Message boards :
Number crunching :
Delay in getting new work units untill all work units have cleared
Message board moderation
Author | Message |
---|---|
Send message Joined: 14 Aug 12 Posts: 10 Credit: 10,052,995 RAC: 0 |
Is there some reason for this problem to exist without the Administrator making some change to fix it? Given that volunteers are offering free computing to the project, which cost the volunteer some amount of money, I simply can't understand why this situation is allowed to exist. Given that I arrived only recently to the discussion I must be missing something. |
Send message Joined: 24 Jan 11 Posts: 715 Credit: 555,880,723 RAC: 42,619 |
The system administrators just recently updated the server software to version 1.04 during the summer and it took them a month to reconfigure everything to somewhat working condition again from the previous server code. Give them some time to figure things out again. They will need to update again since the BOINC server code branch 1.20 has been released. |
Send message Joined: 2 Oct 16 Posts: 167 Credit: 1,008,062,758 RAC: 1,016 |
|
Send message Joined: 14 Aug 12 Posts: 10 Credit: 10,052,995 RAC: 0 |
I read the pages associated with the link. Interesting but I am not geek enough to understand all that. Keith Myers actually answered the question. |
Send message Joined: 24 Jan 11 Posts: 715 Credit: 555,880,723 RAC: 42,619 |
https://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=4424 Can you try the latest 7.15.0 client artifact if you are running Windows. It has the most recent work_fetch module. https://ci.appveyor.com/api/buildjobs/4bvvgoug1ej0x5mh/artifacts/deploy%2Fwin-client%2Fwin-client_master_2019-09-18_15ffc98a.7z |
Send message Joined: 13 Dec 12 Posts: 101 Credit: 1,782,758,310 RAC: 0 |
I think this topic is why I came looking on the forum. I've noticed I keep running out of WU's and more aren't being pushed out? (or my pc isn't fetching more...however it works) So quite often I have one or more PC's just sitting idle until I manually hit refresh.....then it fetches more work. All my PC's are running 7.14.2. It's a bit annoying. Edit - I see people know about the issue. Oh well....I'll just keep truckin. |
Send message Joined: 8 May 09 Posts: 3339 Credit: 524,010,781 RAC: 0 |
I think this topic is why I came looking on the forum. I've noticed I keep running out of WU's and more aren't being pushed out? (or my pc isn't fetching more...however it works) Setup a 2nd gpu project with a zero resource share so when MW runs out it will get tasks and your gpu won't be idle while waiting for MW to refill the cache. PrimeGrid and Collatz are two projects that almost always have tasks, at PrimeGrid you can pick wu's that run very quickly like the MW wu's do and at Collatz if you do use the optimization codes the wu's will run much faster as well. |
Send message Joined: 23 Feb 18 Posts: 26 Credit: 4,744,416,145 RAC: 0 |
Setup a 2nd gpu project with a zero resource share so when MW runs out it will get tasks and your gpu won't be idle while waiting for MW to refill the cache. PrimeGrid and Collatz are two projects that almost always have tasks, at PrimeGrid you can pick wu's that run very quickly like the MW wu's do and at Collatz if you do use the optimization codes the wu's will run much faster as well. This is just a workaround, not a solution.. developers must fix it This is for sure a server-side misconfiguration, in 15years of boinc it never happened with any other project Want your Kids stay off from Drugs? Get them building Crunching PC's and they'll never have enough money for drugs |
Send message Joined: 23 Feb 18 Posts: 26 Credit: 4,744,416,145 RAC: 0 |
Double post Want your Kids stay off from Drugs? Get them building Crunching PC's and they'll never have enough money for drugs |
Send message Joined: 2 Oct 16 Posts: 167 Credit: 1,008,062,758 RAC: 1,016 |
Setup a 2nd gpu project with a zero resource share so when MW runs out it will get tasks and your gpu won't be idle while waiting for MW to refill the cache. PrimeGrid and Collatz are two projects that almost always have tasks, at PrimeGrid you can pick wu's that run very quickly like the MW wu's do and at Collatz if you do use the optimization codes the wu's will run much faster as well. I do this for every client and its always a good idea no matter your main project. |
Send message Joined: 8 May 09 Posts: 3339 Credit: 524,010,781 RAC: 0 |
Setup a 2nd gpu project with a zero resource share so when MW runs out it will get tasks and your gpu won't be idle while waiting for MW to refill the cache. PrimeGrid and Collatz are two projects that almost always have tasks, at PrimeGrid you can pick wu's that run very quickly like the MW wu's do and at Collatz if you do use the optimization codes the wu's will run much faster as well. Of course it is but your post said your gpu was idle while waiting for MW workunits, my solution solves the gpu idle problem. MW must fix their own Project. |
Send message Joined: 13 Dec 12 Posts: 101 Credit: 1,782,758,310 RAC: 0 |
Hi Mikey, So can you set 0 resource share for a specific project? At the moment I have my BOINC settings to store 10 days work + another days days work (probably excessive) but it's all I could think of to try and download as many WU's as possible. The Radeon VII just eats them for breakfast.Are you suggesting that I set this to 0 ? If I did this, would Milkyway keep running out of work or choose Primegrid wu's over Milkyway? I understand what you are suggesting and I'd probably pick PPS Sieve WU's as they take no time at all but I'm not sure how to set it up correctly. I've checked the Primegrid settings and can't see that I can set 0 resource share there. TIA |
Send message Joined: 18 Nov 08 Posts: 291 Credit: 2,461,693,501 RAC: 0 |
I've checked the Primegrid settings and can't see that I can set 0 resource share there. At www.primegrid.com log in to your account and click on "project preferences". At the top under "Primary" you should see Resource share 100 It can be changed by scrolling down and looking for "Edit PrimeGrid Preferences" Change the 100 to 0 and scroll to the very bottom and click on "update project preferences" |
Send message Joined: 13 Dec 12 Posts: 101 Credit: 1,782,758,310 RAC: 0 |
LOL...right at the top of the page! I completely overlooked it. Thank you ! |
Send message Joined: 13 Dec 12 Posts: 101 Credit: 1,782,758,310 RAC: 0 |
Hmm.... with resource share set to 0, I'm finding on 2 of my PC's that BOINC is pausing the Milkyway@Home wu's and running the Primegrid work instead. Should that happen? It's not what we want is it. TIA |
Send message Joined: 18 Nov 08 Posts: 291 Credit: 2,461,693,501 RAC: 0 |
Hmm.... with resource share set to 0, I'm finding on 2 of my PC's that BOINC is pausing the Milkyway@Home wu's and running the Primegrid work instead. Should that happen? I assume you are not using a project manager.. When resources are set to 0 for a project then only one WU is downloaded and only if the primary project is out of data. If the primary project has suspended tasks then there is no download either. That work unit (prime), once started, will run to completion unless it takes more than 60 minutes in which case it will be suspended. Once suspended, there is a possibility it will never complete in time, as it will be serviced only when the primary project is out of data. If it finishes only an upload will occur unless the primary project is still out of data. You should have only one prime work unit total (for each GPU). If you have any queued up to run that is another problem. I assume you have only one prime grid task running per gpu and none queued up to run. Prime grid takes an entire GPU. When I used to run Prime I had to force nVidia fans at %100 with the side case off and a box or tornado fan blowing air in. I assume nothing has changed since then other than Prime opting out of gridcoin mining. Suggest you run Einstein as secondary project rather than prime. It will not need a box fan or an open mining rig for cooling unlike prime and is also eligible for gridcoin rewards. My understating is the gridcoin people are allowing miners to keep their membership in local clubs. I had to drop my Texas A&M membership to mine while crunching. Note that once milkyway downloads new work units, it will NOT stop the secondary task from running You mentioned that a few MW tasks were suspended. That can happen when attaching for first time or changing venue where the new venue has 0 for resources. I have found it requires two "project updates" the first time a connection is made (new project) The first connection gets task but does not set the resource to 0. The second "project update" gets "resource = 0" . Unfortunately, you will be stuck with an unwanted task or two. Using a project manager like !BAM can complicate things as BAM must be told to set the resource to 0 using a "project manager sync". Once a few unwanted Prime tasks arrive they will get serviced because they are there already but eventually they will finish or get suspended and MW tasks will take over. |
Send message Joined: 24 Jan 11 Posts: 715 Credit: 555,880,723 RAC: 42,619 |
My understating is the gridcoin people are allowing miners to keep their membership in local clubs Not fully implemented yet. Still in beta testing for a few outside teams. Going well. Probably won't have the new general release open team clients available till the first of the new year. |
Send message Joined: 24 Jul 12 Posts: 40 Credit: 7,123,301,054 RAC: 0 |
Has anyone seen any response from a Milkyway administrator/developer/moderator on this issue since Jake left the project? Joe |
Send message Joined: 10 Apr 19 Posts: 408 Credit: 120,203,200 RAC: 0 |
We have been monitoring the situation, and it seems like the community has found fixes to some of the problems you are experiencing. Jake said that the problem appeared to be some obscure BOINC setting somewhere, and had asked BOINC forums about it. It looks like this issue disappears in the new beta of a BOINC client, so they must have patched whatever was causing problems. When that is released, hopefully the problem will be resolved. - Tom |
Send message Joined: 24 Jan 11 Posts: 715 Credit: 555,880,723 RAC: 42,619 |
JStateson tried the Windows AppVeyor artifact for the 7.15.0 development client at my suggestion and ran into the same issue of work unit starvation. The work_fetch.cpp module (4/20/2019) that controls work fetch has not changed between the 7.15.0 development branch and the latest 7.16.2 development branch in the past 5 months and which is scheduled to go mainline hopefully soon once the translations are finished. There may be more interactions with the other modules in the client involved with work fetch that have been patched or changed in the current development branch that may fix the issue. But I have my doubts. I believe there are still server side misconfigurations that the new expected client will not fix. |
©2024 Astroinformatics Group