WUs are now back since last night.
What I notice is that their runtime is even shorter than before, credit earned per WU is mostly markedly below 500 :-(

Too many Invalid OPNG's all over the place now. Both with my GTX980, and my iGPU HD4600. It's not just me, but all wingmen getting invalids, with these. Something is wrong with too many WU's
https://www.worldcommunitygrid.org/ms/device/...s.do?workunitId=813561025

Project Name:		OpenPandemics - COVID-19 - GPU
Created:		09/16/2021 06:48:55
Name:		OPNG_0085779_00042
Minimum Quorum:		1
Replication:		6
	

Result Name	OS type	OS version	App Version Number	Status	Sent Time	Time Due /
 Return Time	CPU Time / Elapsed Time (hours)	Claimed/ Granted BOINC Credit
OPNG_ 0085779_ 00042_ 3--	Microsoft Windows 10	Professional x64 Edition, (10.00.19043.00)	728	Invalid	9/16/21 07:02:35	9/16/21 07:06:51	0.05	0.6 / 0.0
OPNG_ 0085779_ 00042_ 4--	Microsoft Windows 8.1	Core x64 Edition, (06.03.9600.00)	728	Invalid	9/16/21 07:02:35	9/16/21 07:07:34	0.07	0.6 / 0.0
OPNG_ 0085779_ 00042_ 1--	Microsoft Windows 10	Professional x64 Edition, (10.00.19043.00)	728	Invalid	9/16/21 06:57:01	9/16/21 07:02:51	0.09	0.6 / 0.0
OPNG_ 0085779_ 00042_ 2--	Microsoft Windows 10	Core x64 Edition, (10.00.19043.00)	728	Invalid	9/16/21 06:57:01	9/16/21 07:02:22	0.06	0.6 / 0.0
OPNG_ 0085779_ 00042_ 0--	Microsoft Windows 10	Professional x64 Edition, (10.00.18362.00)	728	Invalid	9/16/21 06:50:15	9/16/21 06:56:54	0.05	0.6 / 0.0

In the past half hour, I experienced several "server aborted" WUs :-(
That's annoying, of course, since it is that difficult anyway to obtain WUs.

This pattern of work units going Invalid appears to be a repeat of what happened to quite a lot of tasks between 2nd and 6th of June this year, where quite a lot of work units for the same receptor currently being targeted came up as all Invalid or one/two Valid, the rest Invalid...

Presumably, certain ligands push the software outside its comfort zone and the output contains unexpected items or values - however, without being able to see the full output one can only speculate...

I think they're throwing just about every ligand they can conjure up at this target, and that's probably why we have the occasional problem. However, as a proportion of the work units investigating this receptor, the number of Invalid tasks would appear to be quite small - it's just unfortunate that they seem to crop up in clusters. In my case I had 14 Invalids in June and 10 so far this time; that's out of over 10,000 work units I have processed for this one receptor (which is just over half the total number of work units I've processed altogether.)

On the basis that even an Invalid result is valid science (if the reason for it can be determined...), I'd hope that the scientists get to see which jobs went Invalid. One can debate whether credit should be awarded; one probably ought to get the time credited, but if the results look invalid the WCG points calculator probably can't score the task(s)!

By the way, there's a topic about Invalid GPU Work Units; so far, no-one has chipped in there with an explanation either (and given the current situation, I suspect the move takes priority over pretty much everything else, so I guess we just have to wait for it to get past the ligands that are presumably causing the problem...)

Cheers - Al.

[Edited: very small -> quite small.]

Hi, erich!
It seems that you got the fourh copy of the WU to process, all the other four failed and was returned 1-10min before yours