News AMD Suggest Possible Fixes for USB Connectivity Issues

[Admin]

AMD's support gives the company's first official advice on resolving USB connectivity issues.

AMD Suggest Possible Fixes for USB Connectivity Issues : Read more

 

[MarsISwaiting]

1.) Set PCIe mode from Gen4/Auto to Gen 3 in the BIOS

"we sold you Gen4.0 PCIe in Beta Version"

 

[ZolaIII]

I had the problem with Gigabyte B550 board and it whose only related to USB 2.0 controller. At first it got totally overrun, with bios update "Improve connection stability for USB 2.0 ports of USB hub" it only lagged, with second update "Improve USB 2.0 compatibility" it's working as it should.

 

[waltc3]

PCie4 has been running fine in my x570 Aorus Master, 5700XT for the past 17+ months. My GPU and my boot NVMe driver run at PCIe4 speeds. I have zero problems with USB. The only GPUs I've seen that are being shunted back to PCIe3 mode are nVidia RTX-3k GPUs, which seem to gobble up all available bandwidth on the PCIe bus when running in PCIe4 mode.

 

[kiniku]

I have had USB problems with USB 2.0 devices causing problems. In particular, my Creative Katana Sound Bar would freeze and stop Windows. The only remedy was a hard shut off. I have a Gigabyte Aorus X570 Elite (wireless). It looks like my problem may be solved or at least greatly improved by updating my bios and then updating the AMD chipset drivers.

 

[daworstplaya]

The only issue I (and my friends) have had has to do with the HP Reverb G2 VR HMD, that device refuses to work on most USB 3.0 ports on x570, x470, b550, b450 and b350 boards. I finally got mine to work on the front USB 3.0 header after testing all the ports on the back. One of my friends used a powered USB 3.0 hub to get the HMD to work on their AMD motherboard. No other VR headset I own has had issues with these chipsets.

 

[Super_Duty1]

Just recently built (within the past month) a new rig, x570 Aorus Elite mb, Ryzen 5600x cpu , 2tb Aorus Nvme SSD, MSI RX6800 gpu. Any time that I would insert a USB drive, the machine would crash and not reboot to windows (boot drive not found), requiring a windows reload. Fix so far, was as indicated in AMD's comments was to make sure that ALL the latest updates (windows, drivers, bios, etc.) were at the latest levels. Seems to be related to PCIe Gen-4, and all things playing nicely together. My observations have been that you need to defer loading any software until you confirm that all updates have been applied. Meaning to manually request and install all windows 10 updates as they are not all installed immediately upon a new install. although it appears that they have been applied. Since doing so, my rig is mostly stable.

 

[InvalidError]

So, on one hand we have AMD fans bragging about AMD beating Intel to bringing PCIe 4.0 to the mainstream a few years earlier and on the other hand, we have AMD saying that you may need to turn that off to solve intermittent connectivity issues.

The joys of early adoption, you never know when you will get burned.

 

[Super_Duty1]

Personally I was not buying or selling with regard to Intel vs. AMD hardware. Built pretty much all of them over the decades, even Motorola and Cyrix way way back, math co-processors, or setting up hard drives manually configuring # of cylinders, platters, ugh. But yes, early adopters have always had to jump through hoops. Couldn't even guess how many OpenGL gpu driver combinations I've installed back in the 3dfx days... but when it worked, just awesome, thinking about Unreal / Unreal Tournament. Was young then, now I very much like "plug and play", mostly...

 

[UWguy]

"we sold you Gen4.0 PCIe in Beta Version"

How about “you’re holding it the wrong way” defense?

I laughed at the suggestion of moving settings to gen 3. None of their suggestions give me a warm fuzzy feeling, it’s like they don’t even understand what is truly going wrong.

 

[InvalidError]

I laughed at the suggestion of moving settings to gen 3. None of their suggestions give me a warm fuzzy feeling, it’s like they don’t even understand what is truly going wrong.

If AMD knew what went wrong, it would have said exactly what the issue was and whether it was fixable by AGESA update.

Since it won't make any specific comments about the issue, either it hasn't found the root cause yet or it is something that cannot be fixed (ex.: a hard-wired race condition that is more likely to occur at 4.0 speed than 3.0) and it is hoping people will let it slip instead of initiating a class action.

 

[digitalgriffin]

I heard a lot of these problems were to chips that handled more than 1 usb port per channel.

If there's a c state change I could see this causing state issues on the split channel.

It's just a guess.

 

[PlasmaTorus]

How many of those affected are using non-ECC memory? How many have run MemTest to check for memory corruption? Sounds like a common symptom of memory corruption or CPU overclocking.

 

[Flyfisherman]

Well I'm having a Samsung SSD 980 Pro (PCIE 4.0) and the suggestion to change this to PCIE 3.0 is ridiculous on my Asus ROG Crosshair VIII Hero X570 (latest bios AGESA 1.2.0.1 with an AMD Ryzen 9 9350X CPU and 16GB 3200MHz Ram with all the latest chipsets.

And the ASUS TUF Gaming Radeon RX 6900 XT OC graphics card bought just a few week ago here in Sweden - of course the latest graphics drivers.

AMD has to do better than this.

Btw. I've already set the Disable Global C-State in the BIOS.
I'm using @Yuri Bubliy CTR optimus overclockling/under volting tool https://twitter.com/1usmus
ClockTuner v2.0 RC4 for Ryzen (CTR)

And yet, once in a while I'm still hearing the unmistakingly sound of an USB-device being disconnected/connected.
But I've not discover any problem what so ever with this.

Best regards from Sweden.

Ps. Perhaps I should notify AMD about my experiences.

 

[Jim90]

So, on one hand we have AMD fans bragging about AMD beating Intel to bringing PCIe 4.0 to the mainstream a few years earlier and on the other hand, we have AMD saying that you may need to turn that off to solve intermittent connectivity issues.

The joys of early adoption, you never know when you will get burned.

--> The joys of early adoption, you never know when you will get burned.

(This applies to any company/any group of fans. Beware, though, of that loud army of ever present, paid shills)

 

[jepeman]

"we sold you Gen4.0 PCIe in Beta Version"

AMD in nutshell. We sold you a beta AGESA, bios, Gen4.0 PCIe, Graphics cards drivers, etc.

Im also an AMD beta tester, just to let people know.

 

[InvalidError]

with an AMD Ryzen 9 9350X

It is a sad day when people are sufficiently upset with currently available parts that they start building time machines to go shopping in the future instead.

 

[Sam Bi]

Had a problem where a wireless keyboard and mouse combo was losing connection. I realized that this started to occur after I upgraded my HDD to a SSD. The SSD was mounted right behind the USB port that the BT dongle was inserted. Moved the dongle to a USB port on the back of the PC and the problem disappeared. Is it possible that some of the SSD produce RFI with enough RF energy that it would could cause havoc with other components in the PC. Is this problem reported only in PC’s with SSD ? Just wondering.

 

[MarsISwaiting]

One Question : How about Threadrippers with Gen 4.0 PCIE ?

 

[InvalidError]

One Question : How about Threadrippers with Gen 4.0 PCIE ?

The sample size is probably too small to generate a significant number of complaints. TR also uses a different IOD and chipset, so the USB bug may not be applicable at all.

 

[Super_Duty1]

Just an update, since my previous post. I since have written iso/dd images to 5 backup USB 3.0 drives without issue since making sure my rigs software had all the latest updates. This also involved removing and reinstalling my 2.4ghz usb thumb drive for my wireless headset to make the USB port available. I will this thread again should something change.

 

[Flyfisherman]

Yeah, this is a real xxx question for AMD to solve as the title of the thread suggest.

I'm for one.
The suggestion to change this to PCIE 3.0 is ridiculous on my Asus ROG Crosshair VIII Hero X570 (latest bios AGESA 1.2.0.1) with an AMD Ryzen 9 9350X CPU and 16GB 3200MHz Ram with all the latest chipsets.

 

[MarsISwaiting]

The sample size is probably too small to generate a significant number of complaints. TR also uses a different IOD and chipset, so the USB bug may not be applicable at all.

Well , if Threadrippers dont have the same problem , then shame on AMD ... it is not GEN4.0 too early , it is bad Ryzen Design !

 

[InvalidError]

Actually your GPU is deriving no performance boost from PCIe4.

The 4GB RX5500 can be up to 70% faster on 4.0 than 3.0, so 4.0 is starting to pay dividends at the low-end where GPUs need fast access to system memory to offset their VRAM deficit. That makes me optimistic about how well the 3050(Ti) will perform despite still having only 4GB once Nvidia gets around to launching it.

 

[jpe1701]

The only thing I've seen for a problem on my x570 aorus master is just a minor annoyance where one of my usb 3.0 front ports doesn't notify you when something is connected or disconnected. I first thought it wasn't working at all because nothing happened when I connected a flash drive but it was listed in file explorer. There isn't a listing for it to be removed either unless I right click the drive in explorer. The other header on the motherboard works fine though.