X670E Ace Bios Problems

BiscuitBomber

Active member
Joined
Oct 19, 2022
Messages
167
Hardware:
Processor: 7950X
Motherboard: MSI X670E Ace
GPU: EVGA 3080 TI FTW3
MEMORY: G.SKILL Trident Z5 Neo RGB Series AMD EXPO 32GB (2 x 16GB) 288-Pin PC RAM DDR5 6000 (PC5 48000)
PSU: EVGA 1000W P6
Cooler: 360 Phanteks AIO MPH

After updating my BIOS today on my PC. When booting up I was getting BIOS Code A6 and the RAM light was staying Amber. This didn't occur on the previous BIOS 7D69v13. I attempted to clear CMOS but same issue. I ended up downgrading my BIOS back to 7D69v13. I opened a ticket with MSI, but just a heads up in case anyone else runs into this. Before 7D69v142 I had zero issues with RAM and bootup problems.


----------------
EDIT Title
Alan J T
 
Last edited by a moderator:
uhm I see in that reddit thread many changed the cpu 9950x3d and solved the problem
A lot of different boards affected, asrock and asus too
I'm thinking about a possible early batches problem
Are you using only 1 disk and the slot beside the ram?
Try another slot

Other things to test that come to my mind:
try a different windows iso
maybe a simply customized one with rufus, with the hardware requirements removed (TPM)
a different windows build? you trying to install 24h2? go back to previous one or viceversa.

But looking at all the people that solved with cpu swap, a hardware cpu problem doesn't seem so unlikely
Hi Michele,
I have 4 SSDs but I tried disconnecting all the disks, I tried moving them from m.2_1 to m.2_4 slot etc but nothing solved until I reinstalled the old 7950x3d
I tried a different ISO of windows as suggested on Reddit but same problem
Yes using 24h2 but I have older images created with macrium reflect and same problem.
 
The weird thing is that with Linux works well until I install video driver. There must be some incompatibility with Windows.
I switched from a 7950X to a 9950X3D, it turns out the 9950X3D likes totally different memory settings even with the same RAM. I ended up resetting bios several times and finally found memory settings that worked with 9950X3D, since then I''ve changed memory and it works like a charm now and the old memory likes the old settings on another board. I haven't reinstalled, but I cleaned Windows a bit.
I'll keep this in mind even considering that all my troubles happens on stock settings too with ram at 3600mt/s

What about vsoc? Is 1.3V and below still considered a good one? If I set the memory to 6000mt/s it boots ok but default vsoc is only 1.220V that is pretty low considered the defaults I had with the 7950X3D
 
Hi Michele,
I have 4 SSDs but I tried disconnecting all the disks, I tried moving them from m.2_1 to m.2_4 slot etc but nothing solved until I reinstalled the old 7950x3d
I tried a different ISO of windows as suggested on Reddit but same problem
Yes using 24h2 but I have older images created with macrium reflect and same problem.

You have already done a lot of testing, I have exhausted my ideas too.
Maybe last thing is switch from AHCI mode to RAID, even though it doesnt make sense if you don't plan to create any array..
Then try to load the raid driver on windows setup. Just out of curiosity.
But 99% will fail, I'm already convinced it's a pcie problem

At this point I think there could be something weird on the pcie lanes of the new cpus
Not sure if hardware or software side (bios/microcode/drivers)
The fact that even on linux you have problems after installing the gpu drivers (gpu is connected to pcie lanes, different from the ones used for storage, but still cpu pcie lanes) is also something that points in the same direction. A coincidence?
Maybe It's just windows that exhibit the problem earlier, in the setup wizard.

What about vsoc? Is 1.3V and below still considered a good one? If I set the memory to 6000mt/s it boots ok but default vsoc is only 1.220V that is pretty low considered the defaults I had with the 7950X3D
Idk how the soc works in the new cpu, but even on 7950x3d 1.3V was already considered pretty high, despite the motherboards pushing very high by default after enabling expo profiles.
6000/6200 typically are ok around 1.2/1.24V with the old cpus, unless you put more stress like using 4 dimms....

You might also want to seek for an updated VBIOS. Most of the problems are on the NV side... Soooo freaking much issues with the 5xxx series I saw only with Fermi....
there are also many problems on 4xxx/3xxx series
mostly caused by recent 572.xx drivers, there's something fishy about nvidia too (and this is also related to pcie, what a coincidence)
For example, lots of people with troubles waking up the monitor from a deep sleep status (monitor sleep, not pc sleep): pc still active and not crashed, but monitor stays black
It also happened to me a couple times: searching in tech forums it's a well know issue of recent drivers, so I disabled monitor sleep in windows and rely on the proximity sensor of my new monitor.... I think it's the best option for me, but not everybody has a proximity sensor available.

Somebody else fixed by forcing the pcie to gen 4 in bios, how curious....
Going back to 566.36 is another way to avoid this weird black screens problem, but I understand maybe it's not a good idea for 5xxx owners. A VBIOS update can mitigate something for sure, but drivers are pretty much broken right now.
Similar black screens events are happening in different situations, on 5xxx series it may happen while gaming too. Very frustrating.
And both in windows and linux environments.
 
Last edited:
I waiting the RMA of the 9950X3D. Hope that the new CPU will work :(

Good luck with the new one, regarding VSOC I ran 1.23V with my 7950X for the longest time using 32GB 6400MT/2133 c30. Now with 9950X3D at 96GB 6000MT/2000 c30 I'm running 1.20V VSOC. 1.3V sounds pretty high imo and too close to burning out the CPU territory from what I've seen.
 
Good luck with the new one, regarding VSOC I ran 1.23V with my 7950X for the longest time using 32GB 6400MT/2133 c30. Now with 9950X3D at 96GB 6000MT/2000 c30 I'm running 1.20V VSOC. 1.3V sounds pretty high imo and too close to burning out the CPU territory from what I've seen.
I said vsoc from 1.3V and below meaning that 1.3V was the highest possible.
On my 7950X3D I needed 1.270V to be rock solid on 4 sticks 4x16GB at 6000/c30
 
Good luck with the new one, regarding VSOC I ran 1.23V with my 7950X for the longest time using 32GB 6400MT/2133 c30. Now with 9950X3D at 96GB 6000MT/2000 c30 I'm running 1.20V VSOC. 1.3V sounds pretty high imo and too close to burning out the CPU territory from what I've seen.
1.3V is perfectly fine. There is no extra risk. I would be more worried about stability at 1.2V as some CPUs can't handle even 6000MHz properly at that voltage.
With my 9800X3D I had 32GB single rank and now I have 64GB dualrank sticks and it wouldn't be stable below 1.24V Vsoc. Some CPUs are like this.
 
But is 9950X3D like 7950X3D?
Can it handle up to 1.3V on VSOC?
Yes, this is why it is locked to a maximum of 1.3V. This is the highest voltage considered safe for 7000 and 9000 series X3D CPUs.
However, it is still a good idea to keep it as low as possible if you want to increase Infinity Fabric speed. Higher vSoC can reduce stability when pushing Infinity Fabric beyond 2000 MHz. Additionally, there is a negligible power-saving benefit, as the CPU consumes slightly less wattage at lower vSoC, but this effect is insignificant.
 
Still waiting the new 9950X3D from RMA hoping that it will solve my problem.
In the mean time I wrote to MSI asking if there are some known incompatibilities between my components and the ACE
 
well. if you are waiting for rma push that soc to 1.3V
there's a limit of 1.3V soc, and the same applies for 9950x3d too, it's perfectly fine to use that voltage
If it catches fire, you have not done anything wrong (and you have another good reason to rma the cpu that was defective on arrival)

if we see lot of people running low soc on new cpu it's because maybe there are more lucky IMCs around statistically
a lucky IMC on a 7950x3d could theoretically run stable 6000mhz at 1.1/1.15V soc as well, agesa version can also influence the lowest soc you can run without issues.
but that doesn't mean you should run at that low voltage
Btw we are talking about 1:1 configs here, cause when you switch to 2:1 you are in fact reducing the stress on IMC (unless you are in 8000 territory) and that changes the soc requirement
 
Last edited:
well. if you are waiting for rma push that soc to 1.3V
there's a limit of 1.3V soc, and the same applies for 9950x3d too, it's perfectly fine to use that voltage
If it catches fire, you have not done anything wrong (and you have another good reason to rma the cpu that was defective on arrival)

if we see lot of people running low soc on new cpu it's because maybe there are more lucky IMCs around statistically
a lucky IMC on a 7950x3d could theoretically run stable 6000mhz at 1.1/1.15V soc as well, agesa version can also influence the lowest soc you can run without issues.
but that doesn't mean you should run at that low voltage
Btw we are talking about 1:1 configs here, cause when you switch to 2:1 you are in fact reducing the stress on IMC (unless you are in 8000 territory) and that changes the soc requirement
do I need 1.3V to run 3600mt/s?
current 9950X3D does not work even at 3600mt/s
 
No, but...

SOC voltage is useful to stabilize some internal cpu components, including pcie lanes (where gpu/nvme are typically connected)
It's not only for the memory

Go for it, set ram to more classical 4800/5200mhz, push 1.3V on soc and see if it changes anything if you are curious
You are not doing anything out of spec
It makes sense to test high soc voltage for this reason, but even it works at 1.3V SOC it's still a cpu that I wouldn't keep. (with 1.22V soc it should boot at least)

While I'm editing the msg, I was thinking something else to ask:
do you get the prompt to reset fTPM after changing the cpu? Isn't the reset mandatory when you change the cpu?

EDIT;
have you noticed the beta 1.2.0.3a bioses are getting deleted from all the msi support pages?
For the ACE L4 beta disappeared
same for all other mb
😨
 
Last edited:
No, but...

SOC voltage is useful to stabilize some internal cpu components, including pcie lanes (where gpu/nvme are typically connected)
It's not only for the memory

Go for it, set ram to more classical 4800/5200mhz, push 1.3V on soc and see if it changes anything if you are curious
You are not doing anything out of spec
It makes sense to test high soc voltage for this reason, but even it works at 1.3V SOC it's still a cpu that I wouldn't keep. (with 1.22V soc it should boot at least)

While I'm editing the msg, I was thinking something else to ask:
do you get the prompt to reset fTPM after changing the cpu? Isn't the reset mandatory when you change the cpu?

EDIT;
have you noticed the beta 1.2.0.3a bioses are getting deleted from all the msi support pages?
For the ACE L4 beta disappeared
same for all other mb
😨
Isn't that normal, to delete the beta versions when a GA release is made available?
 
Normally the betas are still listed on the support side... weird....
Ehh... just for a better feeling and to prevent my brain from creating weird conspiracy theories I gotta update to 1L I think....
 
Normally the betas are still listed on the support side... weird....
Ehh... just for a better feeling and to prevent my brain from creating weird conspiracy theories I gotta update to 1L I think....
They are doing the same thing for all the models it’s even weirder
1.2.0.3 so broken
But 1L seems fine apparently for me. Played a lot, even remotely with Apollo/moonlight (that also stress the nic, cause I wasnt sure about the nic reliability right after the flash)
Not a single problem so far
 
Last edited:
Also on 1L. No problems so far, but also not able to reach anything above 5200 rock stable. 5400 POSTs but with problems. Anything above -- dreaded "Ab" on the display after training ("15").
 
128GB is still hard for this processors
Always remember this:
the IMC is pretty much the same, but binning is slightly better with the new series (more refined production process probabily)

7xxx9xxx
2x1R
DDR5-5200
2x2R
DDR5-5200
4x1R
DDR5-3600
4x2R
DDR5-3600
2x1R
DDR5-5600
2x2R
DDR5-5600
4x1R
DDR5-3600
4x2R
DDR5-3600
 
Back
Top