When PCIe x16 is set to Gen 4, system doesn't boot

There are some problems on the AM5 platform, besides the issue that was mentioned here.

There is also a bug in the Zen 5 microarchitecture that prevents program activations through iLok License Manager too.

About the second issue, AMD stated that they are working on a new firmware: AGESA 1.2.0.2A, which should come out in early november 2024.

Maybe that firmware update will solve both issues.

So any update on this? Facing this same thing.
RMA'd the GPU, then sent it back today saying its fine. I haven't yet tried to force gen 3 on it. I read somewhere that we require a firmware update on the GPU (4080S in my case). I have an old 1070 runs fine on Auto in BIOs.

https://www.nvidia.com/content/DriverDownloads/confirmation.php?url=/Windows/uefi/firmware/1.2/NVIDIA_UEFI_Firmware_Updater_1.2-x64.exe&firmware=1&lang=us&type=Other

Hi Scott,

Were you able to solve the problem and get your GPU running with Gen4? I had similar issues (lock-ups/frozen screen) when I upgraded from my GTX 1080 to a Palit 4080 Super GamingPro in my recently build rig with a 7800X3D, 32GB at 6000MHz (EXPO), a PCIe 4.0 Kingston Fury SSD, 750W PCIe 5.0 PSU and X670E Steel Legend with BIOS 3.01 (Windows 24H2).

I had almost the same issues you reported, but after 5 or (many) more reboots I was able to get the system running without no issues at all. When that happened, I've tried many games, run Furmark and everything was fine with GPU-Z reporting PCIe 16x 4.0 (no riser cable was used). Then, I rebooted and got again the same problem: the system locked up as soon as I logged in (mouse cursor stuttering, big keyboard lag and a few seconds later the screen froze) or it just froze right after Windows finished its boot with occasional self-reboots in this case. I have noticed that the lock-ups coincided when the GPU entered in a low power state, i.e., when the GPU fans stopped spinning. So, as suggested by other threads, I have tried setting Nvidia 3D Power Management mode to Prefer maximum performance, but I had no success at all.

By the time, I have not thought of changing the PCI interface from auto to Gen3, since (after many reboots) the GPU/system has worked with Gen4. However, I've changed the number of active PCIe lanes in the BIOS, changed the SSD to another M2 slot, changed the HDMI cable, tried a DisplayPort output, removed the EXPO configuration, reset the BIOS to its default values, changed off and on the dGPU only option too, changed the 12x6-2 cable and tried the Y-cable adapter to use the PSU PCIe power connectors and used an external bootable drive with Windows. Everything failed. I've also tried a fresh Windows install (both Windows 10 and Windows 11 23H2) but couldn't finish the installation since the system was locking up as soon as the NVIDIA drive was installed, and I had to reinstall my 1080 to get the installation finished. Then, reinstalled the 4080 and got all the lock-ups again.

Since I thought it was a GPU issue (my 1080 is still running with no problem at all), I've sent the 4080 back to the vendor and got reimbursed (they did not tell me if the GPU really had a problem). I am looking to buy another GPU, but it is quite hard to find a good priced 4080 (if you can find one here, they are almost the same price as a 5080) or a 5080 in stock. However, I am afraid that the issue is with my CPU (like in https://forum.asrock.com/forum_posts.asp?TID=67762&PID=151169𤺁m, which had something very similar to our issues) or with my motherboard. They are both under the warranty period, but I am still hoping that a new GPU will fix the issue.

In many years building rigs and working with many different computer configurations, I've never seen such odd and intermittent issue.

Best regards,
Cristiano