AMD crowd-sources evidence for the fast-emerging 500-series chipset USB issue

The Threadrippers are basically nonexistent, I am wondering if they are going to delay their Threadripper/EPIC launches to get a better handle on their existing parts or if they are going to make the supply of the existing stuff worse by launching the TR & EP's because the margins on those parts are much better... I need a new TR or EPYC system to replace an aging RDP server for the digital design students.

Newegg has threadripper pro in stock. If that is what you are asking about. I don't know if they are the new ones out, but newegg has them as a Feb 2021 release.
 
Newegg has threadripper pro in stock. If that is what you are asking about. I don't know if they are the new ones out, but newegg has them as a Feb 2021 release.
Well, that wasn't there when I last checked I'll have to give CDW a go as New Egg's billing really pisses off accounting, but it's even close to MSRP.
 
When you buy bleeding edge hardware, you accept the risk that it may have teething problems of bugs or other undesirable behavior. Throwing out a whole system based on a USB issue and "going back to intel" like they're perfect is just pathetic. USB since its inception in the late 90s has had wide ranging issues with implementation and bugs. Problems with USB have historically centered around standards changes, like going from 1.0 to 1.1, then 2.0 and 3.0. Now we have a quagmire of successive USB standards all stacked on top of each other with 3.1, 3.2 and all of the stupid connector changes. I'm surprised USB is working at all on new devices.

How about a real serious problem, like when Intel had self-destructing SATA ports on X58 and some early Sandy Bridge chipsets? Or their ongoing snafu of severe processor vulnerabilities that cover basically their entire product line and the patches that reduce performance by as much as 50%.



This is fucking amazing. You swear off an entire company based on the TDP of a single processor when comparable Intel processors of the time were far more power hungry. The Athlon 64 x2 4400+ was either a 89W or 110W part (depending on the die revision), when comparable Intel processors released at the same time, like the Pentium 4 670 was a 115W part. Or the Pentium D 830/840, which were 130W parts that often drew closer to 160W. Even the top end Athlon 64 4800+ was only a 110W part.



I'd hazard a guess you're an Intel/Nvidia fanboy? There's really no other way of looking at it with all of the hate you're throwing at AMD, and saying an entire company is bad based on your wildly inaccurate assumption about a single processor from 16 years ago.
Yeah Intel had a big problem with one of their recent 2.5Gb LAN chipsets which shipped on a bunch of motherboards. Notable packet loss issues. I dunno if they ever fixed it. I do know they released a revised chipset before they fixed the old one, LoL.
New motherboards which released in the midst of it, went with Realtek instead.
 
Yeah Intel had a big problem with one of their recent 2.5Gb LAN chipsets which shipped on a bunch of motherboards. Notable packet loss issues. I dunno if they ever fixed it. I do know they released a revised chipset before they fixed the old one, LoL.
New motherboards which released in the midst of it, went with Realtek instead.

Sort of like the shit Atheros "Killer LAN" used on gaming boards that eventually end up committing suicide and stop working properly. I've had to get PCIe NICs for several boards because of those garbage things. Thank god motherboard vendors stopped using those shitty things and went back to Realtek, Intel or even Broadcom.

The title is probably misleading. It is not necessarily a chipset issue. It could be a bios ir microcode issue. There is no evidence to believe the chipset is defective.

I would say it's probably firmware and/or driver related. I hope AMD irons out the bugs eventually.
 
Yeah Intel had a big problem with one of their recent 2.5Gb LAN chipsets which shipped on a bunch of motherboards. Notable packet loss issues. I dunno if they ever fixed it. I do know they released a revised chipset before they fixed the old one, LoL.
New motherboards which released in the midst of it, went with Realtek instead.
They “fixed” it, there have been software updates that make it less terrible, they recalled the remaining chips but it’s up to the MB manufacturers to recall their boards. They won’t recall the boards.
 
Still trying to get anyone having the issue to try manually setting base clk to 100 and see if it helps. In the past whenever I tried base clk to push megahertz a bit more the symptom was flaky usb functionality. Mainly curious as I really doubt the clk would drift that much but stranger things have happened.
yeah when they enabled bclk control on asus tuf x570 if you even bumped it up 1mhz the whole system would flake out and take literally 5 min to boot. so I could def see that being a problem.
 
So, for those of you with mobos/hardware which suffered from this USB dropout issue: did the bios update fix it for you?
 
So, for those of you with mobos/hardware which suffered from this USB dropout issue: did the bios update fix it for you?
I never had it because I couldn’t find new hardware. But from the lack of new complaints in most of the groups it seems to have overall fixed many issues.
 
When you buy bleeding edge hardware, you accept the risk that it may have teething problems of bugs or other undesirable behavior. Throwing out a whole system based on a USB issue and "going back to intel" like they're perfect is just pathetic. USB since its inception in the late 90s has had wide ranging issues with implementation and bugs. Problems with USB have historically centered around standards changes, like going from 1.0 to 1.1, then 2.0 and 3.0. Now we have a quagmire of successive USB standards all stacked on top of each other with 3.1, 3.2 and all of the stupid connector changes. I'm surprised USB is working at all on new devices.

How about a real serious problem, like when Intel had self-destructing SATA ports on X58 and some early Sandy Bridge chipsets? Or their ongoing snafu of severe processor vulnerabilities that cover basically their entire product line and the patches that reduce performance by as much as 50%.



This is fucking amazing. You swear off an entire company based on the TDP of a single processor when comparable Intel processors of the time were far more power hungry. The Athlon 64 x2 4400+ was either a 89W or 110W part (depending on the die revision), when comparable Intel processors released at the same time, like the Pentium 4 670 was a 115W part. Or the Pentium D 830/840, which were 130W parts that often drew closer to 160W. Even the top end Athlon 64 4800+ was only a 110W part.



I'd hazard a guess you're an Intel/Nvidia fanboy? There's really no other way of looking at it with all of the hate you're throwing at AMD, and saying an entire company is bad based on your wildly inaccurate assumption about a single processor from 16 years ago.

Quagmire?
263y6a.jpg


Check for "zombie" USB devices. I've had USB devices go into a corrupt state where they take down the entire USB subsystem, or cause erratic behavior like slow speeds or new USB devices not enumerating properly. This tends to happen with USB devices that remain plugged in for extended periods of time, especially storage devices.

I can't tell you how many "emergency" calls I've gotten with malfunctioning or slow computers from customers that turned out to be a flash drive stuck in the back of their system that they had forgotten about months ago. The flash drive will eventually go into a state of funk and cause the entire computer to go haywire.

Also check that the port isn't fouled up, ESPECIALLY on newer USB 3.x ports like the shitty USB-C connector. The contacts eventually get tarnished or dirt packed into them and intermittent connectivity issues. I've seen USB 3.x devices fall back to 2.0 compatibility mode due to bad connections and be really slow. Even the contacts on the device itself can get fouled up. I keep a can of Deoxit Gold G5 on hand for tarnished connectors and it does amazing work restoring electronic gear with worn/dirty/tarnished contacts.

This happend to me also! I think they should have designed a reversible USB connector from the get go! I never broke a type-C yet but I do have to blow it (that's what he said LoL) out with compressed air occasionally
I use the De-Oxit to keep my good old audio gear in working order this stuff is 40 years or more older and is built and sounds better then the modern "Walmart" grade JUNK I do have a modern AVR with 7.2 surround sound it sounds OK for the price I paid but it is not good with music

Never had any USB issues with my Gigabyte x570 master.

Me neither!
 
Sort of like the shit Atheros "Killer LAN" used on gaming boards that eventually end up committing suicide and stop working properly. I've had to get PCIe NICs for several boards because of those garbage things. Thank god motherboard vendors stopped using those shitty things and went back to Realtek, Intel or even Broadcom.



I would say it's probably firmware and/or driver related. I hope AMD irons out the bugs eventually.
Yeah I've had to stick NICs in several Fatal1ty mobos over the years. Remember the $250 Bigfoot M1? LMAO! iirc, Intel owns them now.
 
Yeah I've had to stick NICs in several Fatal1ty mobos over the years. Remember the $250 Bigfoot M1? LMAO! iirc, Intel owns them now.

Yeah, it was a big deal when it came out, there was a lot of hype over it, but it quickly fizzled out.

The price killed it, but so did the form factor. It was first released in 2006, three years into PCIe and it was a bog standard 32 bit PCI card. This means that there was no chance of it ever being able to saturate a gigabit connection link because it shared bandwidth with the rest of the PCI bus. Even if it was the only device on the bus, the overhead wouldn't allow it to perform at 100%.

It was a dual keyed card, which could allow it to work in a 66 MHz PCI or PCI-X slot, but those were virtually non-existent outside of workstation and server boards. But I'm not sure if it'd be able to take advantage of the faster bus speed. I have one LGA775 ASUS workstation board with a PCI-X slot that it'd fit in, but I have more interesting things to put in that slot, like RAID controllers.
 
Back
Top