7800X3D will be an utter failure of a CPU

Looks like ASRock was allowing X3D chips to go up to 1.4v (actual CPU voltage, not SoC/DRAM) too with their latest non-beta bios. They have fixed it in their latest beta BIOS. Apparently this was happening even when sitting in the BIOS menu or with idle windows activity, not just at load. That's the same BIOS that wouldn't POST with my Steel Legend. Wonder if it is because the CPU or MB was already detecting problems being caused by overvoltage.

It appears all four major MB maker's latest "stable" BIOS versions on day 1 release of the 7800X3D, several weeks after the 7900/7950X3D release, could potentially overvolt and kill any of the X3D chips. Sounds like AMD either didn't properly inform the MB makers of what needed to be changed and/or rushed out the X3D chips without giving MB makers time to correctly configure BIOS versions. Not looking good for AMD...

Legendary Gamer said:
MSI is also limiting their core voltages to the 7000 X3D series due to chips burning up on their motherboards:

https://videocardz.com/newz/msis-new-bios-restricts-voltage-for-amd-ryzen-7000x3d-cpus
Click to expand...
Gigabyte also deleted all their old BIOS versions, like ASUS did. For the same reasons.
 
Legendary Gamer said:
MSI is also limiting their core voltages to the 7000 X3D series due to chips burning up on their motherboards:

https://videocardz.com/newz/msis-new-bios-restricts-voltage-for-amd-ryzen-7000x3d-cpus
Click to expand...
I didn't think that there were any burnt CPUs on MSI boards? It says in the article "MSI has a dedicated Reddit post to this problem, although no trouble with MSI boards has been reported thus far."

That said, I hadn't planned on upgrading to the beta bios that were released on the 14th because my build has been very stable, but after the update that said it limits the voltage, I decided to go for it.
 
1682425431982.png


Looks like a new bios will roll out today limiting SOC voltage to 1.3V on X3D chips.
 
izzypizzy said:
I didn't think that there were any burnt CPUs on MSI boards? It says in the article "MSI has a dedicated Reddit post to this problem, although no trouble with MSI boards has been reported thus far."
Click to expand...
On Reddit there has been at least one claim of a killed CPU on an MSI board. True? Who knows.
funkydmunky said:
Early adopter on a new arch isn't a great idea IMO. I've done it too many times. Let it settle in, bugs fixed and prices to drop a bit.
Click to expand...
This is now confimed affecting the whole AM5 platform, not just X3D. We are seven months into launch. How much time has to pass for a platform to not be considered "early adoption" anymore lol. But I do agree with the sentiment.

EDIT:
https://www.tomshardware.com/news/a...use-identified-expo-and-soc-voltages-to-blame

All MB makers, including even Biostar, have confirmed failures now.
 
LittleBuddy said:
Also I'll add one of my observations from my board. On my board the SOC defaulted to 1.350V (BIOS 1410 beta) which shouldn't burn the chip out since it's still in safe range. But it is still very high for a default SOC voltage, so anyone setting up their board you might want to dial back the SOC voltage to 1.250V which should run DDR5-6000 (2000 FCLK) kits no problem, you could even go lower on SOC.
Click to expand...

1usmus is recommending 1.15V-1.2V SOC voltage for DDR5-6200. So yea my recommendation of 1.250V SOC Voltage may even be too high for most rigs. I'm going to test 1.15V overnight and see how well it holds up. I trust 1usmus and would follow his recommendations. He also recommends <1.35V on the memory controller.
1682429266458.png

https://twitter.com/1usmus/status/1650810800745922564
 
LittleBuddy said:
1usmus is recommending 1.15V-1.2V SOC voltage for DDR5-6200. So yea my recommendation of 1.250V SOC Voltage may even be too high for most rigs. I'm going to test 1.15V overnight and see how well it holds up. I trust 1usmus and would follow his recommendations. He also recommends <1.35V on the memory controller.
View attachment 566374
https://twitter.com/1usmus/status/1650810800745922564
Click to expand...
Yep, heard the same elsewhere so already moved my SoC down to 1.2. However, for the DRAM voltage I could not boot with 1.3v, so staying at 1.35 for now. Maybe something in-between is possible but have not tried yet.
 
(eco mode - 65watt) Expo I: SOC = 1.288 / mem 5600 / CBR23 = 23814
(eco mode - 65watt) No Expo, set to Auto: SOC = 1.048 / mem 4800 / CBR23 = 25186
 
I just got my friend’s 7800x3D in and installed it. Working on getting windows on the drive then I’ll update the bios. It’s nice that it has built in gpu so I don’t have to put one just yet since the case hasn’t arrived, but I’m using the good old box method and I felt the need to add a fan because this b650 chipset gets very hot with no airflow.
 
1.15V SOC seems good on my PC (2000 FCLK) so far, been running Prime95 large FTT and using my PC while it's going for about 3 hours now, no WHEA errors and no errors in Prime95. I'll probably let it run overnight. BTW going down from 1.25V to 1.15V the CPU uses about 10% less power as well. Went from package peaking at 92W down to 85W and temps came down a bit too.
 
LittleBuddy said:
1.15V SOC seems good on my PC (2000 FCLK) so far, been running Prime95 large FTT and using my PC while it's going for about 3 hours now, no WHEA errors and no errors in Prime95. I'll probably let it run overnight. BTW going down from 1.25V to 1.15V the CPU uses about 10% less power as well. Went from package peaking at 92W down to 85W and temps came down a bit too.
Click to expand...

I’m running this on an MSI B650-P and my soc voltage was at 1.364 with expo on and a 6000mhz kit. I just tried 1.150 like you did, it boots but I’ll wait until I get the case to stress it.
 
LigTasm said:
I’m running this on an MSI B650-P and my soc voltage was at 1.364 with expo on and a 6000mhz kit. I just tried 1.150 like you did, it boots but I’ll wait until I get the case to stress it.
Click to expand...

Wonder why MSI sets the SoC voltage so high for 6000MHz. My Gigabyte B650 sets it to 1.245v and since that's already below the new cap that Asus will set at 1.30v I guess I shouldn't be expecting a new BIOS from Gigabyte since it seems to be running safe already.
 
MistaSparkul said:
Wonder why MSI sets the SoC voltage so high for 6000MHz. My Gigabyte B650 sets it to 1.245v and since that's already below the new cap that Asus will set at 1.30v I guess I shouldn't be expecting a new BIOS from Gigabyte since it seems to be running safe already.
Click to expand...
my guess would be that they were trying to one up each other by pushing it harder to get the best perf. in comparisons.
 
Man, I haven't opened my 7800X3D yet and I may end up just sending it back. Some users are having it die on first boot without even turning EXPO on.

This has got to be a hardware defect on some 7000 series chips.
 
therealjustin said:
Man, I haven't opened my 7800X3D yet and I may end up just sending it back. Some users are having it die on first boot without even turning EXPO on.

This has got to be a hardware defect on some 7000 series chips.
Click to expand...
Wish all this had come out a little sooner. Would have just returned MB/CPU/RAM and waited until the 13700k went back on a good sale again. EDIT: Also questioning if XMP/EXPO is even worth it regardless of the platform. Seems to just be asking for stability/compatibility problems, and in my testing I haven't found any measurable difference in CPU-bound gaming performance by changing timings.

pendragon1 said:
dram voltage and the cpu mc are different things, are they not?
Click to expand...
Well when I change my DRAM voltage on my MSI board (with latest bios) it automatically changes the DRAM VDDQ and CPU VDDIO to match. But I just reduced CPU VDDIO down to 1.3 to see what happens. So far it at least booted. Also reduced the DRAM VDD/VDDQ down to 1.34 because I noticed they were spiking into 1.36 above the manual 1.35 setting.
 
Last edited:
pendragon1 said:
my guess would be that they were trying to one up each other by pushing it harder to get the best perf. in comparisons.
Click to expand...

possibly also possible that the limits AMD gave based on engineering samples ended up not being the same as production SOC's. who knows, there's a lot of variables involved. i just chalk it up as shit happens, at least it's a known issue now so AMD and motherboard manufactures have to replace them.
 
therealjustin said:
Man, I haven't opened my 7800X3D yet and I may end up just sending it back. Some users are having it die on first boot without even turning EXPO on.

This has got to be a hardware defect on some 7000 series chips.
Click to expand...

Just saw a tweet from Buildzoid about that. Now I'm even more sure that EXPO/Voltages have nothing to do with this. Zen 4 has been out way too long for this issue to just suddenly come out of nowhere, even X3D parts have been around for 2 months now.
 
therealjustin said:
Man, I haven't opened my 7800X3D yet and I may end up just sending it back. Some users are having it die on first boot without even turning EXPO on.

This has got to be a hardware defect on some 7000 series chips.
Click to expand...
Do you know if those people had updated the bios before trying to boot with the X3D chip?
 
MistaSparkul said:
Just saw a tweet from Buildzoid about that. Now I'm even more sure that EXPO/Voltages have nothing to do with this. Zen 4 has been out way too long for this issue to just suddenly come out of nowhere, even X3D parts have been around for 2 months now.
Click to expand...
As much as I like AMD generally they need to be quicker to the press and a little more transparent with mistakes. The 7900XTX heatpipe issue was slow and dragged out. Needlessly. Had they just clearly announced that they had discovered a problem, would take care of affected peeps and fix the issue asap, they would have looked golden. I really hope, but have my doubts, that they will get to the bottom of this CPU issue fast and assure purchasers that they will be taken care of.
 
funkydmunky said:
As much as I like AMD generally they need to be quicker to the press and a little more transparent with mistakes. The 7900XTX heatpipe issue was slow and dragged out. Needlessly. Had they just clearly announced that they had discovered a problem, would take care of affected peeps and fix the issue asap, they would have looked golden. I really hope, but have my doubts, that they will get to the bottom of this CPU issue fast and assure purchasers that they will be taken care of.
Click to expand...

Well this time around they have already acknowledged that there is indeed a problem, with the 7900 XTX it took them like 2 weeks just to even say anything. But it's going to take some time before they nail down the exact root cause. Right now everyone is seriously just guessing what the problem is, nobody really knows. I'm sure they will take care of those who are affected by this.
 
MistaSparkul said:
Well this time around they have already acknowledged that there is indeed a problem, with the 7900 XTX it took them like 2 weeks just to even say anything. But it's going to take some time before they nail down the exact root cause. Right now everyone is seriously just guessing what the problem is, nobody really knows. I'm sure they will take care of those who are affected by this.
Click to expand...
Well I hope there were corporate lessons learned from the 7900XTX fiasco going forward then.
 
EXPO pushed my "SOC" voltage to about 1.35v with a 5600 / cl28 kit (on Gigabyte B650m DS3H) which seems unnecessary high actually since it's not even 6000mhz and only 2 x 16GB sticks. I turned EXPO off (and now voltages are really low, SOC at 1.05 etc), which is up to ~10% loss in some CPU bound games even with a 7800 X3D (not nice) - but rather play it safe and let the dust settle for the time being. I have enough games where it makes no difference to keep myself busy.

There is no actual evidence that voltage numbers have anything to do with the failures though. There is even one case of the CPU blowing up with EXPO off... There is just evidence that the failure rate for AMD 7000 CPUs is very alarming and most of those failures look similar with a "bulge" and what looks like extreme heat markings. And X3D chips seem to die in greater number than the other 7000 CPUs that have been around for a while now.

So maybe it's just a pure random manufacturing defect or a horrifyingly bad AGESA bug. Either way this stinks. CPUs are usually that one component you can trust because failures are so damn rare.
And if it's really too high vollages it's not better news, since that would mean the functioning chips have been silently degrading in the meantime.
 
Last edited:
It appears Max SOC is 1.3v and +1.4v was enabled by EXPO and/or extreme overclocking profiles. Asus is limiting SOC to 1.3v now.


OG CROSSHAIR X670E EXTREME BIOS 1302


Version 1302 Beta Version

9.54 MB 2023/04/25

SoC voltage for Ryzen 7000X3D series limited to a maximum of 1.30V to protect the CPU and motherboard.
 
Last edited:
Apparently AMD is bundling Jedi Survivor with 7800X3D purchases, if the store you purchased from offers this promotion, you should contact them for a key. I just chatted with Microcenter and provided them my order number and had the key in my inbox within a few minutes.
 
Mega6 said:
It appears Max SOC is 1.3v and +1.4v was enabled by EXPO and/or extreme overclocking profiles. Asus is limiting SOC to 1.3v now.


OG CROSSHAIR X670E EXTREME BIOS 1302


Version 1302 Beta Version

9.54 MB 2023/04/25

SoC voltage for Ryzen 7000X3D series limited to a maximum of 1.30V to protect the CPU and motherboard.
Click to expand...

On my Gigabyte B650 Aorus Elite AX Rev. F5a, the SoC voltage was set to 1.245v for 6000 CL30 EXPO. Verified this with both HWMonitor and HWInfo64. I'm guessing there's no need to update the BIOS then since I'm already running below the 1.3v cap?
 
kalston said:
EXPO pushed my "SOC" voltage to about 1.35v with a 5600 / cl28 kit (on Gigabyte B650m DS3H) which seems unnecessary high actually since it's not even 6000mhz and only 2 x 16GB sticks. I turned EXPO off (and now voltages are really low, SOC at 1.05 etc), which is up to ~10% loss in some CPU bound games even with a 7800 X3D (not nice) - but rather play it safe and let the dust settle for the time being. I have enough games where it makes no difference to keep myself busy.

There is no actual evidence that voltage numbers have anything to do with the failures though. There is even one case of the CPU blowing up with EXPO off... There is just evidence that the failure rate for AMD 7000 CPUs is very alarming and most of those failures look similar with a "bulge" and what looks like extreme heat markings. And X3D chips seem to die in greater number than the other 7000 CPUs that have been around for a while now.

So maybe it's just a pure random manufacturing defect or a horrifyingly bad AGESA bug. Either way this stinks. CPUs are usually that one component you can trust because failures are so damn rare.
And if it's really too high vollages it's not better news, since that would mean the functioning chips have been silently degrading in the meantime.
Click to expand...

Are you on Rev F5a? That is the latest BIOS which I am on and my B650 Aorus Elite AX only sets SoC voltage to 1.245v for 6000 CL30 EXPO.
 
They removed every BIOS from the support page of my mobo... except the one I'm using since day one which is this one:
1682530361885.png


The version name changes depending on the motherboard model so I guess it's best to look at the AGESA number if comparing. But yes I'm reading (hwinfo and BIOS) over 1.3v SOC with this BIOS when EXPO is on.
 
Last edited:
You must log in or register to reply here.
Back
Top