Random items generated across multiple GPUs (including 5600G iGPU) in the latest release
Random items generated across multiple GPUs (including 5600G iGPU) in the latest release
Hello everyone! I faced a problem with my new build CPU: AMD Ryzen 5 5600G MoBo: AORUS B550M ELITE (F15 BIOS). The RAM was AMD Radeon R9 Gamer Series 16GB (2x8GB) and I had two GPU kits. I used Sapphire Nitro, RX480, GT710, and an iGPU PSU from Cooler Master MWE 750 WHITE. After a fresh Windows 10 installation and drivers, the AMD GPUs started flickering, showing artifacts, wrong icons, freezes, and other issues. The GT710 drivers made the output almost unusable. Device Manager reported error codes like 31 and 43 depending on the condition. Even with official AMD drivers, Gigabyte and AmazonZONE drivers didn’t fix it. My dedicated cards ran perfectly using basic drivers, so the problem wasn’t there. Memtest86 confirmed the RAM was fine. I checked my second PC (I5-3570, 16GB DDR3) with the same SSD from a new system—software seemed okay. It looks like either the CPU or the motherboard module might be faulty. I’m not sure if RAM is the issue, but I can’t rule it out completely. I tried different driver setups: normal installation, disabling auto-updates, and isolating from network. All passed stress tests. The PSU and CPU were new and tested. I’ll check the PSU on my second PC later and update you soon. UPD2: Swapped the PSU to my second PC; after 10 minutes of Furmark, everything worked. But a 750W PSU at 200W load is barely enough.
If several graphics cards are present, consider checking the video cable, motherboard, or power supply. You can verify RAM stability using memtest – https://www.memtest86.com/download.htm. Create a bootable USB and run it to ensure no errors appear during startup for about 20-30 minutes; stable performance indicates healthy RAM. Power supply issues may arise if it delivers incorrect voltages, though integrated graphics should still function adequately. This could point to a faulty cable or monitor. Have you tried overclocking by adjusting the bus speed? That might introduce glitches. Use Aida64, BIOS settings, or CPU-Z for diagnostics. OOOH... Another possibility: Error 43 suggests the card was flashed with a mining BIOS, which can lower default GPU clocks and alter RAM timings, potentially causing VRAM issues. Visit the VGA BIOS archive on techpowerup.com, download the correct original BIOS for your card, and apply it via AMD Flash tool here: https://www.techpowerup.com/download/ati-atiflash/. For RX480, consider using an older version (like 2.93) – find it under blue download options and select “Show older versions.”
Cable seems improbable because I used identical cables and the monitor on another bench; it doesn’t seem to interfere with the system. The screen freezing and unresponsive function keys (like num lock) are concerning. PSU might be the problem too, but since it came new, it should be fine. As noted, I ran memtest during troubleshooting—four passes all passed. No XMP enabled, just a straight out-of-box setup. I also performed several CMOS clearing steps while diagnosing.
The lower configuration lacks complete 16 lanes. Also, the iGPU seems to be behaving oddly.
I discovered this while checking online tutorials and repair guides. Since it functioned well on another setup without re-flashing the GPU BIOS, it’s likely not the primary problem. Also, it shouldn’t impact the GT710 or iGPU.
Removed the item completely. Used a reliable approach to clear everything and reinstall one by one after settling down. It turned out the single RAM stick was defective. 16GB or 24GB setups function properly, regardless of installation sequence.