Hello everyone.
My PC is as follows:
— MSI P55-GD65
— Intel Core i5-750 (about 5% over the factory clock)
— 8 GB DDR3 1333 (about 5% over the factory clock)
— Gigabyte GeForce GTX560 Ti (factory overclock + about 7-8% over that)
— Crucial m4 SSD + some HDDs
— Infrasonic Quartet (semi-professional sound card)
— Corsair HX520 PSU
— Windows 7 Pro x64
— Ascot 6AR PC case (slightly above average, nothing fancy though)
Pretty much all the hardware is 3-3.5 years old (including the 560 Ti). I'm using it for everything, including work and games, and generally it performs very well at the relevant tasks. I also tend to never turn it off as it doubles as a personal web-server/seedbox.
So what happened is that this afternoon the PC suddenly rebooted without any prior sign of failure. When it booted back up, the display started showing shimmering horizontal lines (starting as early as the initial Windows logo screen during the boot-up). Additionally, when it booted the OS the GPU driver was disabled with the error code 43 (general failure), so it ended up working in a legacy mode in 800×600. It didn't, however, reboot or otherwise show instabilities at any point after the initial accident.
First thing I thought is that this sort of image corruption looked exactly as if you had overclocked the VRAM too much, so maybe what happened was that the VRAM modules overheated for some reason (the card wasn't under heavy load prior to the reboot, and it wasn't very hot inside the case), so I thought letting it cool off would do. I turned the PC off and went away for several hours, but upon my return the problem persisted. The ambient temperature in my room is around 20°C, so it's quite cool actually.
Then I thought it could be a software problem, and tried uninstalling and purging my old Forceware driver and DirectX. This didn't help either. I found my previous graphics card which luckily had the same power setup (with two extra internal power cords), and it worked just fine. This made it obvious that it was the 560 Ti that failed.
Now that as much was established, I thought that maybe the (rather modest) overclock has, for one reason or another, pushed the VRAM into a malfunctioning state, and lowering its clock and/or tweaking voltage settings would be able to help. Unfortunately, every over/underclock-related tool I know accesses the GPU via the driver, which, as I've said before, doesn't even load, so I can't access any of the GPU's settings. I figured that maybe reflashing BIOS could at least drop the clocks down to factory defaults, but it didn't help at all either.
So the main question is, is there any way to salvage this card, even if it ends up underclocked? Also, are there any over/underclock tools that access the hardware at a lower level, circumventing the driver? What gives me hope is that if it doesn't show garbage during the BIOS sequence, then maybe not everything is lost. However, that might be because the textmode output doesn't go through the VRAM at all (unfortunately, I'm not too familiar with that level of technical detail).
I would be grateful for every constructive advice and/or expert opinion on the matter. I do eventually plan to upgrade my hardware, but I also had plans for this particular card, so if it's at all possible to save it, that's what I'm looking for in any case.
Any option to salvage a presumably damaged graphics card?
My PC is as follows:
— MSI P55-GD65
— Intel Core i5-750 (about 5% over the factory clock)
— 8 GB DDR3 1333 (about 5% over the factory clock)
— Gigabyte GeForce GTX560 Ti (factory overclock + about 7-8% over that)
— Crucial m4 SSD + some HDDs
— Infrasonic Quartet (semi-professional sound card)
— Corsair HX520 PSU
— Windows 7 Pro x64
— Ascot 6AR PC case (slightly above average, nothing fancy though)
Pretty much all the hardware is 3-3.5 years old (including the 560 Ti). I'm using it for everything, including work and games, and generally it performs very well at the relevant tasks. I also tend to never turn it off as it doubles as a personal web-server/seedbox.
So what happened is that this afternoon the PC suddenly rebooted without any prior sign of failure. When it booted back up, the display started showing shimmering horizontal lines (starting as early as the initial Windows logo screen during the boot-up). Additionally, when it booted the OS the GPU driver was disabled with the error code 43 (general failure), so it ended up working in a legacy mode in 800×600. It didn't, however, reboot or otherwise show instabilities at any point after the initial accident.
First thing I thought is that this sort of image corruption looked exactly as if you had overclocked the VRAM too much, so maybe what happened was that the VRAM modules overheated for some reason (the card wasn't under heavy load prior to the reboot, and it wasn't very hot inside the case), so I thought letting it cool off would do. I turned the PC off and went away for several hours, but upon my return the problem persisted. The ambient temperature in my room is around 20°C, so it's quite cool actually.
Then I thought it could be a software problem, and tried uninstalling and purging my old Forceware driver and DirectX. This didn't help either. I found my previous graphics card which luckily had the same power setup (with two extra internal power cords), and it worked just fine. This made it obvious that it was the 560 Ti that failed.
Now that as much was established, I thought that maybe the (rather modest) overclock has, for one reason or another, pushed the VRAM into a malfunctioning state, and lowering its clock and/or tweaking voltage settings would be able to help. Unfortunately, every over/underclock-related tool I know accesses the GPU via the driver, which, as I've said before, doesn't even load, so I can't access any of the GPU's settings. I figured that maybe reflashing BIOS could at least drop the clocks down to factory defaults, but it didn't help at all either.
So the main question is, is there any way to salvage this card, even if it ends up underclocked? Also, are there any over/underclock tools that access the hardware at a lower level, circumventing the driver? What gives me hope is that if it doesn't show garbage during the BIOS sequence, then maybe not everything is lost. However, that might be because the textmode output doesn't go through the VRAM at all (unfortunately, I'm not too familiar with that level of technical detail).
I would be grateful for every constructive advice and/or expert opinion on the matter. I do eventually plan to upgrade my hardware, but I also had plans for this particular card, so if it's at all possible to save it, that's what I'm looking for in any case.
0 commentaires:
Enregistrer un commentaire