Hey guys, I have an MSI R9 280x 3GB I bought new in 2014. Since I bought it I've had issues with overheating and major artifacting, which now causes BSODs and instability in windows. I have a GTX 650 Ti I've been using since I realized the source of my system instability, but the performance is so lackluster in comparison to the 280x. So I have been pursuing any semblance of stability from this card in any way I can.
1st I replaced the thermal paste, which was essentially non-existent, I could read every letter printed on the die.
2nd I DDU'd my nVidia drivers and reinstalled AMD drivers, of course.
But now, I'm stuck. I wanted to underclock the GPU both in terms of core and memory clock, and possibly undervolt the card, in any attempt to reduce heat and find a stable clock. I assumed that the factory overclock on this card (1020 MHz core, 1500 MHz mem) was unstable for my silicon in particular, causing the problems. However, any slight variation of clock speed below the factory overclock produces MAJOR artifacting, with the screen jumping up and down rapidly and becoming unbearably jittery. I found this strange because I can still overclock the card, just not underclock it. So I followed my overclocking instincts and decided that since my temps were more reasonable now and underclocking was out of the picture, I should give the card more voltage/power to compensate for the overly high factory overclock. This helped for a little while, and I played a couple of games on it, but when I closed Overwatch and tried to reopen steam, it gave me an error, then my PC froze and a BSOD appeared. I was hoping for advice on this, as I would like to get some use out of this card before I call it quits and sell it for parts. My current overvolt is only 6 mV, but I was not comfortable upping the power by such great leagues as a regular overclock because I am already having such huge issues that may be caused by overvoltage. So my questions are: Is there a better way to achieve stability on a corrupted, artifacting card like mine, Are there methods of underclocking that will avoid the screen-jumping/tearing/artifacting, and Are there any recommendations on voltage at my factory overclocked frequencies? Thanks in advance
1st I replaced the thermal paste, which was essentially non-existent, I could read every letter printed on the die.
2nd I DDU'd my nVidia drivers and reinstalled AMD drivers, of course.
But now, I'm stuck. I wanted to underclock the GPU both in terms of core and memory clock, and possibly undervolt the card, in any attempt to reduce heat and find a stable clock. I assumed that the factory overclock on this card (1020 MHz core, 1500 MHz mem) was unstable for my silicon in particular, causing the problems. However, any slight variation of clock speed below the factory overclock produces MAJOR artifacting, with the screen jumping up and down rapidly and becoming unbearably jittery. I found this strange because I can still overclock the card, just not underclock it. So I followed my overclocking instincts and decided that since my temps were more reasonable now and underclocking was out of the picture, I should give the card more voltage/power to compensate for the overly high factory overclock. This helped for a little while, and I played a couple of games on it, but when I closed Overwatch and tried to reopen steam, it gave me an error, then my PC froze and a BSOD appeared. I was hoping for advice on this, as I would like to get some use out of this card before I call it quits and sell it for parts. My current overvolt is only 6 mV, but I was not comfortable upping the power by such great leagues as a regular overclock because I am already having such huge issues that may be caused by overvoltage. So my questions are: Is there a better way to achieve stability on a corrupted, artifacting card like mine, Are there methods of underclocking that will avoid the screen-jumping/tearing/artifacting, and Are there any recommendations on voltage at my factory overclocked frequencies? Thanks in advance