| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Active Research
Forum: OpenPandemics - COVID-19 Project
Thread: Invalid GPU work units |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 171
|
|
| Author |
|
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges: 
|
Here's some another FWIW I've found.
----------------------------------------On my hosts tasks that turn invalid seem to have 2-3x the amount of CPU time than the tasks that validate. Not a % but actual time. Some tasks of the same batch that turn invalid on my Linux host with Nvidia will validate on my windows host with AMD and vice versa. 
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.
----------------------------------------  [Edit 1 times, last edit by nanoprobe at Apr 18, 2021 4:19:41 PM] |
||
|
|
Grumpy Swede
Master Cruncher Svíþjóð Joined: Apr 10, 2020 Post Count: 2165 Status: Offline Project Badges: 
|
Here's some another FWIW I've found. On my hosts tasks that error seem to have 2-3x the amount of CPU time than the tasks that validate. Not a % but actual time. Some tasks of the same batch that error on my Linux host with Nvidia will validate on my windows host with AMD and vice versa. With "error" do you mean that they turn into "error", or do they turn into "invalid"? That's a pretty big difference, since when it comes to these "invalid" we've seen recently, none of them have turned into "error" at least not for me, only to "invalid" |
||
|
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges:
|
Here's some another FWIW I've found. On my hosts tasks that error seem to have 2-3x the amount of CPU time than the tasks that validate. Not a % but actual time. Some tasks of the same batch that error on my Linux host with Nvidia will validate on my windows host with AMD and vice versa. With "error" do you mean that they turn into "error", or do they turn into "invalid"? That's a pretty big difference, since when it comes to these "invalid" we've seen recently, none of them have turned into "error" at least not for me, only to "invalid" Should have worded that better. Edited post.
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.
|
||
|
|
Dark Angel
Veteran Cruncher Australia Joined: Nov 11, 2005 Post Count: 721 Status: Offline Project Badges: 
|
I see there are still large numbers of invalid units being returned. I have about 4 pages of them. While they might be "valid" to the scientists they mess with the volunteers' ability to get further work and waste the electricity and machine life that WE are paying for. If whomever is building these dud units could get their collective excrement in order, that would be great.
---------------------------------------- Currently being moderated under false pretences |
||
|
|
ThreadRipper
Veteran Cruncher Sweden Joined: Apr 26, 2007 Post Count: 1321 Status: Offline Project Badges: 
|
I have two interesting WUs "close" to eachother
----------------------------------------OPNG_0003570_00295_0 DESKTOP-7D63M3N Valid 4/19/21 09:05:48 4/19/21 09:36:27 0.05 / 0.08 0.2 / 120.0 The first is valid, but the second one is strange because wingman has a valid result while mine is invalid. Both WUs crunched on the same GPU non-overclocked and should not return invalid results, and still, it did. This is the link to the WU status: https://www.worldcommunitygrid.org/ms/device/...s.do?workunitId=626928935 EDIT: Also, this one is very interesting as it is also invalid, but not only for me. It is invalid for one of the wingmen too: OPNG_0002322_00040_2 DESKTOP-7D63M3N Invalid 4/19/21 01:36:54 4/19/21 01:41:08 0.03 / 0.06 0.1 / 0.1 However, another wingman got it set to Valid: https://www.worldcommunitygrid.org/ms/device/...s.do?workunitId=618613562 Project Name: OpenPandemics - COVID-19 - GPU How did that guy get it set to Valid when there are three invalids? Shouldn't there be at least two results set to Valid for that to happen? How can you trust that the result was valid from 1 PC only, even if it is trusted, when there already are other invalids?  Join The International Team: https://www.worldcommunitygrid.org/team/viewTeamInfo.do?teamId=CK9RP1BKX1 AMD TR2990WX @ PBO, 64GB Quad 3200MHz 14-17-17-17-1T, RX6900XT @ Stock AMD 3800X @ PBO AMD 2700X @ 4GHz [Edit 1 times, last edit by ThreadRipper at Apr 19, 2021 10:17:37 AM] |
||
|
|
zdnko
Senior Cruncher Joined: Dec 1, 2005 Post Count: 225 Status: Offline Project Badges: 
|
6 wu 3967 invalid or server aborted
 https://www.worldcommunitygrid.org/ms/device/...s.do?workunitId=627332602 https://www.worldcommunitygrid.org/ms/device/...s.do?workunitId=627332600 https://www.worldcommunitygrid.org/ms/device/...s.do?workunitId=627332880 https://www.worldcommunitygrid.org/ms/device/...s.do?workunitId=627332606 https://www.worldcommunitygrid.org/ms/device/...s.do?workunitId=627331934 https://www.worldcommunitygrid.org/ms/device/...s.do?workunitId=627331940 |
||
|
|
True54Blue
Advanced Cruncher Joined: Nov 17, 2004 Post Count: 97 Status: Offline Project Badges: 
|
I've found one job where I was marked invalid but everyone else was valid which seems strange.
----------------------------------------https://www.worldcommunitygrid.org/ms/device/...s.do?workunitId=626034497 and some where 2/4 were invalid and 2/4 were valid. https://www.worldcommunitygrid.org/ms/device/...s.do?workunitId=624937906 https://www.worldcommunitygrid.org/ms/device/...s.do?workunitId=624658618 https://www.worldcommunitygrid.org/ms/device/...s.do?workunitId=624658623  |
||
|
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges:
|
I've found one job where I was marked invalid but everyone else was valid which seems strange. https://www.worldcommunitygrid.org/ms/device/...s.do?workunitId=626034497 and some where 2/4 were invalid and 2/4 were valid. https://www.worldcommunitygrid.org/ms/device/...s.do?workunitId=624937906 https://www.worldcommunitygrid.org/ms/device/...s.do?workunitId=624658618 https://www.worldcommunitygrid.org/ms/device/...s.do?workunitId=624658623 At first I thought it might be GPU related where the newest ones would have the task validated while anything older would not but you have the newest and it went invalid. There seems to be no rhyme or reason for the weirdness going on.
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.
|
||
|
|
sam6861
Advanced Cruncher Joined: Mar 31, 2020 Post Count: 107 Status: Offline Project Badges: 
|
Work units with only invalids.
----------------------------------------OPNG_0003352_00065_0 OPNG_0003352_00090_0 OPNG_0003352_00177_0 OPNG_0003352_00273_0 OPNG_0003466_00019_1 OPNG_0003466_00135_2 OPNG_0003466_00134_4 OPNG_0003472_00297_3 OPNG_0003490_00024_1 OPNG_0003503_00064_1 OPNG_0003729_00354_1 OPNG_0003729_00456_1 OPNG_0003730_00371_1 OPNG_0004166_00148_4 "Too late" with lots of invalids (T = "too late" on me) OPNG_0003352_00061_2 OPNG_0003466_00351_3 OPNG_0003475_00248_1 OPNG_0003727_00311_3 OPNG_0003729_00257_1 OPNG_0003729_00356_4 OPNG_0003729_00435_1 OPNG_0003755_00023_0 OPNG_0003755_00028_0 OPNG_0003755_00031_0 OPNG_0003978_00404_3 OPNG_0003979_00443_4T OPNG_0003979_00449_1 OPNG_0004091_00028_2T OPNG_0004091_00198_2 OPNG_0004123_00055_2 OPNG_0004123_00259_2 OPNG_0004126_00092_4T OPNG_0004137_00383_1 OPNG_0004347_00110_1 Work units with both invalid and valid, low granted credits. OPNG_0003352_00050_0 - 116.3 credits OPNG_0003352_00267_0 - 128.5 credits OPNG_0004123_00112_2 - 84.4 credits OPNG_0004123_00280_1 - 93.8 credits OPNG_0004123_00259_2 - 117.8 credits OPNG_0004091_00152_3 - 117.8 credits OPNG_0003503_00197_1 - 135.5 credits OPNG_0003727_00133_1 - 101.4 credits Lots of invalids happens on all 4 of my GPU with 3 PC. Ryzen 2700x, Win10, NVidia GT 1030 AMD RX 5500 XT, Ryzen 3900x, Win10, AMD RX 550 Intel i7-2600, Linux Debian, AMD RX 580, AMDGPU-Pro Over a past 2 days, looks like about 15% invalids, and 5% valid less then 120 granted credits. Hope to soon see a fix or update to low granted credits and lots of invalids. [Edit 1 times, last edit by sam6861 at Apr 20, 2021 2:23:18 AM] |
||
|
|
Rickjb
Veteran Cruncher Australia Joined: Sep 17, 2006 Post Count: 666 Status: Offline Project Badges: 
|
Fellow crunchers who are reading & posting here:
----------------------------------------*** Don't post links to work unit results *** The results pages are private to you, and other people trying to access them will be blocked. Post WU names instead, or copy a brief version of the relevant text that shows a sample of your issue. - HTH - ------------ PS: Different topic but on thread: I wouldn't be surprised if the techs pull OPNG off-line until the issue of the high proportion of WUs being duds is further investigated. [Edit 1 times, last edit by Rickjb at Apr 20, 2021 4:39:28 AM] |
||
|
|
|