| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Active Research
Forum: OpenPandemics - COVID-19 Project
Thread: Work unit availability |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 822
|
|
| Author |
|
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges: 
|
in the current situation more hosts would actually be more beneficial to snagging more tasks than using a single command line to fix BOINC's broken request behavior Amazing. You've been here less than a year and you think you know how things run/are supposed to run. The 2 minute request delay is to prevent the server from being hammered with endless non stop requests for work when projects only have a limited supply like OPNG. Nothing is broken and the endless whining won't change a thing.
In 1969 I took an oath to defend and protect the U S Constitution against all enemies, both foreign and Domestic. There was no expiration date.
  |
||
|
|
Ian-n-Steve C.
Senior Cruncher United States Joined: May 15, 2020 Post Count: 180 Status: Offline Project Badges: 
|
in the current situation more hosts would actually be more beneficial to snagging more tasks than using a single command line to fix BOINC's broken request behavior Amazing. You've been here less than a year and you think you know how things run/are supposed to run. The 2 minute request delay is to prevent the server from being hammered with endless non stop requests for work when projects only have a limited supply like OPNG. Nothing is broken and the endless whining won't change a thing. The two minute timer is not something I have an issue with. This is something that is configurable from the project side. Some projects run 5 mins. Some run 10 seconds. But again that’s not my concern. And I’ve been a part of BOINC for a long time and know better than most how the BOINC software works and “thinks” through exhaustive observation and testing. My issue with with how BOINC handles repeated schedule requests that end in no tasks being sent. After a few requests that don’t get tasks, BOINC starts artificially extending the schedule request delay. This has nothing to do with the project servers. This is handled on the HOST and independent of the project. If the project has decided they don’t want a host to request quicker than 2 minutes, no problem. Nothing I have done or could do, changes that. Because if I try to make a request sooner, the project already detects this and ignores the request. You actually will end up never getting work if you try to circumvent this 2 minute timer as every request will be classified as “too soon” and you’ll never get work sent to you. So I simply issue a “retry” command every 5 minutes. Longer than the 2 minute default timer. This has the effect of keeping BOINC on its toes and never lets it get into artificial extended backoffs. Literally the only people whining are the ones upset that I’ve worked around the problem. And I’m using tools provided by BOINC themselves to do it. I have absolutely no problem with whatever timer the project decides . I’m just forcing BOINC to stick to that and not extend that for me.  EPYC 7V12 / [5] RTX A4000 EPYC 7B12 / [5] RTX 3080Ti + [2] RTX 2080Ti EPYC 7B12 / [6] RTX 3070Ti + [2] RTX 3060 [2] EPYC 7642 / [2] RTX 2080Ti |
||
|
|
kittyman
Advanced Cruncher Joined: May 14, 2020 Post Count: 140 Status: Offline Project Badges: 
|
Boinc is working as it is intended to. The 'artificial extended backoffs' are built into it so that if the work being requested is not readily available, there are not tens of thousands of computers banging the hell out of the server trying to get some anyway.
----------------------------------------If every computer on the project were running the way you have chosen to set yours up, the project servers would likely be DOSed into a tizzy and anybody would be lucky to get anything out of it. Just because the normal server backoff is 2 minutes, that doesn't mean the system expects every host connected to the project to whack away at it every 2 minutes looking for work that isn't available. Just because you CAN doesn't mean you SHOULD. And no, that's not whining. That just the facts.  [Edit 1 times, last edit by kittyman at Apr 11, 2021 3:34:28 AM] |
||
|
|
Ian-n-Steve C.
Senior Cruncher United States Joined: May 15, 2020 Post Count: 180 Status: Offline Project Badges:
|
Boinc is working as it is intended to. The 'artificial extended backoffs' are built into it so that if the work being requested is not readily available, there are not tens of thousands of computers banging the hell out of the server trying to get some anyway. If every computer on the project were running the way you have chosen to set yours up, the project servers would likely be DOSed into a tizzy and anybody would be lucky to get anything out of it. Just because the normal server backoff is 2 minutes, that doesn't mean the system expects every host connected to the project to whack away at it every 2 minutes looking for work that isn't available. Just because you CAN doesn't mean you SHOULD. And no, that's not whining. That just the facts. The project decides their own back off. Not BOINC. They’ve decided that 2 mins is the limit for whatever reasons. And I’m in no way breaking that. Literally zero harm. It’s no extra load for the servers than if the server was actually giving me work every request, because if there was enough work to fulfill a request every time it tried, it would naturally stick to the 2min schedule. Obviously nothing bad is happening because I’m still sticking to the limits that the PROJECT has set. This fantasy that doing this has ANY quantifiable negative effect is entirely without merit. Anyone who thinks so still doesn’t understand what’s actually being done. That’s the facts. EPYC 7V12 / [5] RTX A4000 EPYC 7B12 / [5] RTX 3080Ti + [2] RTX 2080Ti EPYC 7B12 / [6] RTX 3070Ti + [2] RTX 3060 [2] EPYC 7642 / [2] RTX 2080Ti [Edit 1 times, last edit by Ian-n-Steve C. at Apr 11, 2021 3:47:13 AM] |
||
|
|
sam6861
Advanced Cruncher Joined: Mar 31, 2020 Post Count: 107 Status: Offline Project Badges: 
|
The overall daily points generated is going down.
https://www.worldcommunitygrid.org/stat/viewP....do?projectShortName=opn1 My 3 computers are barely getting 10 OPNG task per day. A few days ago, used to get about 50 OPNG tasks. |
||
|
|
goben_2003
Advanced Cruncher Joined: Jun 16, 2006 Post Count: 145 Status: Offline Project Badges: 
|
I could do that if I was only interested in keeping my GPU busy. I am doing this project because I am interested in contributing to the efforts against COVID. Largely due to the fact that my significant other nearly died from it, and has been fighting it for a month now. I am sorry to hear that and I hope they get better soon! One of my friends just got out of the ICU after being there for over a month due to COVID-19. He still has a long recovery ahead before he can leave the hospital. As I said though, you can contribute to the efforts to COVID with folding@home in the meantime. I run folding@home on my gpu(you can select COVID-19 as the cause) I am currently running a work unit for https://stats.foldingathome.org/project/17339. There are also projects for the Covid Moonshot project such as https://stats.foldingathome.org/project/13447 As well as other ones like https://stats.foldingathome.org/project/16604 So there is plenty of GPU work available if you want to contribute to the efforts against COVID while OPNG gets up to speed.  |
||
|
|
Andrew80431
Cruncher Joined: Nov 25, 2005 Post Count: 36 Status: Offline Project Badges:
|
So I added Environment="CLOVER_PLATFORM_VERSION_OVERRIDE=1.2" to "/etc/systemd/system/boinc-client.service" and did a systemctl daemon-reload Now BOINC's event log shows: OpenCL: AMD/ATI GPU 0: AMD CAYMAN (DRM 2.50.0 / 5.4.0-67-generic, LLVM 10.0.0) (driver version 20.0.8, device version OpenCL 1.2 Mesa 20.0.8, 2048MB, 2048MB available, 768 GFLOPS peak) I'm curious if this will work... I'll keep you posted. Well today I got an AMD WU and unfortunately it did not go too well. While the Radeon card was trying to process said WU, the server locked up twice. So for the time being I removed the HD6950 and am trying to get my hands on an affordable NVIDIA GFX card... (which probably won't happen anytime soon)  |
||
|
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges:
|
So I added Environment="CLOVER_PLATFORM_VERSION_OVERRIDE=1.2" to "/etc/systemd/system/boinc-client.service" and did a systemctl daemon-reload Now BOINC's event log shows: OpenCL: AMD/ATI GPU 0: AMD CAYMAN (DRM 2.50.0 / 5.4.0-67-generic, LLVM 10.0.0) (driver version 20.0.8, device version OpenCL 1.2 Mesa 20.0.8, 2048MB, 2048MB available, 768 GFLOPS peak) I'm curious if this will work... I'll keep you posted. Well today I got an AMD WU and unfortunately it did not go too well. While the Radeon card was trying to process said WU, the server locked up twice. So for the time being I removed the HD6950 and am trying to get my hands on an affordable NVIDIA GFX card... (which probably won't happen anytime soon) You might have better luck with AMD on Linux by upgrading to 20.04 and following the tutorial here
In 1969 I took an oath to defend and protect the U S Constitution against all enemies, both foreign and Domestic. There was no expiration date.
|
||
|
|
floyd
Cruncher Joined: May 28, 2016 Post Count: 47 Status: Offline Project Badges:
|
Well today I got an AMD WU and unfortunately it did not go too well. While the Radeon card was trying to process said WU, the server locked up twice. I can only say that I processed dozens of tasks onAMD VERDE (DRM 2.50.0, 4.19.0-16-amd64, LLVM 7.0.1) (Radeon HD 7750) without any problem and the results were 100% valid. But that was on a system without X.trying to get my hands on an affordable NVIDIA GFX card... (which probably won't happen anytime soon) Wait until the project needs more computing power (which probably won't happen anytime soon). |
||
|
|
biini
Senior Cruncher Finland Joined: Jan 25, 2007 Post Count: 334 Status: Offline Project Badges: 
|
I'd need help for getting some GPU tasks. For past five days, I've got only four tasks.
----------------------------------------It seems that my client wont poll for new tasks too often... it's already on hour from last poll. Is there a way to make client connect more often? 11.4.2021 18:38:30 | World Community Grid | Requesting new tasks for NVIDIA GPU 11.4.2021 18:38:32 | World Community Grid | Scheduler request completed: got 0 new tasks 11.4.2021 18:38:32 | World Community Grid | No tasks sent 11.4.2021 18:38:32 | World Community Grid | No tasks are available for OpenPandemics - COVID 19 11.4.2021 18:38:32 | World Community Grid | No tasks are available for OpenPandemics - COVID-19 - GPU 11.4.2021 18:38:32 | World Community Grid | Tasks for CPU are available, but your preferences are set to not accept them 11.4.2021 18:38:32 | World Community Grid | Project requested delay of 121 seconds 11.4.2021 19:13:56 | World Community Grid | Sending scheduler request: To fetch work. 11.4.2021 19:13:56 | World Community Grid | Requesting new tasks for NVIDIA GPU 11.4.2021 19:13:58 | World Community Grid | Scheduler request completed: got 0 new tasks 11.4.2021 19:13:58 | World Community Grid | No tasks sent 11.4.2021 19:13:58 | World Community Grid | No tasks are available for OpenPandemics - COVID 19 11.4.2021 19:13:58 | World Community Grid | No tasks are available for OpenPandemics - COVID-19 - GPU 11.4.2021 19:13:58 | World Community Grid | Tasks for CPU are available, but your preferences are set to not accept them 11.4.2021 19:13:58 | World Community Grid | Project requested delay of 121 seconds (local time now 20:13... so again an hour from last poll) |
||
|
|
|