Should i return this GPU

M

motqalden

[H]ard|DCOTM x5
Joined
Jun 22, 2009
Messages
3,167
So i bought my EVGA 1070 (ICS cooler edition) card about 2 months back.
It seems to happily run games for hours at 2100MHZ.
Folded for about a week boosting to 2050+ with no errors. Around the time that i started doing WCG i started getting this type of error.

07:43:36:WARNING:WU00:FS00:FahCore returned an unknown error code which probably indicates that it crashed
07:43:36:WARNING:WU00:FS00:FahCore returned: UNKNOWN_ENUM (127 = 0x7f)

This typically doesn't fail the work unit, but if it happens enough on the same work unit i have seen if return no credit.
I have been gradually decreasing my overclock and the errors seems to go away or show up less frequently but after a while they would return and the interval would get shorter.

As of now i am actually underclocking the GPU -50 on the core (running no memory OC) and i am still getting these errors!

I read a thread somewhere that this might be caused by virtual box so i uninstalled that but it has not helped.
Also should mention that card typically doesn't get over 65c and never over 70
 
are you overclocking the vRAM?. what driver are you using? what OS?..
 
Try running an older driver. Or ditch F@H and run GPU BOINC projects.
 
Also, are you leaving a full core available to "feed" the GPU? I'm not as experienced with FAH work units so not sure if it is failing due to over commitment of the resources. Someone that runs FAH GPU with BOINC CPU work at the same time may want to chime in on that one.

You may also want to tun all BOINC as Skillz suggested just to see if it is perhaps FAH not playing nice with BOINC. Also, you may want to test with just running straight FAH again to make sure it is related to running BOINC alongside it or if the card finally just gave out.
 
Gilthanis said:
Also, are you leaving a full core available to "feed" the GPU? I'm not as experienced with FAH work units so not sure if it is failing due to over commitment of the resources. Someone that runs FAH GPU with BOINC CPU work at the same time may want to chime in on that one.

You may also want to tun all BOINC as Skillz suggested just to see if it is perhaps FAH not playing nice with BOINC. Also, you may want to test with just running straight FAH again to make sure it is related to running BOINC alongside it or if the card finally just gave out.
Click to expand...


Thanks for the suggestions. I hate to lose production but i will try foldind with wcg off to test that. I have tried older drivers and that did not help. Not really keen on dropping fah. I was dedicating 2 cores to fah because i had heard 1 is not always enough
 
I have looked at wcg logs and i see suspicious entries related to gpu usage despite disabling it fully in settings
 
WCG doesn't use GPUs on any of it's projects. What are these entries you are referring to?
 
Yeah...you should copy and paste your event log showing the entries...
 
If anything they just acknowledge the fact that you have a GPU. If you have "allow GPU work" from them then it might tell you something along the lines of "no projects with GPU work found" or something to that effect.
 
Skillz said:
If anything they just acknowledge the fact that you have a GPU. If you have "allow GPU work" from them then it might tell you something along the lines of "no projects with GPU work found" or something to that effect.
Click to expand...

mine say things as "resuming GPU computation" and "Suspending GPU computation - computer it's in use" However it doesn't ever do anything with the GPU, GPU activity remains 0% - 1% while idling.
 
I think that's from the BOINC manager. If you have your settings set to suspend GPU when the computer is in use then it's probably just logging when the computer begins idling and when it comes out of idling. Just a guess as I don't see those in my logs, but I don't have that setting enabled on my setup(s) either.
 
Skillz said:
I think that's from the BOINC manager. If you have your settings set to suspend GPU when the computer is in use then it's probably just logging when the computer begins idling and when it comes out of idling. Just a guess as I don't see those in my logs, but I don't have that setting enabled on my setup(s) either.
Click to expand...

this is an example it does happens several times a day but it actually does nothing, you are right, it just sense the machine being idle (I just set an idle time of like 3minutes IIRC).
example1.PNG
 
I wouldn't even enable that option if you aren't using the GPU for BOINC. Would remove that from your log. As long as none of the projects you are running have GPU work then it wouldn't ever run on the GPU. If those projects do have GPU then it would be best to just disable GPU from each project that has them individually.

Either way it doesn't matter. As you stated nothing is running on the GPUs when that happens anyway.
 
Thanks Araxie, That was the log entry i was talking about but i could not find a recent one for some reason... Anyways i think this thread can be closed or whatever. I will either figure it out or send it in for RMA. I just figured i would post something here in case someone else has had similar experiences with the error. I had been googling it and reading a ton of forums but didn't come up with anything concrete which is why i made this post. I am certainly going to take my time before sending anything back since the amount of points i would lose waiting for replacement would be way way more than the occasional failure i am getting now :D
 
If you were getting that error, then I would double check to make sure BOINC did not have a project with GPU work turned on. Because if it did, it may be trying to load it while FAH was running GPU work at the same time which could cause an error. In BOINC manager with the advanced view, click on Activity and then click on Suspend under the GPU section. This will prevent BOINC from running GPU work regardless of project.
 
Yeah i had it disabled like you are suggesting in that section, but i also had the checkmark for suspend gpu when cpu not idle. I have removed the idle checkmark
 
See of they offer advanced RMA. They'll send you a replacement card before you send them the defective card.
 
Hated to do it but i stopped my WCG for now and the folding errors went away... I even put my OC back on the GPU and no issues. guess my cheap AMD with budget MSI can't handle being fully stressed out with both GPU and CPU load? seems silly since it would run all out for 10+ days with no restart and not once would i get any hangs or errors.... Oh well I have a new Ryzen 7 / x370 on the way which should fix this issue. don't think the AMD system will have any issues running WCG only once its all said and done. I plan on keeping both systems up and running but i am going to undervolt / underclock the 8320 a bit to save on power.

Glad i didn't return the GPU :D
 
Not sure if i do or not but its gonna be a moot point come tomorrow. I have my new ryzen almost built in prep for its ram arriving tomorrow and i plan to harvest my 1070 and put it with the ryzen. I got a cheap nvidia gt210 to throw in the 8320 rig and i plan to keep it running WCG only so that basically solves my problem (assuming i don't have errors on the new setup) Funny story. i installed the MB and my cooler and then closed up my new case with nice tempered glass windows. I have pet rats and they were running around and then i notice that one of them was inside my new case! for the life of me i couldn't figure out how he had got in there but then i remembered i hadn't put the power supply in yet, so he had crawled in the whole in the back and climbed over the PSU shroud :D really funny to see him through the glass panel. little douche then started reaching for a fan cable and i quickly unscrewed the panel to let him out....
 
Are you trying to run FAH *AND* WCG on the same GPU?

VERY VERY BAD IDEA if so - I'm amazed that it would even try to work.

If you're trying to run WCG on your CPU, set it up so it doesn't touch at least one FULL core of the CPU per 1080 GPU - folding NEEDS that core to keep the GPU fed.
 
QuintLeo said:
Are you trying to run FAH *AND* WCG on the same GPU?

VERY VERY BAD IDEA if so - I'm amazed that it would even try to work.


Yeah i am not running CPU for FAH and not running GPU for WCG. Anyways my errors all went away as soon as i moved the GPU to the new system. 0 issues and definitely a bit faster moving to PCIE 3.0

All issues solved!
Click to expand...
 
You must log in or register to reply here.
Back
Top