What do you use for PSU (power, cable management, etc...)
Motherboard model ?
powered risers and/or mobo with added molex ?
Tried the card solo ?
Topic: Is my graphics card dying - advice please! (Read 191 times)
tried underclocking / having at stock. similar effect - slowly gets worse then crashes.
will try it in another slot
will try it in another slot
Swap the position of the card with that of another good card.
Does the bad performance move with the card or is it any card in a particular position that has issues?
As the other poster said, suspect the riser, they have a poor reputation.
Baz
Does the bad performance move with the card or is it any card in a particular position that has issues?
As the other poster said, suspect the riser, they have a poor reputation.
Baz
I have a box with 6x560's (yes, they're slower than the 580's).
All are undervolted / running cool at 12.5MH/s. They have been OK for ~2 months.
Over the last week, 1 card starts jumping up and down, and running at 11MH/s, and keeps slowing down over maybe 3-4 hours to ~10.5MHs, then crashes.
I've updated the BIOS back to stock (now it runs a lot hotter - ~80 degrees instead of ~60), thinking if I need to return it, it should have stock bios (still within warranty).
So,
- I think the card is dying, but it's not dead yet. It has not yet crashed since being returned to stock - but looking at the trend / spikey-ness, I suspect it needs a half day to drop down enough to crash.
- Is this RMA'able? At what point can I RMA? I do not want to keep running a card this hot. How would I even describe this fault?
All are undervolted / running cool at 12.5MH/s. They have been OK for ~2 months.
Over the last week, 1 card starts jumping up and down, and running at 11MH/s, and keeps slowing down over maybe 3-4 hours to ~10.5MHs, then crashes.
I've updated the BIOS back to stock (now it runs a lot hotter - ~80 degrees instead of ~60), thinking if I need to return it, it should have stock bios (still within warranty).
So,
- I think the card is dying, but it's not dead yet. It has not yet crashed since being returned to stock - but looking at the trend / spikey-ness, I suspect it needs a half day to drop down enough to crash.
- Is this RMA'able? At what point can I RMA? I do not want to keep running a card this hot. How would I even describe this fault?
You simply may be pushing it to hard.
If all cards do 12 except for one you could set clocks on that card lower then the other five.
Msi afterburner or smos will let you do that.
Also the riser could be failing
I have a box with 6x560's (yes, they're slower than the 580's).
All are undervolted / running cool at 12.5MH/s. They have been OK for ~2 months.
Over the last week, 1 card starts jumping up and down, and running at 11MH/s, and keeps slowing down over maybe 3-4 hours to ~10.5MHs, then crashes.
I've updated the BIOS back to stock (now it runs a lot hotter - ~80 degrees instead of ~60), thinking if I need to return it, it should have stock bios (still within warranty).
So,
- I think the card is dying, but it's not dead yet. It has not yet crashed since being returned to stock - but looking at the trend / spikey-ness, I suspect it needs a half day to drop down enough to crash.
- Is this RMA'able? At what point can I RMA? I do not want to keep running a card this hot. How would I even describe this fault?
All are undervolted / running cool at 12.5MH/s. They have been OK for ~2 months.
Over the last week, 1 card starts jumping up and down, and running at 11MH/s, and keeps slowing down over maybe 3-4 hours to ~10.5MHs, then crashes.
I've updated the BIOS back to stock (now it runs a lot hotter - ~80 degrees instead of ~60), thinking if I need to return it, it should have stock bios (still within warranty).
So,
- I think the card is dying, but it's not dead yet. It has not yet crashed since being returned to stock - but looking at the trend / spikey-ness, I suspect it needs a half day to drop down enough to crash.
- Is this RMA'able? At what point can I RMA? I do not want to keep running a card this hot. How would I even describe this fault?
Jump to: