Does SETI@Home no longer support a Nvidia Quaddro FX 4800 GPU? If any additional info would help, just ask. Thanks.

Should be OK. I looked at your computer and SETI sees the card and the 342.0 driver. As I don't see any OpenCL listed, I would imagine that Cuda 32/42/50 would be the best you'll be able to do, but you should be able to do something as long as that PC is set at SETI to get GPU work, not just CPU tasks. I can't remember for sure, but I don't think you'll be doing OpenCL work on a card that old. I would suggest you go to the NVidia driver download site and see if there's something newer for that card. Seems to me I recall SETI having issues with 342, but I could be wrong there.

I do see a bunch of Cuda32/42/50 tasks that have been in work since mid-April. Do you see them in BOINC Manager as well?

---

The Quadro FX 4800 is a glorified GTX260 and also pre-Fermi so any drivers from 340.xx onwards only gives legacy support only for them, but they still do well in a pre-Win10 environment with pre-340.xx drivers. Read thread @Pre-FERMI nVidia GPU users: Important warning.

Using drivers supplied by Nvidia will show the OpenCL component while the driver supplied by Win10 will not, but again refer back to linked thread.

With your setup you should be able to do Cuda32 work, but if I was you I'd suggest replacing that card with a more modern 1 that will give you much better performance while using less power.

Cheers.

Thanks for quick reply! For the moment the problem is solved. I did a clean install of the nVidia driver (342.00) and it found the GPU w/ OpenCL support. (Thanks Bob!) I had asked a question in the support forum about an update notice that I was getting. Bob suggested the clean install of the nVidia supplied drivers, it's seems that windows update had installed it's own version of 342.00 that didn't have the OpenCL support.

I hadn't noticed those tasks. Now that the GPU processing is working, tasks will process more quickly. Do you have a suggestion to free those tasks, or should just leave things as they are? Any suggestions are welcome, especially if there is a better way doing things.

Thanks,
Jack

---

https://boincstats.com/signature/-1/user/40040505/sig.png

Thanks for quick reply! For the moment the problem is solved. I did a clean install of the nVidia driver (342.00) and it found the GPU w/ OpenCL support. (Thanks Bob!) I had asked a question in the support forum about an update notice that I was getting. Bob suggested the clean install of the nVidia supplied drivers, it's seems that windows update had installed it's own version of 342.00 that didn't have the OpenCL support.

I hadn't noticed those tasks. Now that the GPU processing is working, tasks will process more quickly. Do you have a suggestion to free those tasks, or should just leave things as they are? Any suggestions are welcome, especially if there is a better way doing things.

Thanks,
Jack

. . If M$/Windows ever updates your video drivers you will lose OpenCL support, best to avoid that is you can ...

. . It's a realy pain.

. . If you want to recover those old tasks have look in ''Panic Mode On (112) Server Problems", it is being discussed quite thoroughly there.

Stephen

:(

Is there a major difference in how many cpu's SETI uses between 8.20 and 8.22? It seems to me.

---

Is there a major difference in how many cpu's SETI uses between 8.20 and 8.22? It seems to me.

. . Having just changed from 8.20 to 8.22 I can say definitely NO!

. . Sorry about the hiccough!

Stephen

. .

Is there a major difference in how many cpu's SETI uses between 8.20 and 8.22? It seems to me.

. . Having just changed from 8.20 to 8.22 I can say definitely NO!

. . But if you had some tweaks when using 8.20, such as -use_sleep and you lost that config when changing to 8.22 then it is running with sleep OFF and yes it will then use more CPU time. But that is due to the sleep function not the app version.

Stephen

. .

Thank you Stephen. How do I make the change in the 8.22 so it also uses about 0.04 cpu for each gpu doing SETI?

---

Thank you Stephen. How do I make the change in the 8.22 so it also uses about 0.04 cpu for each gpu doing SETI?

You can't. The application will use as much cpu as it wants. The only way to limit the cpu resource is with the application tuning parameters like the -use_sleep that Stephen mentioned.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

All available options and their meaning listed in ReadMe_MultiBeam_OpenCL.txt file in project directory.

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

@ Raistmer

. . Since this morning I am having problems with SoG r3557. It is running OK, but taking a long time, but the problem is I am getting lots of these messages in every task.

<core_client_version>7.6.33</core_client_version>
<![CDATA[
<stderr_txt>
occurs; FFTLength=8192
WARNING PulseFind w/o TripletFind occurs; FFTLength=8192
WARNING PulseFind w/o TripletFind occurs; FFTLength=8192
WARNING PulseFind w/o TripletFind occurs; FFTLength=8192
WARNING PulseFind w/o TripletFind occurs; FFTLength=8192
WARNING PulseFind w/o TripletFind occurs; FFTLength=8192
WARNING PulseFind w/o TripletFind occurs; FFTLength=8192
WARNING PulseFind w/o TripletFind occurs; FFTLength=8192
WARNING PulseFind w/o TripletFind occurs; FFTLength=8192

. . After that the rest of stderr.txt seems normal.

http://setiathome.berkeley.edu/result.php?resultid=6817235206

. . I have closed BOINC manager/client and restarted but the problem remains.

. . I am about to reboot the machine but I was wondering if you might have any idea why this is happening?

Stephen

[edit] . . The system reboot may have been successful. When it restarted there was a disk error detected/repaired on my SETI drive. The tasks currently running are not showing that message. It seems there was a glitch on that drive.

? ?

This Looks like an GPU issue to me.

Dust bunnies ????

---

With each crime and every kindness we birth our future.

This Looks like an GPU issue to me.

Dust bunnies ????

. . More a drive error it seems ...

Stephen

. . Wrong again,

. . After the reboot and before I went out it seemed to be OK, but on return it turns out it was still doing the same thing. On all the Arecibo tasks. I have now suspended the remaining Arecibo tasks and it is running the GBT tasks which so far seem to not have the problem.
? ?

. . I will keep you posted.

Stephen

:(

. . More a drive disk error it seems ...
Stephen
. . Wrong again,
. . After the reboot and before I went out it seemed to be OK, but on return it turns out it was still doing the same thing. On all the Arecibo tasks. I have now suspended the remaining Arecibo tasks and it is running the GBT tasks which so far seem to not have the problem.
? ?
. . I will keep you posted.
Stephen

. . OK, not seeing the problem on the GBT tasks at all but still on the Arecibo tasks ....

. . I am wondering if I should abort them in case I am generating bogus results, though they are validating ... ???

Stephen

?

. . I moved the offending Arecibo tasks to the CPU, they are not producing any errors there so I am concluding that the files are not corrupted. I moved a few back to the GPU and confirmed they are still getting those errors when run under r3557.

. . Hopefully they are not causing bad results to get into the science database :(

Stephen

:(

How about looking in this short thread from last year. It happened back then too with some 03ja07aa tasks.
https://setiathome.berkeley.edu/forum_thread.php?id=82399

. . thanks Tut,

. . I refilled an almost empty cache and got over 100 of these tasks on my Windows box. I haven't noticed any on my Linux boxes ...

Stephen

. . Good to know it isn't just me ...

:)

. . thanks Tut,
. . I refilled an almost empty cache and got over 100 of these tasks on my Windows box. I haven't noticed any on my Linux boxes ...
Stephen
. . Good to know it isn't just me ...
:)

Well, it can for some reason happen with SoG, extremely rare though. I don't think we ever got an explanation from Raistmer for it though, back in December 2017.
They do validate OK though.

. . Yep, they are all validating OK now as well. But I have left the rest of them on the CPU as they do not do it there.

Stephen

<shrug>

Has anyone been playing with this setting -pref_wg_num_per_cu ?

Has anyone been playing with this setting -pref_wg_num_per_cu ?

Of course as the alpha tester of this app i did.
And what is the question ?

---

With each crime and every kindness we birth our future.

What are your settings with the -pref_wg_num_per_cu?

Im getting quite an improvement on my 1080ti increasing this setting, especially when running for example 3 of 4 units at once in combination with cpu_lock.

I did not find much information about it, so i thought id ask.