In cc_config.xml You should add:

<exclude_gpu> <url>www.primegrid.com</url> <device_num>1</device_num> </exclude_gpu>

That should work, just remember to restart boinc.

(If 1050 is device 1, restart boinc and look in event log which number it is)

9/12/2021 11:31:26 AM | PrimeGrid | Config: excluded GPU. Type: all. App: genefer. Device: 1


It sees the command but ignores it.
I rebooted the computer, still ignores it.
I included the app name genefer
But nothing works.
I tried to "repair" BOINC, that doesn't help.
I even added in the 'type' line and that doesn't change anything either.

You don't use type and device num at the same time in the wording or it could confused also in your since both are Nvidia both would be excluded. Also you MUST use the names of the projects from within Boinc Manager not from the websites themselves, I pulled my hair out until I figured that out. What I did was pretend to re-add the Project and then copy and pasted that info into the cc_config file. No idea why the exact same info didn't work for me when using the website but it didn't.

Here's one I used awhile back:
<cc_config>

<options>
<use_all_gpus>1</use_all_gpus>

<exclude_gpu>
<url>http://einstein.phys.uwm.edu/</url>
<device_num>0</device_num>
</exclude_gpu>

<exclude_gpu>
<url>http://milkyway.cs.rpi.edu/milkyway/</url>
<device_num>1</device_num>
</exclude_gpu>

</options>

</cc_config>

In cc_config.xml You should add:

<exclude_gpu> <url>www.primegrid.com</url> <device_num>1</device_num> </exclude_gpu>

That should work, just remember to restart boinc.

(If 1050 is device 1, restart boinc and look in event log which number it is)

How can Woodall(LLR)(mt) and 9.01 PPS (LLR) (mt) run at the same time? (10 cores each)
I have 16 cores. 10 are used by PG and the others are spread over 4 tasks from another project and 2 controlling my GPU (combined .36 of a core)

This is physically impossible unless both projects are sharing the same 10 cores?

I've seen that on my own pc's before and eventually one of the tasks should say 'waiting to run' as like you say there is not enough resources to run it.

Actually, to be the bearer of bad news, if you're talking about the Ryzen 2700 on your account, you only have 8 physical cores. The other 8 threads are logical. There are many threads on this forum about this discrepancy and how to get the most out of such a system. Search for "hyperthreading".

Regardless of that, to address the matter at hand, I don't think it is actually doing both at the same time, but rather switching between the two units. This happens a lot when you have mixed long (e.g. Woodall) and short (e.g. PPS) tasks. When you get near the end of a task, Boinc downloads a new one, but when they are vastly different sizes, it starts to switch back and forth because of a) different deadline setups and b) weird logic, until the first finally finishes.

Or, I could be off base. Can you take a screenshot of your Boinc window showing both tasks?

It appeared to be about 95% done when it was aborted.

Now that's weird!
Everything I read showed it was not even really started.
Where did you read this information at so I can read it next time I get a weird reading?

Well, BOINC should have been showing you the correct completion time. I'm not sure why it wasn't. BOINC is notorious for this, however.

I looked in the stderr.txt output. While the task is running, this is in a file in the slot directory where the task is running. Once the task is over, it's included in the task page: https://www.primegrid.com/result.php?resultid=1044031858

Near the top of the stderr output, it says that the expected run time is 117 hours. This is usually pretty accurate.

While it doesn't tell you directly how far along it is, you can make a decent guess by reading between the lines.

Each time the task was stopped and restarted, which happened many times, it tells you how many iterations are remaining. The first time it restarted, there were 66 million iterations to go. The final time it restarted, there were 4 million iterations to go. So it had to be at least about 95% done.

I was doing driver updates the other night around the time this task started having problems. The studio driver updated ok, but I was having issues with game driver installing.
Would this cause problems?

It's hard to say if this did cause problems, but it certainly can cause problems.

I recommend never upgrading drivers -- or anything -- while there are tasks in progress.

I have a 1050ti and a DYFL task takes just over 7 days and a GFN22 task takes just over 5 days.
That's likely an inaccurate measurement.
Like Kellen said, use a 3rd party program to check the GPU usage. Also check your temperatures to make sure things aren't throttling down to prevent melting.

https://www.primegrid.com/result.php?resultid=1044031858

This is the most absurd task I have gotten.
Even though it looks like in the output file that it was using the maximum capability of my card, I watched it in BOINC stats and it was barely moving .0002 every 6 seconds.
In all my time on BOINC I have never seen a 100+ day completion time.

I looked in the PG stats for GPU and this is a #27 card, but still...114 days? Crazy!

Is this a DYFL task? (we can't see the task in the link until it is finished running). Even if so, that is likely an incorrect estimate or there is something wrong with your computer. I ran some GFN-22 tasks, which are almost as long as DYFL, on a GT 1030 and they took 11 days, so I would expect DYFL on a GTX 1050 ti to take somewhere around 6 or 7 days.

If you can use HarwareInfo64 or some similar program to check the utilization of your GPU it would be good. (Task Manager in Windows will not report GPU utilization accurately, btw, so additional software is required). If GPU usage is close to 100% then it is likely just an overestimate by BOINC.