Should I RMA this 7950

[jimbit]

I have just built a rig with 4 x 7950 (Gigabyte WF3).

3 out of the four are absolutely solid, but one gives persistent hw errors.

I've moved it to other slots to rule out the mb (the hw errors follow the card).

Tried reducing the intensity, but end up with  poor hashrates on the card (and still get hw errors)

Should I rma it?  And if so, what do I give as the reason for returning it?  Are hw errors on mining a valid reason for return?

Only other thing I can think of is taking all the other cards out and running some kind of stress test on it to try and make it throw an error (suggestions as to best program?)

thanks for reading

J

[1715096913]

1715096913

[1715096913]

1715096913

[1715096913]

1715096913

[1715096913]

1715096913

[chalupa_batman]

run it on its own, see if there are any errors. try running a game on it, if it has issues then you can rma

[cardcomm]

You don't say whee you bought it, but most retailers won't give you trouble replacing a defective card. I usually just say I'm getting screen artifacts.

[Kaepora]

I have just built a rig with 4 x 7950 (Gigabyte WF3).

3 out of the four are absolutely solid, but one gives persistent hw errors.

I've moved it to other slots to rule out the mb (the hw errors follow the card).

Tried reducing the intensity, but end up with  poor hashrates on the card (and still get hw errors)

Should I rma it?  And if so, what do I give as the reason for returning it?  Are hw errors on mining a valid reason for return?

Only other thing I can think of is taking all the other cards out and running some kind of stress test on it to try and make it throw an error (suggestions as to best program?)

thanks for reading

J

run OCCT http://www.ocbase.com/ and stress test the GPU make sure to check off scan for errors. In the progroam just go to GPU:3D set complexity to 7 and check off scan for errors. Let it run for a while maybe 10,20 or 30+ minutes to make sure.

[jimbit]

Great - will do, thanks for the info.

[jimbit]

I have just built a rig with 4 x 7950 (Gigabyte WF3).

3 out of the four are absolutely solid, but one gives persistent hw errors.

I've moved it to other slots to rule out the mb (the hw errors follow the card).

Tried reducing the intensity, but end up with  poor hashrates on the card (and still get hw errors)

Should I rma it?  And if so, what do I give as the reason for returning it?  Are hw errors on mining a valid reason for return?

Only other thing I can think of is taking all the other cards out and running some kind of stress test on it to try and make it throw an error (suggestions as to best program?)

thanks for reading

J

run OCCT http://www.ocbase.com/ and stress test the GPU make sure to check off scan for errors. In the progroam just go to GPU:3D set complexity to 7 and check off scan for errors. Let it run for a while maybe 10,20 or 30+ minutes to make sure.

Quick question - when you say "check off scan for errors" do you mean leave it checked (default) or take the check off (unchecked)?

[jimbit]

..also,  DX11 or DX9 test.

(I get what looks like a stationary image on both - should it be moving?)

[Kaepora]

..also,  DX11 or DX9 test.

(I get what looks like a stationary image on both - should it be moving?)

DX11 and leave scan for errors checked