AMD patches botched firmware that disabled cores on Ryzen 5 7600X

AlphaX

Posts: 98   +20
Staff
What just happened? Every so often, AMD releases a new update to the AGESA firmware that improves performance, increases stability, adds additional CPU support, and so on for AMD motherboards. Typically, these updates go off without a hitch and work flawlessly. Unfortunately, the latest AGESA software update was anything but flawless.

Last week, AMD released AGESA ComboAM5PI 1.0.0.4, the newest software update for motherboard manufacturers that boasted typical stability upgrades. Most importantly, it also added support for the recently launched Ryzen non-X CPUs and the upcoming X3D processors, set to reach shelves in February.

Unfortunately for AMD, hardware leaker "chi11eddog" revealed that the new software was accidentally disabling one to two cores on specific Ryzen 5 7600X processors. 7600X CPUs with a single core-complex die (CCD) were unhindered by this bug, however dual-CCD processors weren't so lucky.

Users with a dual-CCD processor who updated their BIOS during the brief period when the downloads were available either experienced significant performance loss or, worse, a computer that simply would not boot.

Certain motherboards initialize the boot process strictly through one core, typically the first one, known as Core 0. The biggest issue is that Core 0 was the core that was disabled, which meant the PC wouldn't boot. There were no specific numbers regarding performance loss, but disabling an entire core (or two) is guaranteed to cause a noticeable drop.

As shown above, MSI and Asrock were quick to pull any BIOS updates that featured the new firmware from their respective support pages following the reports. When chi11eddog's tweet went live, Gigabyte hadn't yet removed downloads for BIOS updates with the affected software. However, the links were no longer available later that day.

Thankfully, AMD was quick to send out a patched version of the AGESA ComboAM5PI 1.0.0.4 firmware that does not inadvertently disable any cores. So far, no bad reports have come out regarding the new software, so hopefully the second time's the charm. At the time of this writing, MSI has already released BIOS updates with the fixed firmware. However, Asrock and Gigabyte have currently not done so.

Overall, this was a weird miscue from AMD, and it's rather odd how this bug made it through testing and beta BIOS updates featuring the firmware. The issue wasn't discovered until official BIOS releases went live but thankfully, manufacturers were prompt to ensure users wouldn't potentially download it and suffer any consequences. Kudos to AMD for the quick patch work as well.

Permalink to story.

 
At this point, I think that AMD is trolling the tech community.

I mean, everyone is super quick to write and make negative videos when something bad happens with AMD but never to praise them for something good or even give them fair reviews, so they are doubling down on "bad" things to keep them even busier. :-D
 
Last edited by a moderator:
Something I think should be mentioned, as part of the new AM5 platform, AMD has built into the chipset BIOS flashback. As a result, users will always be able to flash and update the BIOS to their AM5 boards, regardless of the CPUs supported by or the operational status of the current motherboard.

It's unfortunate that this happened but at least you can fix it without having to replace the motherboard or replace a BIOS chip or anything like that.
 
Something I think should be mentioned, as part of the new AM5 platform, AMD has built into the chipset BIOS flashback. As a result, users will always be able to flash and update the BIOS to their AM5 boards, regardless of the CPUs supported by or the operational status of the current motherboard.

It's unfortunate that this happened but at least you can fix it without having to replace the motherboard or replace a BIOS chip or anything like that.
See? I didnt know that!

Maybe I missed it on the barrage of negative cr@p posted in all the reviews or perhaps they simply didnt bother in mentioning it because they couldn't turn it into something negative.

Either way, that is indeed a nice feature to have.
 
Last edited by a moderator:
Something I think should be mentioned, as part of the new AM5 platform, AMD has built into the chipset BIOS flashback. As a result, users will always be able to flash and update the BIOS to their AM5 boards, regardless of the CPUs supported by or the operational status of the current motherboard.

It's unfortunate that this happened but at least you can fix it without having to replace the motherboard or replace a BIOS chip or anything like that.
Having bought a Gigabyte AM4 X570 MB for a build I did for my wife, they have a feature on the MB that allows bios flashing without a CPU. It's called "Q-Flash Plus" https://www.gigabyte.com/Motherboard/X570-AORUS-ELITE-rev-10#kf It looks like they also have this on at least some of their AM5 MBs.
 
Having bought a Gigabyte AM4 X570 MB for a build I did for my wife, they have a feature on the MB that allows bios flashing without a CPU. It's called "Q-Flash Plus" https://www.gigabyte.com/Motherboard/X570-AORUS-ELITE-rev-10#kf It looks like they also have this on at least some of their AM5 MBs.
AMD have added their own version to the AM5 platform, Every single AM5 motherboard has it without failure. I believe motherboard manufacturers can choose not to enable it. Which is fine for Asus, Gigabyte, Asrock etc... As they all have their own versions. It just means the hardware is there to do this regardless of motherboard vender, it's just up to the vender to enable AMD's version.

And you're absolutely right about past motherboards having the same feature, Asus call it BIOS Flashback for example. Very handy feature for AMD motherboards especially, since they actually release CPU's way down the line for the same chipset and therefore, it's basically a requirement to be able to update the BIOS to support newer chips.

I have a vague memory of some older AM4 boards requiring a BIOS update to support the 5000 series and the only way to do that on some boards was to get a supported CPU installed, update the BIOS, then you can install the new CPU, which obviously is a bit of a faff. This gives motherboard manufacturers no excuses this time.
 
I have a vague memory of some older AM4 boards requiring a BIOS update to support the 5000 series and the only way to do that on some boards was to get a supported CPU installed, update the BIOS, then you can install the new CPU, which obviously is a bit of a faff. This gives motherboard manufacturers no excuses this time.
You're right about that. I have three AM4 motherboards. I started with an ASRock X370 Killer SLI and then upgraded to an ASRock X570 Pro4. One day while I was in Canada Computers, I saw (of all things) a Biostar A320MH motherboard on clearance for CA$40. Needless to say, I couldn't say no to that so now I have three boards. :laughing:

To update the BIOS of an ASRock Killer SLI 300-series AM4 motherboard you first need an Athlon X4 900-series, A6/8/10-9000-series or a Ryzen 1000-series CPU (I used my R7-1700). These processors will only be usable up to BIOS v5.40. For a 400-series AM4 motherboard, you need to start with an Athlon 300 or 3000-series or a Ryzen 2000-series CPU. These are also only good until BIOS v5.40. For BIOS v5.50-7.30, you'll need a Ryzen 3000-series CPU.
ASRock X370 Killer SLI BIOS List

Meanwhile, the X570 requires a Ryzen 3000-series CPU to do the update that allows for the Ryzen 5000G-series APUs to work.
ASRock X570 Pro4 BIOS List

The Biostar A320MH v6.x motherboard has similar requirements for BIOS updates as the X370 Killer SLI.
Biostar A320MH BIOS List

BIOS updates went smoothly for all three motherboards. I thought initially that I bricked my X570 Pro4 which confused the hell out of me because I've done close to 100 BIOS updates in my life without ever having a problem. The system wouldn't post, it just sat there at the splash screen. By blind chance, I remembered similar behaviour from my old Gigabyte 990FX motherboard and it turned out to be a faulty SATA cable.

I had to weed through six SATA cables before discovering the offender. I had to unplug them one by one until the system posted. Relieved that my precious X570 board wasn't bricked, I finished the BIOS updates. The next day, I went to Canada Computers and bought a dozen SATA cables for about $36 and just changed all of them. I wasn't about to worry about saving a few dollars after going through that migraine!

I just put that anecdote in there so that if anyone has a similar problem, it might save them from turfing a perfectly good motherboard. SATA cables fail so rarely that most people have never encountered a bad one and would never think of it being the problem. Maybe that means I have really crappy luck because I've encountered it twice now. At the same time, I've been building since 1988 so I guess that it had to happen sooner or later. Before SATA, I'd never seen an RLL or IDE cable fail. :D

Oh well, at least the DIMMs in my skull are still good! :laughing:
 
Last edited:
Back