AMD Fixes Intermittent USB Connectivity Issues on 500 Series Chipsets, BIOS Update Arrives in April

[efikkan]

While I can understand that not every bug is fixed on launch day, stability issues plaguing the platforms months after a product release is simply not good enough. And keep in mind that we are not talking about "theoretical" problems here, but something breaking system stability which is a deal breaker for anyone using a computer for something productive. It took about ~3 months before Zen 2 became reasonable stable, and Zen 3 is four months in and still getting regular stability improvements, and then I believe this particular issue is chipset related. AMD needs to step up their game here and do more extensive testing, even if this means products will need a month delay and larger groups of testers.

 

[AsRock]

Proof on that, or it's hearsay.
I mean, it's not as if Intel haven't had their fair of issues over the years.
In fact, at one point, SiS made a better chipset for Intel than Intel did, not long after Intel banned third party chipsets.

All so would of cost me a extra $100 to at the time of buying my mobo and CPU too. I am using but one USB port and not a single issue.

 

[Fouquin]

Intel has always been the more stable platform, dating back to the days when third parties were making chipsets for their CPUs. It seems AMD is keeping that tradition alive.

Yeah like nForce 680i. Man I remember when forums were filled with "My motherboard killed my RAID array/DRAM!" due to the FSB bug on those boards.

Or i820 and the MTH issue which led to a market wide recall of all the chips and boards using it.

Or i850s PCI write delay issue that limited bus bandwidth, or i850's RDRAM null cycle issue where it would prematurely terminate IIO to RDRAM and cause the whole system to lock up.

Or Northwood and the entire SNDS debacle. $400 CPUs falling like flies left and right.

Or how about Mobile 965 Express's IGPU issues.

Or H67/P67's Rev.B SATAII PLL defect?

Maybe want an asterisk on the "always" in your statement there.

 

[TheLostSwede]

Yeah like nForce 680i. Man I remember when forums were filled with "My motherboard killed my RAID array/DRAM!" due to the FSB bug on those boards.

Or i820 and the MTH issue which led to a market wide recall of all the chips and boards using it.

Or i850s PCI write delay issue that limited bus bandwidth, or i850's RDRAM null cycle issue where it would prematurely terminate IIO to RDRAM and cause the whole system to lock up.

Or Northwood and the entire SNDS debacle. $400 CPUs falling like flies left and right.

Or how about Mobile 965 Express's IGPU issues.

Or H67/P67's Rev.B SATAII PLL defect?

Maybe want an asterisk on the "always" in your statement there.

You got a seriously good memory.

I mean, it's just dumb to defend any company that sells you products, as there's no such thing as a flawless, faultless, perfect product, even less so when it come to technology.

That said, AMD should really spend a bit more time on getting their platforms ready before they launch them. Hopefully that'll happen with the move to AM5, as they should have a bit more money in their coffers now.

 

[Fouquin]

You got a seriously good memory.

I mean, it's just dumb to defend any company that sells you products, as there's no such thing as a flawless, faultless, perfect product, even less so when it come to technology.

That said, AMD should really spend a bit more time on getting their platforms ready before they launch them. Hopefully that'll happen with the move to AM5, as they should have a bit more money in their coffers now.

I use a lot of older platforms day to day and still have to step around these issues when compatibility testing.

Certainly, and don't confuse that list as defense of AMD. It's entirely to demonstrate Intel's capability to make mistakes. Most recently on that list I could add all the Spectre/Meltdown and now the new ring bus security exploits but honestly those should be fresh in everyone's mind.

 

[newtekie1]

VIA had to respin KT266 into KT266A and KT400 into KT400A, for example. And that's just otoh.
Mind you, they both ended up as excellent chipsets, but the initial problems is something Intel managed to avoid.

Remember when Intel had to respin their 6-Series chipsets because the SATA controller would degrade over a short time? Intel isn't free from issue, but AMD tends to run into more because they also tend to push the edge more than Intel. But X570 is AMD's first actual chipset in a really long time.

 

[TheinsanegamerN]

Please provide supporting evidence because I'm pretty sure that is a total myth. I know I've read an article or two about it in the past but can't seem to find one at the moment.

You want him to provide evidence of a negative? That's a logical fallacy. "Yes please prove something DOESNT exist".

 

[BigBonedCartman]

Ill take a USB issue over countless CPU security flaws any day

 

[sutyi]

Just ANUTHA reason not to buy Rev1 of any pc hdwr.....

The REAL question is (IMHO): why this was not discovered & fixed PRIOR to release ?????

Had ZEN1 + B350 for 3 years have not had any USB issues. Same CPU and mobo are still running 3000MT XMP at my nephew.
Had ZEN2 + B550 for 6 months running PCIe 4.0 graphics, 3200MT XMP had no USB issues. Always on the latest AGESA code.

Since ZEN3 has the same I/O die as ZEN2 I guessing there are a lot of dudes running unstable infinity fabrics and get all sort of anomalies... One of our biggest hardware forums has many ZEN3 owners and there weren't many USB stability complaints or I can't find them... A quick search there from 10.000 comments for "USB" only brings up 71 hits and only one of those had USB2.0 ports going completely inert after flashing a beta ASUS BIOS before it got pulled.

Don't know what they found or how did they fix it, but I'm guessing it must have been some magical combination of hardware and overlocks where the interrupts / signaling on the USB controller got wonky.

 

[evernessince]

Intel has always been the more stable platform, dating back to the days when third parties were making chipsets for their CPUs. It seems AMD is keeping that tradition alive.

Given that my Z170 Intel platform had USB issues with my Oculus Rift back in the day until I upgraded to AMD, I'd say the USB issue thing isn't exclusive to AMD. In fact I know it is, you can google it, tons of people had the same problem.

You want him to provide evidence of a negative? That's a logical fallacy. "Yes please prove something DOESNT exist".

It's called evidence of absence. A logical fallacy is that you can't disprove a negative.

Then again, Bug never provided any evidence to support his claim in the first place anyways so you really don't need evidence of absence to dismiss it in the first place. If you can't provide proof to support your own argument, you haven't make a argument in the first place. If you are taking this example and this example only as proof that AMD's platform is inferior, you are making an extremely poor case of it. As other posters have layed out, both companies have had issues in the past.

When you lay things out logically, it doesn't look so silly as "Intel good, AMD trash".

 

[MikeMurphy]

Yeah like nForce 680i. Man I remember when forums were filled with "My motherboard killed my RAID array/DRAM!" due to the FSB bug on those boards.

Or i820 and the MTH issue which led to a market wide recall of all the chips and boards using it.

Or i850s PCI write delay issue that limited bus bandwidth, or i850's RDRAM null cycle issue where it would prematurely terminate IIO to RDRAM and cause the whole system to lock up.

Or Northwood and the entire SNDS debacle. $400 CPUs falling like flies left and right.

Or how about Mobile 965 Express's IGPU issues.

Or H67/P67's Rev.B SATAII PLL defect?

Maybe want an asterisk on the "always" in your statement there.

Oh this is a walk down Memory Lane. Great post.

 

[thevoiceofreason]

Intel has always been the more stable platform, dating back to the days when third parties were making chipsets for their CPUs. It seems AMD is keeping that tradition alive.

It wasn't long ago that Intel's chipsets had a bug where they would simply stop working https://www.techpowerup.com/forums/...products-–-and-its-not-just-cisco-hit.230510/

I also remember the USB problems on early Haswell chipsets https://www.techpowerup.com/182462/intel-fixes-8-series-chipset-usb-3-0-erratum

 

[freeagent]

I doubt any of them can provide any proof they just prefer intel platforms.

I was on socket 1366 for 10 years and now AM4 1 1/2 years and this platform is just as stable as my previous intel rig.

Yup I rocked 1366 up until 3 years ago when I upgraded to a 3rd gen lol.. my 5600x/B550 combo is mint. Just as (if not more) stable than my last 2 rigs.

Better in every way..

I'm not going to say AMD fo life y0 or anything like that.. because if Intel comes back with something that is more enticing of course I will go back.. but for now.. this is alright..

Just need a GPU

 

[Geofrancis]

there was also the SATA problems on the intel 6 series chipsets

Intel Discovers Bug in 6-Series Chipset: Our Analysis

www.anandtech.com

 

[bug]

Uhm, the A versions weren't just plain respins though. It's like calling 200-series of Intel chipsets a respin of the 100-series.
Yes, the KT266 had an issue with AGP 4x, but I don't know what the KT400 was supposed to suffer from.

It was back in 2003, I worked at MSI in the UK at the time. It must've been the SiS 655 which had dual channel RAM. Unfortunately it's so long ago that I can't find any reliable benchmarks...
But yes, it was a bit of a one off from SiS, but it showed that Intel could face some real competition on the chipset side of things.

I'm pretty sure just about everybody could give Intel and AMD some real competition. My impression is that Intel is usually better with handing out documentation for others to build their stuff. Of course, in the meantime they locked everybody out of building chipsets. But you can still see the signs in USB and Thunderbolt implementations (the client part of Thunderbolt, so far).

 

[mechtech]

"USB 2.0 audio crackling (e.g. DAC/AMP combos),"

I think I have that on my Gigabyte X370........unless it's just a crappy DAC/AMP...........which is very possible lol

 

[deu]

Another reason to buy Intel. Stability from the get go.

Troll joke or hardware rookie not seeing the irony of the level of security bugs intel have dragged for years. Lets get mad about AMD fixing something!

 

[Panther_Seraphin]

Another reason to buy Intel. Stability from the get go.

Intel Discovers Bug in 6-Series Chipset: Our Analysis

www.anandtech.com

Why yes.....They are always perfect even in V1

 

[GamerGuy]

I have the excellent X570 Aorus Xtreme rev 1.0 and I have no issue with USB connections. In fact, I believe I've used all my USB 2.0/3.0 ports on back IO of my mobo, except for one remaining USB Type-C port. My front panel has 4x USB 3.0 and one Type-C ports which I'd use occasionally for external drives (2.5" portable HDD, Thumbdrive, etc). No problem at all....

To the guy who'd claimed that Intel is more stable, please furnish proof, just because you'd said so doesn't make it true.

 

[windwhirl]

Lets get mad about AMD fixing something!

Don't get AMD, get MAD? Or Don't get MAD, get AMD? LMAO

In all seriousness, though, both companies have their history of fucking up something here or there. No point in getting all defensive for one or the other.

 

[ixi]

Another reason to buy Intel. Stability from the get go.

Did you really just now say this? I don't even know whit what should I start.

 

[Patriot]

Intel has always been the more stable platform, dating back to the days when third parties were making chipsets for their CPUs. It seems AMD is keeping that tradition alive.

I guess you never used x79 or x99... both had significant teething issues.

I will take a USB disconnect over bsod anyday. Also seems to be a signal/noise issue, mainly have disconnects on front case ports not rear.
Pcie/gen4 timing bug? I only encounter it on the rig with rtx3xxx. I guess finding this bug in beta would require more gen4 cards for testers lol.
Happy to have a fix, its not a big enough of an issue often enough to roll back to earlier bios.

Zen 2 seemed to have a cleaner launch than Zen 3, stability issues were with those trying to get high IF clocks.

Newest Intel Side-Channel Attack Sniffs Out Sensitive Data

A new side-channel attack takes aim at Intel's CPU ring interconnect in order to glean sensitive data.

threatpost.com

Ring bus timing attack, Appears to have only tested skylake and coffee, Hoping intel can shake these off with Alders lake and on.

 

[freeagent]

I guess you never used x79 or x99... both had significant teething issues.
I will take a USB disconnect over bsod anyday.

Yes indeed, it all started with X58 lol. As much as I love mine, it was probably the most moody, emotional, bi-polar wreck of a PC that I have owned. I love it.. so it sits in a box on a shelf in the dark..

 

[Patriot]

Yes indeed, it all started with X58 lol. As much as I love mine, it was probably the most moody, emotional, bi-polar wreck of a PC that I have owned. I love it.. so it sits in a box on a shelf in the dark..

Oh god, I had forgotten... I got off that platform really fast. Pity those using the SR2 it had a rather unsavory dog flavored nickname....
Went to the 1155 and found it far more stable than 1366 ever was. 2nd gen 1366 boards were pretty alright, and I had a solid 2p 1366 workstation Z600 that was rock solid. But the consumer boards.... not so great experience.

 

[utmode]

February chipset driver update fixed frame dropping issue with my system when I connect through docking stations. Thanks AMD.