I experience artifacts and BSODs, but only during tests with OCCT.
I experience artifacts and BSODs, but only during tests with OCCT.
So I was assisting with a GPU issue and planned a similar stress test on OCCT to help diagnose the problem. However, whenever I run the GPU stress test at 1080p with shader 7, I immediately start seeing artifacts. Then I either encounter a BSOD or drivers crashing, followed by a full recovery. I get the message "AMD display drivers have crashed and has recovered." I tested at 1178/1574 clock speeds, and initially thought it was an unstable OCCT, but it keeps happening even after lowering to 1100/1500. I’ve confirmed the 1178/1574 version works fine with Kombustor, OpenGL, and FurMark. Is this normal or is something wrong with my setup? I know OCCT can be demanding, but maybe not in this case. Also, MSI Afterburner now won’t change my clock, so I’m using TRIXX—does that seem unusual? Can someone explain what’s going on?
It seems the setup remains unstable.
Check if the graphics card was the sole component adjusted.
Have you ever run a stability check on your gear prior to overclocking? If not, you truly lack clarity about whether your system was stable beforehand, making it hard to determine if any part was questionable or damaged from overclocking.
Which AMD model operates at 1100 stock and is still overclocked? Verify its performance using AMD's reference timing before assuming a defect. Not every chip behaves the same, or even consistently.
Is your power supply a Tier 1 or Tier 2 unit? Any issues with quality or capacity could be contributing factors.
It seems the setup remains unstable.
Check if the graphics card was the sole component adjusted.
Have you ever run a stability check on your gear prior to overclocking? If not, you truly lack insight into whether your system was stable beforehand, making it hard to determine if any part was compromised or damaged from the process.
Which AMD model operates at 1100 stock and is still overclocked? Verify its performance using AMD's reference timing before assuming a defect. Not every chip behaves identically or consistently.
Is your power supply a Tier 1 or Tier 2 unit? Any issues with quality or capacity could be contributing factors.
yea, it should appear on the signature. It's a 650G2 and a 7970, and I made the changes myself. It functions at a bit below 1100, but not exceeding it. I've already checked stability beforehand. I'm familiar with OCP procedures, though this particular section feels a little unusual.
Yes, your signature refers to a 7970, but I don't interpret signatures to understand what questions people ask. People might use multiple rigs and not always work on their own gear, and a person's signature could be quite long if you consider it. 7970 never reached 1100. If it improves below that number, that's your stable overclock. Increasing the voltage to the GPU might boost your stable clock speed, though it could also generate more heat.