Is this like the third time I've said that in this thread?....;) Yes. It's been running consistently in PCIe4 mode for the last 14 months, if not longer. When I bought the 5700XT and the x570 Aorus Master on 7/7/19, it took either one or two months before PCIe4 functioned properly with the 5700XT. x570 mboard were the first to support PCIe4, and the 5700XT was the first family of 3d cards to support PCIe4. So, PCIe4 mode--tested and confirmed with 3dMark's little PCIe bandwidth tests--has been running fine with the 5700XT for 14-16 months. And even when GB was getting the initial AGESA's from AMD and just beginning to debug it's x570 Master bios versions--I never heard any interference of any kind with my PCIe sound device.
PCIe4 devices consume double the bandwidth of PCIe3 devices. If a device sitting on the PCIe bus is trying to pull down too much bandwidth for itself then other PCIe devices may suffer. Before PCIe this used to be common in the old days with PCI buses--GPUs often "hogged the bus" creating bus contention with the sound devices/card on the PCI bus, causing crackling and popping. The reason why he no longer has a problem when putting his card in PCIe3 mode is because it's now limited to pulling down half the bandwidth it's trying to pull in PCIe 4 mode. If it wasn't simply a matter of bandwidth then moving to PCIe3 mode would not solve the crackling/popping. In PCIe4 mode his GPU is trying to pull too much bandwidth--that's all it amounts to.
Since these are nVidia's first PCIe4 mode drivers for its first PCIe4 GPUs, I think it's pretty much a sure thing that a driver revision should fix this for running in PCIe4 mode! I hope it's clear this time--if I wasn't running in PCIe4 mode myself I'd likely not be saying so...;)
OK makes sense. I don't see how bandwidth would be an issue though - the GPU is on the direct link to the cpu, the audio card is on the chipset link, sound cards only use pci gen 1 x1 anyway so need very little bandwidth.
