UPDATE 2: after running vram test for 50 mins i got 3 errors 1 at 30 mins and 2 and 48 and 1 at 54min and 1 at 55min and 1 at 58.
So my gpu is broken?
UPDATE: no errors found after 1 hour of occt 3d adaption. Now running vram test its been 10 mins and no error yet
Specs: RX 470 XFX (with RX 570 XXX Edition cooler), i5-13400F, Colorful H610M-E, 8GB DDR4, 240GB SATA SSD, 500GB NVMe, asus 550W Tuff Gaming 80+ Bronze PSU.
I was getting random artifacts (small green/red pixels) during light tasks and gaming, along with crashes and driver errors (Code 32/43). Temps were high (~85°C), and I suspected a BIOS mismatch—GPU-Z showed a single-fan RX 470 BIOS, but my card has a dual-fan cooler. After flashing the BIOS, I ran Heaven Benchmark and other tests without errors, but Minecraft caused artifacts the next day.
I flashed a different official BIOS from TechPowerUp, and now it seems stable. I’ve been running OCCT 3D Adaption for 35 minutes with no errors (before, I’d get millions of errors). I also ran FurMark 2 1080p benchmark and didn’t see any artifacts. I ran Heaven Benchmark for a few scenes, and no issues appeared.
In the past, I ran a VRAM test in safe mode (using VMSD) on the previous BIOS, and there were no errors. I haven’t tested VRAM with the current BIOS yet. Currently, I’m still running the OCCT test for 1 hour. I’ve noticed the GPU voltage is between 1.11V and 1.14V, and wattage fluctuates between 87W and 120W. I haven’t checked the exact GPU temp yet, but I estimate it’s around 79°C to 83°C, possibly reaching 80–88°C under stress (I’m not touching the PC to check right now). I’m unsure if OCCT stresses the card like FurMark does.
I haven’t seen any artifacts so far, and temps seem normal. Previously, I would see small pixel artifacts (1–2) after a while, but that hasn’t happened yet with the new BIOS. I also disabled hardware acceleration in Chrome, just to be thorough.
I’m still concerned, as the card worked fine for a bit before, then artifacts returned. Could overheating VRAM or GPU cause these intermittent artifacts, or was it a BIOS/driver issue? Everything seems stable now, but I’m unsure. Should I continue testing, or is it safe to assume the GPU is fine? Or should I consider selling it?
(Chatgpt wrote this as English is not my mother language)