I have 2 different NVIDIA cards "1080Ti and a 980. I've run this type of work before in Win 10 and Moo finished them.
Now it gets stuck and non of the GPUs are doing anything. I aborted it and have it try another task again.

Windows 11, BOINC 7.22.2, GPU driver 537.13, CUDA 12.2.0 Moo client 1.03 (cuda 31)

---

I see one valid task in your list, run few more and see how they work. However it was running very slow, it needed about 86 seconds per block, my (single) ancient GTX 275 needs 15.5 seconds per block if I let it use a full CPU core, with CUDA client core #10 (nearly no CPU usage at all) it still needs just 17.5 seconds/block. Do you leave one CPU core free for it?

Maybe it's necessary for you to force the OpenCL version via app_info.xml. I've posted a working one for Linux in this message, for Windows you'll just need the right files and their names and of course the right platform name.

---

Na, I just not going to run Moo projects. I'm just not editing any setting just to get something to work. Moo! Wrapper need to fix this. I just want them to know there's a problem.
I did a few more and it got stuck the same way.

Same here.

1 WU runs on 2 GPU.

Weird
That's a GTX980 and a GTX980 Ti
Detected as 2x GTX989Ti (???)

Results:

https://moowrap.net/results.php?hostid=1300227&offset=0&show_names=1&state=4&appid=

Not sure if it makes a difference for credits ?

these 2 also have 2 GPU but run 1WU / GPU.....

https://moowrap.net/show_host_detail.php?hostid=1317434
https://moowrap.net/show_host_detail.php?hostid=1326673

Not sure if it makes a difference for credits ?

No, you get simply 8 credits per crunched block.

---

https://moowrap.net/show_host_detail.php?hostid=1317434
https://moowrap.net/show_host_detail.php?hostid=1326673

Because they run the OpenCL and not the CUDA app, OpenCL uses only one GPU per task. You can create an app_info.xml file to run only the OpenCL app as suggested above, if you post the file names for all files that belong to that application (like in the Linux thread), I can help you with that (my GTX 275 runs only the CUDA app, so I don't have the files for the OpenCL app).

---

thx, the machine did get some opencl tasks and yes they run 1WU/GPU.

I'll give it a go

I used to have the same issue on a spesific machine, it has a GTX 760 and a GT 730, runs fine all the way up to about 93% where according to msi afterburner, the gtx 760 stops computing and the 730 is left computing alone for a while, this has since stopped an
d it now sends single tasks for each card.

I used to have the same issue on a spesific machine, it has a GTX 760 and a GT 730, runs fine all the way up to about 93% where according to msi afterburner, the gtx 760 stops computing and the 730 is left computing alone for a while, this has since stopped an
d it now sends single tasks for each card.

Thank for the update. I'll try it again.

No. Projects still get stuck 98% now.
Remaining time is over an 1hr. Was only 4:43 when started.
21hrs go by then I aborted.
Running 2 GPUs for 1 Project.
I have a GTX 1080 and 980.

If project can't run on 2 different GPUs then exclude projects when 2 GPUs don't match. But still gather work but make it run 1 GPU at a time when GPUs don't match.
Or find a way to have 1 project work together on 2 different GPUs.

I will not change any settings for work arounds because we shouldn't have to mess with settings when it should be a set it and forget it.
And if problem arises then report.

Ok, after creating this message I let 1 more Moo project run 1 PJ on 2 GPUs and finish at 4:43.
I still have plenty of projects to run but I'm going to about if 1 of them goes over 6 mins.
Seem like an on and off issue.