i came across this video and related article on the Gamers Nexus website.
There seem to be a few reported problems with a certain batch of 9800X3D cpu´s failing,
to be initialized on Asrock and other brand boards after a indefinite amount of time.
I don´t particularly think that this is a Asrock issue.
Because people with other brand boards have also experienced it but in lower quantities.
I guess it might be a certain batch of cpu´s or maybe an AGESA issue.
Because i have not read anything about other 9000 x3d series cpu´s specifically.
Also people reported that they got their system working again after downgrading the bios to v3.10.
Which lead me to think that this might be a AGESA related issue.
Asrock currently has bios v3.20 released which should resolve the issue.
However for the X870E Taichi boards specifically that bios version is still listed as Beta.
That might be an error on their support page but (be cautious anyways).
So it´s probably not the cpu´s that are actually failing here.
I based that on the amount of cpu + board combo´s that are sold.
And how low failure report rate currently is unlike the 7800X3D and Asus debacle.
Yeah, same here. If the 9900X3D and 9950X3D are also affected it’s still early for reports of similar failures to appear. Since tech stuff usually launches in beta I figure at least 3-6 months for most problem discovery and fixes.
ASRock appears to be the most common AM5 choice, followed by Asus. So, statistically speaking, ASRock and then Asus having more reports from a more or less constant AGESA failure rate seems plausible. Small n, so hard to tell.
I’m not in a hurry to build 9950X3D. Probably doesn’t matter much for 9900X3D right now as it’s priced high relative to 9950X.
Generally ASRock updates Taichi first, so probably they’ll be the first to exit beta. The beta’s also not fully propagated down stack yet.
Spot checking a few other mobos (X870E Carbon, ProArt, AI Top), if Asus, MSI, and Gigabyte have fix candidates they don’t seem to be public yet.
I helped a couple people with weird 9800x3d problems. planning to swap their cups for new ones to see if I can get a dodgy one but so far no catastrophic failures.
and the one dodgy 9800x3d cpu I swapped for has been perfectly stable on every board I have and the CPU I sent back has been stable for the other person on all their existing gear whereas the one I got from them was unstable on their board.
I guess we have to wait and see how it evolves.
If it is a AGESA or bios related issue a update will likely fix it.
Asrock already released a new bios that should address the problem.
I have not read about other manufacturers yet but the report rates on those seem to be lower.
There are a few reports of actual dead cpu´s but that is likely due to a different cause,
like improper installation.
So far no issues, but i’m scared. Not sure if I should check the socket for burn marks or not touch it, since it works…
Wouldnt be surpriced, if its bad socket batch.
Like with first gen threadripper (which i was “lucky” to have), where the TR socket from foxconn had screw, which was too short and you basically had to bend the socket or use like 70KG of force to catch the thread…
I would not be worried too much about it.
Asrock already released a bios update v3.20 which should address the boot issues,
with a certain batch of cpu´s.
I have a ASRock board and as of a week ago restarts take forever and sometimes after long stretches of use it becomes laggy. Mainly apps take forever to open and sometimes apps are extremely slow to operate. On a reboot it is fine. I do not know if it is the board, or ram or SSD. It just has been having issues in the last week to two. I have a 7950x. ASRock x670 steel legend. Also seems to memory train a lot. So restarts take upward of 15minutes.
I have noticed on a daily basis multiple CPUs both older and newer having stability issues with ASRock. I have had my system since 2022.
We’ve had Windows 11 updates provoke much the same. Seems linked to the OS forgetting how to talk to drives. Reimaging has helped, albeit sometimes only in simplifying diagnosis.
7950X we’ve got on X670E Pro RS is fine but it’s on 10. It’s a quad DIMM overclock and trains in a few seconds too. So might be worth a look to see if your memory config’s gone unstable.
This is kinda interesting there seems to be a little bit more to the story.
Voltage spikes on the soc could be a plausible cause since we have seen this before.
Yeah. Just took a look at the 9900X and 9950X I have on 3.20 and Vsoc (at least as reported by HWiNFO) only varies between 1.185 and 1.190 V. First thing I did after the 9950X posted was flash 3.20 but I don’t recall seeing Vsoc > 1.2 V with the 9900X before 3.20 released.
The BIOS update referred to here is v3.25 or 3.26, depending on which board I’ve checked, which ASRock notes also contains an update to AGESA 1.2.0.3d. That might hint at an AGESA issue under ASRock’s PBO error.
ASRock says not. The article’s pulling from GN’s linked interview with Chris Lee (ASRock VP, motherboards) and gets all the major stuff. I went through GN’s segment and, well, seems like ASRock’s focused doing an adequate job rather than a good one.
This is kinda huge i wanted to buy a Asrock board with a 9700X or 9800X3D.
But i´m going to wait this whole situation out for a bit.
From what i read in the article there seem some reports of dying cpu´s,
even with the 3.25 bios.
However that is another thing that needs to be investigated.
At least Asrock does kinda takes responsibility however they could have been a bit more transparent about it.
Thanks for adding this additional information.
@wendell did you experience any issues on this Asrock PBO debacle?
Yeah. Though MSI, Asus, and Gigabyte also have failures and have communicated zero so far as I know. IMO Chris’s responses weren’t unreasonable but less upper management wafflespeak would have been nice.
FWIW, since the database thread hasn’t updated in a while I went through some of the recent claimed deaths on r/ASRock. Didn’t get through that many before reddit started showing the same threads repeatedly but it doesn’t seem out of the question 3.25 could get ASRock fairly close to background failure rate. Not seeing complaints of RMA hassles by either ASRock or AMD which, reddit being reddit, seems notable.
Yes. At minimum the background failure rate’ll continue. Chris said degradation isn’t a concern but, given the varying times to failure and stated failure causes, I’d rather expect the opposite where PBO damage pre-3.25 may lead to failure after update to 3.25.
Same for one of the Granite Ridge builds I have. Haven’t gotten the other one yet, though it shouldn’t matter as neither’s PBO. If I needed to do a third I’d be fine using ASRock again but I build dual CCD and duals seems to be at background failure rate.
Be nice if ASRock’d get their load saved USB profile to work across BIOS versions so you didn’t have to re-enter all the settings.
