hard time with a few thousand of us 'back seat drivers' to keep humored ;) Thanks for sharing some the inside details and present problems that are being worked on. Can I add an analogy of there only being two jugglers for 20 hot potatoes?! ;) Keep hacking, Regards, Martin See new freedom: Mageia Linux Take a look for yourself: Linux Format The Future is what We all make IT (GPLv3) |
AndyK Send message Joined: 3 Apr 99 Posts: 280 Credit: 305,079 RAC: 0 |
@AndyK No probs, dude. You are welcome. But I wanted to count my pending credits and therefore I made a little php app on my own. The difference is: - it's php! No problems with local number formats - You can set the number of hosts you want to query, the totals are displayed as well! :-)) I've tested it only with up to three hosts, so I don't know how long it takes with more. The limit in execution time is 60s for the script. If you encounter the timeout error, try again with a lower number of hosts. The script should be accessible at: http://ztb.dyndns.org/pc.php Try it and give me some info if it's doing it's job, how you like it, etc... Andy Want to know your pending credit? The biggest bug is sitting 10 inch in front of the screen. |
Roswitha Send message Joined: 27 Dec 01 Posts: 10 Credit: 269,207 RAC: 0 |
I've tried it with the Host-ID 612070 (with Comma as decimal seperator) and got this: Pending credit: 2548,7141 Granted credit: 30644,56 Total credit: 33193,2741 My System: WinXP Pro SP2, German regional settings Thanks CJ for this nice Tool! Greetings Roswitha |
Team Cake Boy's Send message Joined: 24 Jul 99 Posts: 22 Credit: 6,923,381 RAC: 0 |
i have 23 hosts and it keeps timeing out!! @AndyK |
Team Cake Boy's Send message Joined: 24 Jul 99 Posts: 22 Credit: 6,923,381 RAC: 0 |
Just tried the downloadable one Simon 3218 Anton 4162 Pikachu 2236 Grenadier 1023 Tux 0000 Steamer 0028 Drowzee 1532 Rob 1874 Mary 0000 Latitude 0387 Isengard 0083 Evo 0817 Leah 1801 Workshop 0374 Office 0051 17581 Outstanding Credit i have 23 hosts and it keeps timeing out!! |
AndyK Send message Joined: 3 Apr 99 Posts: 280 Credit: 305,079 RAC: 0 |
i have 23 hosts and it keeps timeing out!! As i posted: execution time of this script is limited by the server. If you get timeout, try a lower number of hosts. I increased the execution time to 300 seconds and added a routine that the script will abort collecting data 5 seconds before timeout. So you will have at least the results for the already collected data. edit: I made the test with your 23 hosts and it worked now. It took 150 seconds. hostID pending credit granted credit total credit 936845 2,700.60 45,916.95 48,617.55 1213576 3,918.87 21,478.66 25,397.53 943343 2,853.23 34,338.29 37,191.52 1208403 737.85 4,847.44 5,585.29 1052772 0.00 14,687.58 14,687.58 1150978 0.00 3,622.53 3,622.53 1236979 1,751.13 29,211.34 30,962.47 941042 2,120.00 13,281.36 15,401.36 1202279 0.00 2,155.34 2,155.34 1005846 368.03 4,745.43 5,113.46 1220075 105.14 4,561.61 4,666.75 1112483 28.10 818.59 846.69 1265317 837.10 19.51 856.61 1262604 2,468.59 80.51 2,549.10 60498 66.02 269.43 335.45 1253462 0.00 60.06 60.06 32957 67.76 6,127.12 6,194.88 165761 0.00 3,870.87 3,870.87 131427 44.31 4,949.75 4,994.06 1265163 538.35 0.00 538.35 1267821 106.89 0.00 106.89 1294024 84.97 0.00 84.97 1296373 0.00 0.00 0.00 Total 18,796.94 195,042.37 213,839.31 Andy Want to know your pending credit? The biggest bug is sitting 10 inch in front of the screen. |
David @ TPS Send message Joined: 30 Sep 04 Posts: 70 Credit: 11,323,275 RAC: 0 |
THANK YOU MATT... finally someone says it like it is. I have been told multiple times both on this board and the classic board that the competition in classic is what caused the "cheating" so needed to be disposed of. In the almost a year that I have been on classic, I have not SEEN any "cheating", but that does not mean it didnt happen, HOWEVER, I resent ALL classic crunchers being branded as CHEATERS because a very few found a way to "trick the system" which seems to be what has happened. When the time comes that I bring my 30 computer "farm" to BOINC will I be accused of cheating because my rate HOPEFULLY will be high? Yes, when I put in all new P-4 HT's and my rate went from 15 WU a day to over 100, there were some raised eyebrows...... I am not interested in crunching for other "programs", I joined SETI and that is what I want to do, so I WILL bring my farm to SETI BOINC, but I will probably still have a few WU's in the quene when the plug is finally pulled. The one thing I do not understand is ANYONE wanting to be the 4th recipient of a Boinc WU which will not get credited. If competition is encouraged, why would anyone want to spend the cycles crunching something that will not receive any credit? If my understanding of that is wrong, I would LOVE to hear to the contrary! I want to crunch for seti, I started with Classic, and I will see it to its end. Then I will join SETI BOINC and do what I can. If I can strive for goals and have the friendly competition that I am used to, it will make the investment of time and resources more enjoyable. David |
Plasmo Send message Joined: 8 Jul 02 Posts: 16 Credit: 1,086,722 RAC: 0 |
My computers have not been able to contact SETI for 1-2 weeks, problems with timing out or schedulers are not responding. I cannot even report my completed workunits. Now the results are too late (past the due date) and I will never get credit for work done. I am switching my computers to einstein at home until SETI becomes more reliable. Visit www.greenpeace.org |
Terry Send message Joined: 17 Sep 00 Posts: 153 Credit: 1,805,202 RAC: 0 |
The one thing I do not understand is ANYONE wanting to be the 4th recipient of a Boinc WU which will not get credited. If competition is encouraged, why would anyone want to spend the cycles crunching something that will not receive any credit? If my understanding of that is wrong, I would LOVE to hear to the contrary! I want to crunch for seti, I started with Classic, and I will see it to its end. Then I will join SETI BOINC and do what I can. If I can strive for goals and have the friendly competition that I am used to, it will make the investment of time and resources more enjoyable. Where did you obtain such misinformation? The 4th recipient gets credit as long as he/she returns their result before the deadline. |
MJKelleher Send message Joined: 1 Jul 99 Posts: 2048 Credit: 1,575,401 RAC: 0 |
The one thing I do not understand is ANYONE wanting to be the 4th recipient of a Boinc WU which will not get credited. If competition is encouraged, why would anyone want to spend the cycles crunching something that will not receive any credit? If my understanding of that is wrong, I would LOVE to hear to the contrary! I want to crunch for seti, I started with Classic, and I will see it to its end. Then I will join SETI BOINC and do what I can. If I can strive for goals and have the friendly competition that I am used to, it will make the investment of time and resources more enjoyable. Ok, your understanding is wrong! Here is a link to the Rules as set out in the Wiki. In brief, so long as that fourth work unit is returned by the time limit, it will still get credit. If it gets submitted after the establishment of a quorum, it'll get the credit set by the first three units. If the quorum hadn't been validated, then its results will contribute to the setting of the credit for that work unit. Some units won't receive credit, because their results don't come near the others returned, or in cases of errors in processing. Usually that's the result of problems with the host computer, overclocking being one source of problems. But that has nothing to do with the order in which the results are returned. It's in keeping with the policy of giving credit for valid science returned. Hope this relieves your concerns! MJ |
Contact Send message Joined: 16 Jan 00 Posts: 196 Credit: 2,249,004 RAC: 0 |
|
Colossus Send message Joined: 8 Jul 05 Posts: 73 Credit: 23,618 RAC: 0 |
My computers have not been able to contact SETI for 1-2 weeks, problems with timing out or schedulers are not responding. I cannot even report my completed workunits. Now the results are too late (past the due date) and I will never get credit for work done. I am switching my computers to einstein at home until SETI becomes more reliable. I'm wondering if the scheduler issues for some users are related to their internet connection. I'm using a cable modem & rarely get the "scheduler not responding" message. When it does occur, the next try usually succeeds. Also, I'm not using a proxy. Perhaps it is timing out like a ping that doesn't respond in time. This is the voice of world control. I bring you peace. It may be the peace of plenty and content or the peace of unburied dead. The choice is yours. |
MJKelleher Send message Joined: 1 Jul 99 Posts: 2048 Credit: 1,575,401 RAC: 0 |
My computers have not been able to contact SETI for 1-2 weeks, problems with timing out or schedulers are not responding. I cannot even report my completed workunits. Now the results are too late (past the due date) and I will never get credit for work done. I am switching my computers to einstein at home until SETI becomes more reliable. What have you tried to correct this? Could you post the relevent entries on the messages tab in BOINC Manager? There had been many problems for a while, for some of us, in connecting, which were resolved by going through a proxy. Those seem to have been cleared up in the past few days, though. MJ |
Drew Send message Joined: 31 May 04 Posts: 67 Credit: 95,871 RAC: 0 |
Priorities, dependencies, and possibilities are all part of the equation. We apologize for the annoyance, but the validation queue is one tiny part of a really big picture. And it's being dealt with. Thanks Matt for the great info! |
[B@H] Ray Send message Joined: 1 Sep 00 Posts: 485 Credit: 45,275 RAC: 0 |
In the almost a year that I have been on classic, I have not SEEN any "cheating", but that does not mean it didnt happen, HOWEVER, I resent ALL classic crunchers being branded as CHEATERS because a very few found a way to "trick the system" which seems to be what has happened. When the time comes that I bring my 30 computer "farm" to BOINC will I be accused of cheating because my rate HOPEFULLY will be high? Yes, when I put in all new P-4 HT's and my rate went from 15 WU a day to over 100, there were some raised eyebrows...... David There are others doing around a hundred a day on BOINC and I have never heard one being called a cheater. But I am sure that most others wish that they could do that many also. I just came over from Classis in early June, and no one has called me a cheatter because I was there longer. I never used the boards at classic and when I came over here thare was nothing on the main page about BOINC, I heard about this from a friend. How many other Classic users never use the web site to see that BOINC is here? I have no idea but there are a lot of them. I was the first from my team to find out about it, and the only one to change over. So please let others know, there are a lot who don't know of this. If you join Boinc Synergy they will include your classic units on the signature like my 5101 are below. Cheers Ray Pizza@Home Rays Place Rays place Forums |
Matt Lebofsky Send message Joined: 1 Mar 99 Posts: 1444 Credit: 957,058 RAC: 0 |
HOWEVER, I resent ALL classic crunchers being branded as CHEATERS because a very few found a way to "trick the system" which seems to be what has happened. I hope that's not how I came across. It's pretty easy for me to tell cheaters from non-cheaters in classic, and I know full well that most people with high numbers did so honestly. - Matt -- BOINC/SETI@home network/web/science/development person -- "Any idiot can have a good idea. What is hard is to do it." - Jeanne-Claude |
KB7RZF Send message Joined: 15 Aug 99 Posts: 9549 Credit: 3,308,926 RAC: 2 |
Priorities, dependencies, and possibilities are all part of the equation. We apologize for the annoyance, but ... it's being dealt with. But according to others who like to bash and bitch and complain that nothing works, this statement don't matter. I agree, its a very good slogan, cause its true. I have always enjoyed crunching for SETI, and when BOINC came out and I understood it more, I installed it, and very happy I did. Now I happily crunch Seti and Einstein. I'm still learning, and still trying to understand how it all works, but my computers are happy, and I'm happy I'm donating my computers time. Jeremy |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.