| 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 |
|
|
Falconet
Master Cruncher Portugal Joined: Mar 9, 2009 Post Count: 3265 Status: Offline Project Badges: 
|
Running a few OPNG on Google Colab with an Nvidia Tesla P100-PCIE-16GB and no invalids so far. Batch 4896.
----------------------------------------Lets hope these invalids are indeed past us. Thanks uplinger. AMD Ryzen 5 1600AF 4C/8T 3.2 GHz - 85W AMD Ryzen 5 2500U 4C/8T 2.0 GHz - 28W Intel Z3740 4C/4T 1.8 GHz - 6W [Edit 1 times, last edit by Falconet at Apr 20, 2021 7:08:57 PM] |
||
|
|
Grumpy Swede
Master Cruncher Svíþjóð Joined: Apr 10, 2020 Post Count: 1884 Status: Offline Project Badges: 
|
Good batches flowing into my 2 stone age computers at a pretty good rate now. 90 WU's for the last 12 hours.
----------------------------------------Only one invalid, and that was a resend from a known faulty batch.   |
||
|
|
spRocket
Senior Cruncher Joined: Mar 25, 2020 Post Count: 234 Status: Offline Project Badges: 
|
Just chewed through a block of nine (#4467), and all came back valid. I just wish there were more than nine WUs in that block.
 I'll happily take what I can get, though... |
||
|
|
adriverhoef
Master Cruncher The Netherlands Joined: Apr 3, 2009 Post Count: 1984 Status: Offline Project Badges: 
|
First Invalid above batch# 5000 ...
----------------------------------------workunit 632723626generated by: wcgstats -wQt2 -aOPN1 -m2 -sI -HI -l1Seeing nothing strange happening inside the task ... OPNG_0005553_00232_0-- Linux Fedora 728 Invalid 4/23/21 02:58:05 4/23/21 07:12:47 0.19 1.1 / 1.1(a part of what has been) generated by: wcgstats -wQt2 -aOPN1 -m2 -sI -HI -l1 -w[Edit 2 times, last edit by adriverhoef at Apr 23, 2021 11:38:32 AM] |
||
|
|
Richard Haselgrove
Senior Cruncher United Kingdom Joined: Feb 19, 2021 Post Count: 360 Status: Offline Project Badges: 
|
First Invalid above batch# 5000 ... The four computers which completed the task were:workunit 632723626(successful) Haswell GT2 Mobile HD Graphics P630 Haswell Ultrabook GT3 Mobile (unsuccessful - yours) UHD Graphics 630 - all slight variations on the intel_gpu theme. While working at SETI@Home, we encountered many problems getting the intel_gpus to return results with a sufficiently close mathematical output to validate within the degree of accuracy required. After a weekend of exhaustive testing on my HD Graphics 530 (this was some years ago...), we concluded that a compiler optimisation called '-cl-mad-enable' was to blame - this is noted in the developer notes as producing 'faster but less precise' applications. As the years have passed, and as drivers have improved, I've noticed that the 'fast' (optimised) applications have diverged further and further from the slower 'precise' applications. -cl-mad-enable is in use in the OPNG applications here. I've already let Uplinger know of my concerns. |
||
|
|
gio777
Advanced Cruncher Georgia Joined: Dec 8, 2004 Post Count: 69 Status: Offline Project Badges:
|
I don't get any GPU units at all. not available... :(
---------------------------------------- |
||
|
|
adriverhoef
Master Cruncher The Netherlands Joined: Apr 3, 2009 Post Count: 1984 Status: Offline Project Badges:
|
First Invalid above batch# 5000 ... The four computers which completed the task were:workunit 632723626(successful) Haswell GT2 Mobile HD Graphics P630 Haswell Ultrabook GT3 Mobile (unsuccessful - yours) UHD Graphics 630 - all slight variations on the intel_gpu theme. While working at SETI@Home, we encountered many problems getting the intel_gpus to return results with a sufficiently close mathematical output to validate within the degree of accuracy required. After a weekend of exhaustive testing on my HD Graphics 530 (this was some years ago...), we concluded that a compiler optimisation called '-cl-mad-enable' was to blame - this is noted in the developer notes as producing 'faster but less precise' applications. As the years have passed, and as drivers have improved, I've noticed that the 'fast' (optimised) applications have diverged further and further from the slower 'precise' applications. -cl-mad-enable is in use in the OPNG applications here. I've already let Uplinger know of my concerns. Thanks Richard, I'm writing this beacuse I wanted to let you know that I have also had 4 Invalids in the early GPU stages after the BETA was well over. (I didn't participate in the BETA phase because I was unaware that my device was able to run GPU tasks at that time.) $ wcglog -0xe OPNG | grep ' 0.0 161' | head -5(The task in the middle, the third one in the listing of 5 tasks above, was Server Aborted.) |
||
|
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges: 
|
Looks to me as though the invalid issue has returned to what it was prior to 4/17. Breakdown for me:
----------------------------------------4/18 2 4/19 1 4/20 0 4/21 1 4/22 30 4/23 49 so far and counting.
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.
  |
||
|
|
JohnDK
Advanced Cruncher Denmark Joined: Feb 17, 2010 Post Count: 77 Status: Offline Project Badges: 
|
Have to say it's getting frustrating and annoying with all these invalids...
----------------------------------------
Intel i7-6850K / 16GB / RTX 3090 / 2x RTX 3080 Ti / RTX 3070 Ti
AMD Ryzen 9 5950X / 32GB / RTX 2080 Ti |
||
|
|
Caranthir
Cruncher Joined: May 7, 2016 Post Count: 8 Status: Offline Project Badges:
|
I also got two invalids and one too late:
For the invalid one the log looks : Result Name: OPNG_ 0004324_ 00444_ 2-- And for the one that is too late:
I don't really understand what the problem might be. The tasks took about 2 minutes to complete. |
||
|
|
|