AMD acknowledges USB connectivity issues on X570 and B550 motherboards

[Irata]

The only reason why you are seeing more RTX is because it is the best selling product.

Could be a reason, but it‘s odd that there are no 5000 series Radeon cards (those are PCIe 4) listed.

 

[mongeese]

Could be a reason, but it‘s odd that there are no 5000 series Radeon cards (those are PCIe 4) listed.

It seems (to me) that the severity of the disconnection issue is related to the amount of information passing through the bus. So even though they're PCIe 4.0, I don't think they're powerful enough to trigger the issue. I am curious about why the 6000 series doesn't trigger the issue, but I imagine it's related to the architectural differences. I disagree with MaxSmarties assertion that there aren't enough AMD GPUs being sold.

As I've said above, I do think AMD should have done more compatibility testing with Nvidia GPUs when they released, because the issue seems pretty widespread.

 

[Irata]

As I've said above, I do think AMD should have done more compatibility testing with Nvidia GPUs when they released, because the issue seems pretty widespread.

The difficult part as always with AMD issues is gauging how widespread it actually is.
Gamers Nexus said they believe it is more than a few (as AMD mention in their Reddit post) but they also say the issue is very difficult for them to replicate.

If you had users actually owning the hardware and actually encountering the issue discussing this without a deluge from trolls flooding any discussion with fud it would be much easier to troubleshoot and get a feeling how widespread the issue is. It certainly exists.

But alas, as always this is not happening. Sadly.

 

[HardReset]

you should try DLSS before speaking...
but everyone knows you and your bias.

Several reviews pointed out there are situations where DLSS is slightly better than original resolution, because DLSS isn't just "upscaling" ...

Examples please?

DLSS is mainly upscaling and that's more than enough to say it cannot be better than original. Simply said, 720p picture "scaled" (guessed is better term) to 1440p could not be better than 1440p original. Except for some very special cases where original is something it should not be.

It seems (to me) that the severity of the disconnection issue is related to the amount of information passing through the bus. So even though they're PCIe 4.0, I don't think they're powerful enough to trigger the issue. I am curious about why the 6000 series doesn't trigger the issue, but I imagine it's related to the architectural differences. I disagree with MaxSmarties assertion that there aren't enough AMD GPUs being sold.

As I've said above, I do think AMD should have done more compatibility testing with Nvidia GPUs when they released, because the issue seems pretty widespread.

That logic does not really work here. Single PCIe 4.0 NVMe drive easily makes around 6 GB/s traffic. Something that is very hard to achieve with USB devices, no matter how many. Also AMD GPU's using SAM should create much more PCIe traffic than Nvidia GPU's do.

I predict problem is on USB power saving, power supplies or just software bugs (Windows, Linux, drivers, BIOS, AGESA...).

 

[Guillermo Belli]

Reading the symptoms people have, I noticed something similar a few weeks agoin on old Llano computer I have (A8-3870 with GA-F2A88X-UP4).

I did a cleaning install of Win 10 and updated everything. As I had this on a desk with no ethernet plug nearby I used a USB WiFi adapter. I noticed the adapter disconnected periodically for no reason, but it worked fine on another computer. The mouse (wireless) also felt unresponsive at times, but I blamed USB3 RF interference for that, although I had no USB3 devices connected at the time. Tried different ports and the issues was less pronounced in some ports that others, but it was still there.

I'm using the integrated GPU, no dedicated video here.

Might not be related to the Ryzen USB issue, or maybe it is and it's something to do with a Win 10 update. Maybe the USB controller on this board is just failing.

 

[terzaerian]

Frankly I wonder if it's even so much an AMD/Nvidia/whatever issue as a USB issue. USB is becoming a bit of a clown show, from their baffling nomenclature for 3.0 releases to some of the issues brought up by people in this thread. I myself had a 3.0 disk enclosure where the 3.0 data lines failed (but the older ones didn't) followed later by a complete failure of the USB controller. The replacement "worked" but would go idle and basically sit for a minute any time I opened the disks to navigate folders, and did this no matter what I tried (re: disabling USB suspend settings). The whole experience prompted me to build a NAS instead.

 

[Mr Majestyk]

5800xt with 3080 on an Asus B550-F and fortunately I never had the issue.

But how many USB ports do you use? The problem does not affect everyone which makes it all the more weird, and some MB's seem worse such as Gigabyte Aorus.

Anyway I got an Asus X570 Gaming Tuf Pro and 3700X along with 2080 S since I couldn't get newer gen CPU or GPU at sensible prices.

 

[Tantor]

Gigabyte B550 Aorus Pro AC.

Ubuntu Linux - Zero problems with USB. It's rock solid.
Windows - USB locks up when using bios F13a. Switched back to bios F12 and the USB problem disappears.

This is a Bios problem, and only manifests in Windows.

 

[Irata]

But how many USB ports do you use? The problem does not affect everyone which makes it all the more weird, and some MB's seem worse such as Gigabyte Aorus.

Anyway I got an Asus X570 Gaming Tuf Pro and 3700X along with 2080 S since I couldn't get newer gen CPU or GPU at sensible prices.

It seems to be common with VR headsets. Playing with an RTX 3xxx plus using a VR headset over USB should load PCIe and USB considerably with all the data that needs to be moved.

Perhaps that‘s why you don‘t see many graphics cards with a USB port for VR in the spreadsheet.

 

[Biostud]

I have the exact same problem with a B450 board and a 3800X. The problem arose when I updated the Bios to support 5xxx series, so I think the problem is in the AGESA code.

 

[Avro Arrow]

Well this is just....weird. I mean, things like USB controllers and audio chips aren't exactly things that could be considered "new". I guess that it might have something to do with the fact that the A/B/X5xx series chipsets are the first ones to use USB 3.2 instead of just 3.1 but that doesn't explain the audio issues unless the USB controller isn't sufficiently shielded to prevent electrical interference.

Still.....weird.

 

[Stan1ey]

Same problem here, corsair K63 keyboard or logitech G305 mouse randomly disconnecting for few seconds, I have tried all of my usb ports also hubs, bios updated, newest drivers but still the same problem. I hope this can be fixed as the hardware cost a fortune this days and you would expect a stable operation obviously. I have an Asus rog strix X570-I mini-ITX motherboard, Ryzen 5900x and Asus RTX 3090 TUF OC gpu

 

[Avro Arrow]

What motherboard are you using? Must be some cheap crap if you fan is whining because my wife's board is fine. As to the RT I have a 3080 and could totally live without it the RT, I'm using a 165Hz monitor and obviously I can't hit these frames with RT enabled in any game, I can agree on the drivers side though as only last week we sold my wife's 5700XT and got a 3070 mainly because of the drivers but from what I'm hearing RDNA2 doesn't have a any issues atm

I'm not 100% sure that it was a driver problem with the RX 5700 XT. I say this because originally, my RX 5700 XT was crashing left and right but those were crashes that caused a system restart. Generally, crashes like that are a result of a device drawing more power than the PSU can supply. I knew that my PSU wasn't the problem because it's an EVGA 1000 G2 Supernova and it had no problem with two R9 Furies. Each R9 Fury draws more power than any RX 5700 XT and when I swapped it out for one of my R9 Furies, without changing drivers, all problems immediately ceased.

The RX 5700 XT that I had was an XFX RX 5700 XT Triple-Dissipation. I sent it in for an RMA and XFX sent me back a THICC III. When I went to install it, I noticed that the THICC III had two 8-pin connectors while the Triple Dissipation had an 8-pin and a 6-pin. My crashes immediately ceased and it didn't matter which driver package that I was using. I have a feeling that these problems came from having an 8 and a 6 instead of a couple of 8's. I also think think that reviewers didn't see the issues because AIB partners always send them their top-of-the-line versions of the card and those probably all had the twin 8-pin connectors on them. It's also possible that the 8 and the 6 get almost maxed out and are fine as long as there are no inherent defects in them, defects that could be hidden by having 8-pins in both connectors and therefore always allowing enough juice to get through.

 

[Avro Arrow]

It does not "irk" me, you can judge them however you want but in my 17 years of gaming nVidia and Intel were the bigger scumbags usually trying to screw AMD every chance they get, maybe if nVidia gets a taste of its own medicine they will stop this nonsense, I'm not stupid and I know they only care about my money but I do care how they go about this and nVidia does not go about it they right way, also there is nothing wrong with AdoredTV I linked his videos because they were the best on the topics but he is not the only one who did videos on this. nVidia purposely destroyed performance of AMD's cards and their OWN in many games just to get people to buy new GPU's , they have paid many developers to make their games work better on GeForce and been doing this for years and you bring out God Fall? a game that nobody even cares about

And the only small minded sheep here does seem to be you if you really think that AMD is on the same level of scummines as nVidia or Intel

Shadowboxer says that AdoredTV is full of lies and half-truths? I happen to know Jim personally and he's the best investigative tech reporter since Charlie Demerjian worked at theinquirer. Ignore him, he's so out to lunch that even Steve Walton, one of the most diplomatic men on Earth, made it clear that he considered him to be... (trying to think of a nice way of saying this) less than knowledgeable.

Steve saying something like that is akin to me saying something that would be deleted for violation of the TOS.

 

[Avro Arrow]

Could be a reason, but it‘s odd that there are no 5000 series Radeon cards (those are PCIe 4) listed.

Well, keep in mind that nVidia probably only has shortages because the yields from Samsung's 8nm process are most likely inferior to the yields coming out of TSMC's 7nm process. AMD has a shortage because Zen 3, RDNA2 and all the console APUs are coming from the same place and it's probably overloading TSMC's manufacturing capacity. That would result in far worse shortages than nVidia is experiencing.

 

[LeslieS]

Is this problem occurring with Linux systems. It seems to be a Microsoft Windows issue.

I was going to purchase an MSI x570 motherboard, and now I am hesitating. Is the problem related to a motherboard vendor or is it a software timing issue.

 

[TempleOrion]

This article is for the purpose of "calling out" AMD, in addition to spreading awareness of the issue. Bonus fun fact, my main rig uses an Nvidia GPU....

TBH You shouldn't have even given Shadowboxer the time of day, he's beyond irredeemable...

 

[LeslieS]

TBH You shouldn't have even given Shadowboxer the time of day, he's beyond irredeemable...

AMD is not a "fly-by-night" company. They do try to offer quality hardware and the like. If there is a USB problem, it could be a problem with the card design, it could be a timing issue with sofware, such as the bios (from the bios vendor), and a combination of all three.
Since the issue is random, and seems to be Windows related, as opposed to Linux issues, the problem may even be with microsoft.

Give AMD a chance to diagnose the problem and to provide a fix.

 

[Revrant]

These sound like the issues surrounding X570 and RTX from a way back? I had them, they claim the companies fixed them in some AGESA, but I was experiencing BSODs or sudden no post restarts and it was the PCI bus.

 

[Karl Hungus]

I have an MSI Gaming Plus B550 with a 5600x. I had a Sapphire Pulse 5700XT in it, and now have an EVGA XC3 Ultra 3070 as I moved the 5700XT to a spare parts PC to mine. I've never had a single issue with the 5700XT, the MB, or the 3070.

 

[Mugsy]

I purchased my new x570 MoBo a year ago and immediately ran into trouble getting my two USB TV tuners to work. I contacted Asus at the time but they were of little help, eventually suggesting I remove my MoBo and mail it to them.

I wasn't about to go through the insane headache of disassembling my (otherwise) working PC, put it in the mail, and be w/o a computer for weeks just to diagnose something so "minor".

Instead, I purchased a PCIe tuner and gave up on trying to get the USB tuners to work.

It's amazing to learn nearly a year later that there was indeed a problem with the motherboard's USB controller after all (best I could tell at the time, it seemed to be a "voltage" issue.) Did not matter if I used the device on USB 2.0 or USB 3.x ports.

Hopefully, the problem can be fixed with a BIOS update when they figure out the cause.

 

[TechnoIlliterat]

The attempts at shifting blame from AMD in this thread forced me to register (and I'm a Ryzen CPU and X570 board user myself). It seems that most people did not actually bother to read through the relevant thread on Reddit before making their conclusions. So here is the summary of some information from it:
1. People confirm the existence of the problem on Linux = not a Windows problem
2. Disabling PCI-E 4.0 or disabling global C-states (Buildzoid/Actually hardcore overclocking workaround) DOES NOT solve the problem for many owners of these boards = PCIe is just a symptom, not the cause of the problem
3. The issue is not limited to the systems with PCIe gen 4 devices like the 3000 series NVidia GPUs or NVMe drives - people are reporting having issues on the systems with, for example, RX 500 series GPUs. So using Radeon GPUs solves nothing.
4. People are reporting having USB droupouts with exactly the same symptoms not only on 500 series motherboards, but also on 400 series and even 300 series boards. Possibly, even Threadripper boards are affected. But there it's less widespread.
5. The existence of the issue is confirmed for all major motherboard manufacturers and a wide spectrum of motherboards = not specific manufacturer or board issue.
6. The issue likely went under-reported for a long time - as a lot of people are commenting "oh, I thought it was just me" and blamed the issue either on their USB devices and or their specific board.

 

[Adi6293]

Shadowboxer says that AdoredTV is full of lies and half-truths? I happen to know Jim personally and he's the best investigative tech reporter since Charlie Demerjian worked at theinquirer. Ignore him, he's so out to lunch that even Steve Walton, one of the most diplomatic men on Earth, made it clear that he considered him to be... (trying to think of a nice way of saying this) less than knowledgeable.

Steve saying something like that is akin to me saying something that would be deleted for violation of the TOS.

He was, sadly because of people like Shadowboxer he quit.... I really miss his takes on tech : - (

 

[Shadowboxer]

Shadowboxer says that AdoredTV is full of lies and half-truths? I happen to know Jim personally and he's the best investigative tech reporter since Charlie Demerjian worked at theinquirer. Ignore him, he's so out to lunch that even Steve Walton, one of the most diplomatic men on Earth, made it clear that he considered him to be... (trying to think of a nice way of saying this) less than knowledgeable.

Steve saying something like that is akin to me saying something that would be deleted for violation of the TOS.

AdoredTV is absolutely full of crap. Just because you know him personally doesn’t make any difference. Honestly I would love to sit you down and educate you on how much misinformation is in his videos. It’s impressive how much effort he put into the propaganda. I really hope he got paid by AMD.

And Steve Walton diplomatic? Lmao! That guy is literally the most abrasive tech tuber around. Name another tuber who has managed to infuriate Nvidia. Although, other tubers don’t deliberately attempt hide Nvidias advantages.

One thing is for sure, I know a hell of a lot more than you do. You don’t have a clue..

 

[Adi6293]

AdoredTV is absolutely full of crap. Just because you know him personally doesn’t make any difference.

And Steve Walton diplomatic? Lmao! That guy is literally the most abrasive tech tuber around. Name another tuber who has managed to infuriate Nvidia. Although, other tubers don’t deliberately attempt hide Nvidias advantages.

One thing is for sure, I know a hell of a lot more than you do. You don’t have a clue..

Give me an example of Jim lying or hiding the truth from him viewers about anything? Because right the only person full of crap here is you......