I'm troubleshooting an ASUS ROG Astral RTX 5080 OC that consistently reports PCIe 5.0 x8 under load instead of PCIe 5.0 x16. The system uses an ASUS TUF Gaming Z790-Plus WiFi motherboard with BIOS 1836, an Intel Core i9-14900KF, 32 GB of DDR5-6000 with XMP enabled, and a Samsung 990 Pro 2TB PCIe 4.0 SSD. The power supply model still needs to be confirmed.
I've updated the BIOS and NVIDIA drivers, cleared CMOS, run GPU-Z's render test for more than a minute, tested Steel Nomad and several games, and monitored the system with HWiNFO. Windows reports no WHEA errors, the SSD performs normally, and GPU temperatures and power consumption look normal. The GPU support bracket is also installed.
The motherboard manual indicates that the primary M.2 slot should not reduce the main PCIe x16 slot to x8. Since the SSD is PCIe 4.0 rather than PCIe 5.0, I'm unsure whether lane sharing, a BIOS problem, an improperly seated CPU, bent LGA1700 socket pins, a motherboard fault, or the graphics card itself is responsible.
Has anyone run this motherboard and RTX 5080 combination at PCIe 5.0 x16? Could a BIOS version affect lane negotiation, and which component would be the most likely cause?
3 Answers
The next step is to inspect the LGA1700 socket carefully under strong light or magnification. A slightly misaligned pin or uneven cooler pressure can affect PCIe lanes. After reinstalling the CPU with even mounting pressure, check the motherboard’s System Agent and PCIe configuration screens to see whether the slot is being detected as x8 or x16. Testing another graphics card in the primary slot would also help: if another card links at x8, the CPU/socket or motherboard becomes much more likely than the RTX 5080. It’s also worth checking bifurcation settings and temporarily removing the extra M.2 drives.
Another owner reported the same x8 behavior with an ASUS ROG motherboard and an Astral RTX 5080. They tried the recommended M.2 slots and even removed the additional drives, but the card still operated at PCIe 5.0 x8. Performance remained very good, so the practical impact may be small, although the persistent link width still warrants checking the socket, BIOS settings, and board warranty. If the system passes all diagnostics and another GPU behaves the same way, an RMA would be reasonable.
Some ASUS boards share lanes between particular M.2 slots and the main graphics slot, so moving the SSD to a different documented slot is worth testing even if the manual suggests the primary slot should remain x16. Temporarily removing every NVMe drive except the boot drive—or booting from a live USB—can rule out lane allocation and bifurcation issues. The 14900KF has a limited number of CPU-connected PCIe lanes, so the exact slot layout matters.
The computer is currently being checked by a repair shop, so I can’t inspect it myself. I’ll ask the technician to examine the socket pins, verify the PCIe and bifurcation settings, and test another GPU if possible. If a second card also negotiates at x8, I’ll focus on the motherboard, socket, or CPU rather than the 5080.