MPG X870E CARBON WIFI Beta BIOS

[Svet]

Spoiler: Older Versions

>>> E7E49AMSI.1A1E <<<

>>> E7E49AMSI.1A1H <<<

>>> E7E49AMSI.1A21 <<<

>>> E7E49AMSI.1A22 <<<

>>> E7E49AMSI.1A24 <<<

>>> E7E49AMSI.1A25 <<<

>>> E7E49AMSI.1A26 <<<

>>> E7E49AMSI.1A28 <<<

>>> E7E49AMSI.1A29 <<<

>>> E7E49AMSI.1A31 <<<

>>> E7E49AMSI.1A32 <<<

>>> E7E49AMSI.1A33 <<<

>>> E7E49AMSI.1A34 <<<

>>> E7E49AMSI.1A41 <<<

>>> E7E49AMSI.1A44 <<<

Notes:

• * Unstable RAM OC
• * Hang 0D when saving BIOS settings or doing M-Flash

>>> E7E49AMSI.1A62 <<<

>>> E7E49AMSI.1A65 <<<

>>> E7E49AMSI.1A66 <<<

>>> E7E49AMSI.1A93 <<<

>>> E7E49AMSI.1A99 <<<

.rar SHA-256: 0da3ebedfa9e1f53f3e3aa320c1fd32f770cae0b32f112fe87750f1f88031487
BIOS SHA-256: 07a227bb302e9eac43bca0e18a43cfe58e043302eebfb13b08c7b8653a8ad636

---

>>> E7E49AMSI.1AA1 <<<

.rar SHA-256: cf63c3ad59580532257612b108db651ec3341a517c22efb20127dd44c94d8ae1
BIOS SHA-256: 4BD52D73DE3112768E2C6C62CA85C423F0963359A4C4B1CFCFEF1C04294DE1D2

AGESA ver:

---

 

[Indabox]

Also.. the new MSI MPG x870e Carbon "MAX" was released and guess what's different from the existing Carbon? 64mb BIOS size support and the inclusion of a dedicated clock generator chip... That's it.

Maybe the real reason for the addition of the clock generator chip is to stabilize the board because they found issues they could not fix... Who knows..

 

[Leefort]

Also.. the new MSI MPG x870e Carbon "MAX" was released and guess what's different from the existing Carbon? 64mb BIOS size support and the inclusion of a dedicated clock generator chip... That's it.

Maybe the real reason for the addition of the clock generator chip is to stabilize the board because they found issues they could not fix... Who knows..

This situation is beyond crazy now. I went for the new CPU just because I'd already bought 3 Carbons and a new ram kit. Just before that I searched thru asus forums and found out that their boards are also affected but the issue is way less common.

I guess there really are some weird hardware combos that can cause the issue.

Just hope Intel will still exist when I decide to upgrade my setup. Really tired of this bs

 

[kekolopez55154a02d1]

No, I have the Carbon WiFi and my bus has been locked at 100 and never changed to anything else. If you don’t have issues, don’t touch it. I’m pretty much convinced by now that some Carbon WiFi motherboards just have bad clock generator components that don’t hold the frequency as they should, while others have excellent ones. This varies by hardware revision and may be random, not really traceable from release batch to release batch, and the only one to blame is MSI’s QC, which seems to be nonexistent.

Thank you. I apologize for my English; I'm from Spain. I think I have a problem?:

 

[DarkOmenX]

Thank you. I apologize for my English; I'm from Spain. I think I have a problem?:

I dont think that is a problem in your case , however :
"AI Overview
It is perfectly normal for a motherboard's base clock (BUS speed) to report slightly below 100 MHz, such as
99.7 MHz, 99.8 MHz, or 99.95 MHz, rather than a perfectly flat 100 MHz. This is rarely a fault, but rather a deliberate design choice or a limitation of measurement.

Here are the primary reasons why this happens:

1. Spread Spectrum (EMI Reduction)
The most common cause is a BIOS feature called Spread Spectrum.

• Purpose: To pass Electromagnetic Interference (EMI) regulations, motherboards must not create high spikes of radio frequency noise.
• Mechanism: Spread Spectrum intentionally "stretches" or dithers the bus clock frequency around the 100 MHz target (e.g., oscillating between 99.5 MHz and 100.5 MHz).
• Result: By spreading the energy, the EMI peaks are reduced, but this causes monitoring software to report a fluctuating or slightly lower-than-100 base clock.

2. Physical Clock Drift (Crystal Oscillators)
Motherboards use a physical quartz crystal to generate the base clock.

• Components Tolerance: Quartz crystals have natural variations based on quality, temperature, and age. A 50 parts per million (ppm) drift on a 100 MHz clock can lead to a reading of 100.5 MHz or slightly below, which is within tolerance.
• Not an Atomic Clock: Consumer motherboards do not use hyper-precise (and expensive) atomic clocks, resulting in small inaccuracies.

3. Monitoring Software Inaccuracy
Tools like CPU-Z or HWInfo might display fluctuations, but that does not mean the actual hardware is performing poorly.

• Measurement Timing: Any software application is taking a reading at a specific moment along the bus path, which can show variance.
• Reading Accuracy: The further the monitoring software is from the actual source clock generator, the less accurate the reading can be.

4. Windows Virtualization (Hyper-V)
If you have Virtualization technologies (such as Hyper-V in Windows, Windows Sandbox, or Windows Subsystem for Linux/WSL2) enabled, they can interact with the system clock, causing the reported bus speed to drop below 100 MHz (sometimes to around 98 MHz).

---

Does this matter?
No, it does not affect performance.

• A difference of 0.05 MHz or even 0.5 MHz is negligible (0.25%-0.5%).
• It does not mean your CPU is slower in any practical way.
• If it bugs you, you can usually disable "Spread Spectrum" in the BIOS to get a flatter 100 MHz, though this is not recommended if you are experiencing stability issues.

 

[DarkOmenX]

My spread spectrum setting is always on AUTO, so I have no idea whether it is enabled or disabled. Virtualization is disabled in the BIOS. I have never seen a BUS clock other than 100 unless I set it manually; when set to AUTO, it always stays at 100. Perhaps I am just fortunate to have a quartz oscillator that maintains the 100 precisely as it should.
Maybe MSI does not view slight bus frequency drifts as a hardware or other type of issue therefore they do not really address this "issue" that might not even be a real issue .
For quite some time, users have voiced complaints about this issue, yet MSI has neither addressed it nor provided any clarification. After lookin into it a bit deeper, I've found that MSI is not alone users of ASUS, Gigabyte, and ASROCK have reported similar problems, although MSI appears to receive the highest number of complaints.
While regular users and gamers may not notice a difference, overclocking enthusiasts might find this particularly frustrating. Adjusting overclocking values becomes more challenging due to the shifting bus, which means an overclock that works in one session may become unstable after a restart, even if the values remain unchanged.
Unfortunately, it seems there is no solution at this time, and I'm not certain whether MSI would accept this "issue" for RMA.

 

[Sko-Induktor]

Anyone that got the PCI bug can you try FCH=ON + BUS=100.3125) and report back for me this fixed the issue having any any problems

The first ~7–8 cold boots worked fine. Now I keep getting stuck at PCIe speed 4.0 (16.0 GT/s). In addition, the bus clock setting adjusts your memory speed, UCKL, and FCLK (pic here). There're only a few motherboards that have separate clock generators for RAM and the CPU (I think the Dark Hero has that).

I’m pretty much convinced by now that some Carbon WiFi motherboards just have bad clock generator components that don’t hold the frequency as they should, while others have excellent ones.

That's possible, but it doesn't explain why this issue only occurs during a cold start and why everything works normally EVERY TIME after a reboot.

I switched to the new Asus x870e Crosshair Dark Hero, using the same exact memory kit, CPU, GPU, etc.. and guess what. No problems at all.. everything just works. So in my opinion there is a physical hardware issue on some of these Carbon boards that no firmware update is going to fix.

That's currently the only known solution that has been proven to work. There are also people on the HWL Forum who have switched to a different (non-MSI) motherboard using the exact same hardware and settings and haven't had any problems at all.

I guess there really are some weird hardware combos that can cause the issue.

I can definitely say that my PCIe and cold boot issues started when I switched to the 5080 (PCIe 5.0). With the 4080 (PCIe 4.0), no matter which BIOS I used, I didn't have any problems at all. Before I replace my CPU or GPU, I’m going to throw out the board. That's for sure.

I think I have a problem?:

No. 7900XTX runs fine: 16x 4.0. Bus clock is normal.

report slightly below 100 MHz

When everything is running normally, the bus clock speed in CPU-Z fluctuates slightly below 100 MHz or minimal above 100 MHz. HWinfo reports a bus clock speed of 100.0 MHz. With FCH Spectrum on and bus clock set to Auto, the bus clock fluctuates mostly, from 99.70 to 99.90 (HWinfo reports 99.8 MHz). I tested various combinations a few pages back.

 

[kekolopez55154a02d1]

DarkOmenX​

Thanks for replying. Indeed, it seems to be normal behavior in my case. In fact, it fluctuates between 99.96 and 100 MHz.
It seems that if the motherboard isn't defective, it works very well. MSI should take better care of its users, according to what I've read here.

 

[DarkOmenX]

The first ~7–8 cold boots worked fine. Now I keep getting stuck at PCIe speed 4.0 (16.0 GT/s). In addition, the bus clock setting adjusts your memory speed, UCKL, and FCLK (pic here). There're only a few motherboards that have separate clock generators for RAM and the CPU (I think the Dark Hero has that).


That's possible, but it doesn't explain why this issue only occurs during a cold start and why everything works normally EVERY TIME after a reboot.


That's currently the only known solution that has been proven to work. There are also people on the HWL Forum who have switched to a different (non-MSI) motherboard using the exact same hardware and settings and haven't had any problems at all.


I can definitely say that my PCIe and cold boot issues started when I switched to the 5080 (PCIe 5.0). With the 4080 (PCIe 4.0), no matter which BIOS I used, I didn't have any problems at all. Before I replace my CPU or GPU, I’m going to throw out the board. That's for sure.


No. 7900XTX runs fine: 16x 4.0. Bus clock is normal.


When everything is running normally, the bus clock speed in CPU-Z fluctuates slightly below 100 MHz or minimal above 100 MHz. HWinfo reports a bus clock speed of 100.0 MHz. With FCH Spectrum on and bus clock set to Auto, the bus clock fluctuates mostly, from 99.70 to 99.90 (HWinfo reports 99.8 MHz). I tested various combinations a few pages back.

"That's possible, but it doesn't explain why this issue only occurs during a cold start and why everything works normally EVERY TIME after a reboot."
Some components perform optimally at stable working temperatures. Perhaps the oscillator does not function well when the components are cold, but operates as intended once they reach their operational temperature. Hard to say. During reboot the system is already at its working or operational temperature, meaning all components have reached their optimal conditions. In contrast, a cold start begins with the system at a lower temperature, requiring a warming-up process before reaching operational levels so some components may act iffy during this stage.
At work, we use professional electronic equipment, and some devices require at least 15 minutes of dry run before they are ready for use. For example, certain HPAs need between 5 and 10 minutes to reach optimal operating conditions and power output. While this isn't uncommon in professional setups, it can be somewhat inconvenient and is rarely seen with consumer electronics. However, increased complexity comes at a cost; today's motherboards are vastly different from those of a decade ago.

 

[lysyn12155d02d4]

Bios Asus NEW
1. Update to AGESA ComboAM5 1.3.0.1(BETA)
2. Optimized for the 9950X3D2 Dual Edition
3. Add Crosshair Tweak item (When using Mode 2 in the Crosshair series, the same memory rules as the test BIOS 99xx are apply)

 

[dlomond]

I think that some board must be ready bad mind always works with fch enable and bus at 100.3 never has cold boot issues now. If I turn it off wait 8 hours it anyway does it.

 

[dlomond]

I will go back to gigabyte master when I buy the next update.

 

[maddeniel159602e0]

No issues here I have about 12 USB devices connected across 3 hubs: one from the monitor, one from the keyboard, and an external USB hub, plus an external USB NVMe. For wireless devices, there’s an Xbox dongle, a Corsair Elite Void dongle, and a Logitech unified receiver. Add to that 3 onboard NVMe drives, a 5090, and a Creative AE-5 Pro sound card, and still zero issues no ghosting on any USB devices, all wireless devices stay connected at all times, never a single disconnection or any other anomaly. In fact, I’ve never seen my 5090, or previously the 4090, drop to a lower PCIe state and on the A95 or any other BIOS for that matter of fact, so how? However, I know what you are talking about as I experienced the exact same crap on X570 boards and there were people like me now stating "everything is working perfectly fine" and Im sure now it was for them however I was pulling my hair and thinking how they are talking [***CENSORED***] ... for once I'm on this side of the fence

i'm glad that everything works this well for you. i had that with my intel system back in the day with my 8086k.
that thing purs to this day, just now as my bedroom PC.

my AMD experience has been a nightmare mostly. but my second system burned me for another reason. with my current system, i had to undervolt my 5090, to not have the PC short out.
that problem was the powersupply tho, which i got a replacment for. and i should've then remembered what the guy which got me into PC gaming told me "when everything works, never upgrade what you don't have to, especially when it comes to BIOS".

but as ikarus, i decided to fly too close to the sun, and jumped on the AGESA 1.3.0.0 train and them wings got burned hard

i'm also am doubtfull that anybody experienced the problem i'm describing, as my PCIe doesn't degrade to 3.0 or 2.0 on reboot sometimes, but my PCIe goes nowhere anymore from 5.0 @ 8x since A95. it's stuck on that, when it should have the full x16 lanes which it also had everytime i checked, as i left the one slot open, and only have 3 m2.SSDs on my carbon board, even whilst i would actually love to populate all slots XD

as for the USB, i can deal with the mouse micro stutters as it's really only if i "push the dongle" with many tiny operations, and my controllers are luckily somehow magically excempt from those problems.
this whole AM5 plattform is like life, like a box of chocolates, and you never know when it's actually not chocolate XD

 

[marceluss19]

i'm glad that everything works this well for you. i had that with my intel system back in the day with my 8086k.
that thing purs to this day, just now as my bedroom PC.

my AMD experience has been a nightmare mostly. but my second system burned me for another reason. with my current system, i had to undervolt my 5090, to not have the PC short out.
that problem was the powersupply tho, which i got a replacment for. and i should've then remembered what the guy which got me into PC gaming told me "when everything works, never upgrade what you don't have to, especially when it comes to BIOS".

but as ikarus, i decided to fly too close to the sun, and jumped on the AGESA 1.3.0.0 train and them wings got burned hard

i'm also am doubtfull that anybody experienced the problem i'm describing, as my PCIe doesn't degrade to 3.0 or 2.0 on reboot sometimes, but my PCIe goes nowhere anymore from 5.0 @ 8x since A95. it's stuck on that, when it should have the full x16 lanes which it also had everytime i checked, as i left the one slot open, and only have 3 m2.SSDs on my carbon board, even whilst i would actually love to populate all slots XD

as for the USB, i can deal with the mouse micro stutters as it's really only if i "push the dongle" with many tiny operations, and my controllers are luckily somehow magically excempt from those problems.
this whole AM5 plattform is like life, like a box of chocolates, and you never know when it's actually not chocolate XD

I learned that there is no reason to dealing with bad mobo, trying to fix it via bios updates etc…. Bad mobo is like cancer. I just buy another mobo from other vendor and that’s it. Sure I am losing some money, but fcuk it, I need working PC and peace.

 

[maddeniel159602e0]

I learned that there is no reason to dealing with bad mobo, trying to fix it via bios updates etc…. Bad mobo is like cancer. I just buy another mobo from other vendor and that’s it. Sure I am losing some money, but fcuk it, I need working PC and peace.

i'd agree with you since my strix 670e-f, but this time, with the carbon, i know exactly that PCIe being stuck at 5.0 @ x8, instead of x16, happend after i updated to A95. and also since A95, my PC also got stuck trying to reboot the system twice. if the next BIOS update doesn't fix it, i'm gonna downgrade to the last version before the AGESA 1.3.0.0 update happend, and if either of those would fix the issue, i'm just gonna ignore BIOS and AMD chipset drivers from then on out and jsut enjoy the PC XD

as for the problems i had with my strix mobo, that was a long nightmare, where i just didn't want to bite the bullet, and accept that it wasn't just the BIOS. there were a few ppl with very similiar problems in the ASUS forum, to my problems back then, so i was keeping hopes up that it could be a BIOS issue after all.

as for the carbon, before A95, that thing ran perfectly and i just upgraded out of habit, which i developed from the strix board, hoping that i could make perfect better.
hence why i mentioned ikarus XD

 

[DarkOmenX]

cold boot after about 48hrs didn't use the PC and this is the result immediately after entering Windows

 

[Alyred]

i'd agree with you since my strix 670e-f, but this time, with the carbon, i know exactly that PCIe being stuck at 5.0 @ x8, instead of x16, happend after i updated to A95. and also since A95, my PC also got stuck trying to reboot the system twice. if the next BIOS update doesn't fix it, i'm gonna downgrade to the last version before the AGESA 1.3.0.0 update happend, and if either of those would fix the issue, i'm just gonna ignore BIOS and AMD chipset drivers from then on out and jsut enjoy the PC XD

as for the problems i had with my strix mobo, that was a long nightmare, where i just didn't want to bite the bullet, and accept that it wasn't just the BIOS. there were a few ppl with very similiar problems in the ASUS forum, to my problems back then, so i was keeping hopes up that it could be a BIOS issue after all.

as for the carbon, before A95, that thing ran perfectly and i just upgraded out of habit, which i developed from the strix board, hoping that i could make perfect better.
hence why i mentioned ikarus XD

This really, really sounds like your mainboard thinks either M2_2 or PCI_E2 is populated, or you have bifurcation enabled. When you update your BIOS, are you making sure to do a complete BIOS flush (power down, switch off or unplug Power Supply, hold down the BIOS Reset button on the back (middle button) for 10-15 sec, then go into the BIOS and for good measure, load bios defaults then save and restart before going into your BIOS. Early on, MSI had some issues with retaining some old settings in different fields and it required the BIOS to be reset during updates. Not sure if that screwball behavior has returned. I'm still on A64.
If you haven't installed anything new in PCI_E2 or moved one of your drives into M2_2 and forgotten (Hey, we've all done something like that, especially when trying ANYthing to get something workinng again), check to make sure you have the M2_2 and PCI_E2 set to Auto (I don't remember if you can disable them) and make sure bifurcation isn't set to 8/4/4/4 or anything? Even with all the cold-start/cold-resume issues I've seen (Booted up to PCI v3 just now) I've never seen the lane width change, especially since it's keeping PCIe v5.

 

[maddeniel159602e0]

This really, really sounds like your mainboard thinks either M2_2 or PCI_E2 is populated, or you have bifurcation enabled. When you update your BIOS, are you making sure to do a complete BIOS flush (power down, switch off or unplug Power Supply, hold down the BIOS Reset button on the back (middle button) for 10-15 sec, then go into the BIOS and for good measure, load bios defaults then save and restart before going into your BIOS. Early on, MSI had some issues with retaining some old settings in different fields and it required the BIOS to be reset during updates. Not sure if that screwball behavior has returned. I'm still on A64.
If you haven't installed anything new in PCI_E2 or moved one of your drives into M2_2 and forgotten (Hey, we've all done something like that, especially when trying ANYthing to get something workinng again), check to make sure you have the M2_2 and PCI_E2 set to Auto (I don't remember if you can disable them) and make sure bifurcation isn't set to 8/4/4/4 or anything? Even with all the cold-start/cold-resume issues I've seen (Booted up to PCI v3 just now) I've never seen the lane width change, especially since it's keeping PCIe v5.

this is exactly the type of message i wanted, where somebody tells me things i can try and do till the next BIOS comes out, and you also informed me probably about the best, last and most stable AGESA 1.2.X.X BIOS version with A64 i could revert back to, if the next BIOS update still says no x16.
i'm newer to the carbon BIOS so i still have some learning to do, as my brain has been on ASUS for almost 12 years and i have a harder time dealing with good names for the tech, but the descriptions from MSI even when there's so much space to write sooooo much, only really gives you a surface level description of what each change does, AT BEST XD

what i can tell you right now after waking up and looking into BIOS, when i force the 5.0 option, the options for choosing what lane config i want, the highest it tells me i can pick is x8 + x8 aside from the auto option which should probably pick x16 which it doesn't, and according to the manual which i checked, the option x16 should be available. but maybe auto also could just mean *the highest possible/doubles as x16". that one was also not described all that well. also, i ccouldn't find a bifurcationn setting, but i seen the "8/4/4/4" you mentioned when i search for PCIe, and there is an option like that where i also found the "x8 + x8" setting.

as for having it populated, nope, last time i opened up the PC, i just switched the PSU as it was broken, what i didn't do is the whole BIOS procedure you explained, which wasn't necessary with any BIOS yet, but as you said, back on ASUS, pressing that button on the back of the MOBO was sometimes alleviating the issues for a moment, and it was imperial when flashing backwards in versions. next time the oportunity arises, i'm gonna do that. the setting i change are not many anyway. the old ASUS mobo has completly turned me off from learning how to do tighter RAM timings or the whole PBO thing, cause of the immense RAM issues the ASUS BIOS had anyway, where only EXPO tweaked kept it from getting stuck in memory training every time, normal EXPO functions otherwise were FUBAR. and as said, the same RAM is in the new PC now, and those issues never happend aside from now with A95 with the two reboot stucks i had.

i luckily copied the part that i read about the power savings thing and i just found it in my notebook which i meant, which i haven't tried yet which is this quote "In my case the GPU-Z render test uses 65–70W on my 5080, and the PCIe link speed jumps straight to 32.0 GT/s. Of course, i can select “Max Performance” in NVCP and disable the C-states (ASPM) in the BIOS, but then the system constantly consumes unnecessary power (50w+) without any significant benefit."

y'all think that trying to disable c-states is also an idea to try to force my PC into using x16? in HWINFO it says about the PCIe link speed which it only shows "32.0 GT/s" in all slots, but when it would be x16 according to google confusing me hellishly, it says it should say "64.0 GT/s". but in the quote, it sounds like "32.0 GT/s" is correct.

i also googled in how far the x8 would actually slow me down, and it says should be negliable. one oddity regarding my graphics card is, i'm using the nvidia app in beta since 2 days, so that i could've tried the new dynamic frame gen feature, but when i try to change that particular setting there, it says an error occured and it can't change that setting. i'm obviously now OCD'ishly thinking that that doesn't work cause my PCIe is not at x16, cause of the causal chain, but could also just be that the beta doesn't work for me for some other reason which wouldn't be the first time, and nvidia drivers are also not the cream of the crop rn either.

sorry for me being so detailled and trying to stuff as much info density into one message, but i take time to reply cause i try to mention every eventuality, so that somebody can help me which knows the matter better than i do, with more precision, as my knowledge pool thanks to the terrible ASUS board i had, is now moreso not on AM5 skills, but on ASUS mobo troubleshooting, and i have a hard time to let go, cause i subjected myself to that broken mobo for a little over 2 years.
i know, calling it tech-PTSD is a lil much, but thanks to my health, i'm kinda welded to my PC for most of the times, and that is why this whole topic riles me up to the extent that it's on my mind too often, causing me some digital anguish, as the last two "systems" i had, in total, now only didn't have any issues, for like 4 weeks from changing the PSU, to upadating to A95 and i'm back in problem county, population, me, certainly XD
i jsut try to minimize the amount of message that need exchanging, that's why i'm so info dense is what i'm trying to say.

but, your quote is also in my notebook now, and i hope that in the next reply, y'all either give me good info with that c-state thing which may fix the problem for now, or that i had the oportunity to climb behind my PC which is a lil cumbersome cause i'm really bad at "climbing", but that that may do the trick.

thanks for all the help btw.

 

[Alyred]

this is exactly the type of message i wanted, where somebody tells me things i can try and do till the next BIOS comes out, and you also informed me probably about the best, last and most stable AGESA 1.2.X.X BIOS version with A64 i could revert back to, if the next BIOS update still says no x16.
i'm newer to the carbon BIOS so i still have some learning to do, as my brain has been on ASUS for almost 12 years and i have a harder time dealing with good names for the tech, but the descriptions from MSI even when there's so much space to write sooooo much, only really gives you a surface level description of what each change does, AT BEST XD

what i can tell you right now after waking up and looking into BIOS, when i force the 5.0 option, the options for choosing what lane config i want, the highest it tells me i can pick is x8 + x8 aside from the auto option which should probably pick x16 which it doesn't, and according to the manual which i checked, the option x16 should be available. but maybe auto also could just mean *the highest possible/doubles as x16". that one was also not described all that well. also, i ccouldn't find a bifurcationn setting, but i seen the "8/4/4/4" you mentioned when i search for PCIe, and there is an option like that where i also found the "x8 + x8" setting.

as for having it populated, nope, last time i opened up the PC, i just switched the PSU as it was broken, what i didn't do is the whole BIOS procedure you explained, which wasn't necessary with any BIOS yet, but as you said, back on ASUS, pressing that button on the back of the MOBO was sometimes alleviating the issues for a moment, and it was imperial when flashing backwards in versions. next time the oportunity arises, i'm gonna do that. the setting i change are not many anyway. the old ASUS mobo has completly turned me off from learning how to do tighter RAM timings or the whole PBO thing, cause of the immense RAM issues the ASUS BIOS had anyway, where only EXPO tweaked kept it from getting stuck in memory training every time, normal EXPO functions otherwise were FUBAR. and as said, the same RAM is in the new PC now, and those issues never happend aside from now with A95 with the two reboot stucks i had.

i luckily copied the part that i read about the power savings thing and i just found it in my notebook which i meant, which i haven't tried yet which is this quote "In my case the GPU-Z render test uses 65–70W on my 5080, and the PCIe link speed jumps straight to 32.0 GT/s. Of course, i can select “Max Performance” in NVCP and disable the C-states (ASPM) in the BIOS, but then the system constantly consumes unnecessary power (50w+) without any significant benefit."

y'all think that trying to disable c-states is also an idea to try to force my PC into using x16? in HWINFO it says about the PCIe link speed which it only shows "32.0 GT/s" in all slots, but when it would be x16 according to google confusing me hellishly, it says it should say "64.0 GT/s". but in the quote, it sounds like "32.0 GT/s" is correct.

i also googled in how far the x8 would actually slow me down, and it says should be negliable. one oddity regarding my graphics card is, i'm using the nvidia app in beta since 2 days, so that i could've tried the new dynamic frame gen feature, but when i try to change that particular setting there, it says an error occured and it can't change that setting. i'm obviously now OCD'ishly thinking that that doesn't work cause my PCIe is not at x16, cause of the causal chain, but could also just be that the beta doesn't work for me for some other reason which wouldn't be the first time, and nvidia drivers are also not the cream of the crop rn either.

sorry for me being so detailled and trying to stuff as much info density into one message, but i take time to reply cause i try to mention every eventuality, so that somebody can help me which knows the matter better than i do, with more precision, as my knowledge pool thanks to the terrible ASUS board i had, is now moreso not on AM5 skills, but on ASUS mobo troubleshooting, and i have a hard time to let go, cause i subjected myself to that broken mobo for a little over 2 years.
i know, calling it tech-PTSD is a lil much, but thanks to my health, i'm kinda welded to my PC for most of the times, and that is why this whole topic riles me up to the extent that it's on my mind too often, causing me some digital anguish, as the last two "systems" i had, in total, now only didn't have any issues, for like 4 weeks from changing the PSU, to upadating to A95 and i'm back in problem county, population, me, certainly XD
i jsut try to minimize the amount of message that need exchanging, that's why i'm so info dense is what i'm trying to say.

but, your quote is also in my notebook now, and i hope that in the next reply, y'all either give me good info with that c-state thing which may fix the problem for now, or that i had the oportunity to climb behind my PC which is a lil cumbersome cause i'm really bad at "climbing", but that that may do the trick.

thanks for all the help btw.

C-states deal with processor power modes for power usage. Disabling C-states means your processor is never going into a power savings mode when it doesn't need to be running at full power, and definitely shouldn't be impacting your PCIe lane width.

It's a bit unclear if you are still on A95 or if you've gone back to an earlier BIOS, but if you're still in A95 and since you are not doing any custom timings or settings, I would definitely try a full BIOS reset now to make sure there's no old settings that are still stuck and being applied to the wrong fields. Completely cut power to the power supply, via the switch on the back. Hold down that middle button on the back marked CMOS with the arrow around it for 15 seconds. Boot back up go into the BIOS, and go to the save and exit page and select "load bios defaults" and click save and exit. This should make sure ALL settings are not just cleared, but are set to MSI's default settings. This second step may not be absolutely necessary, but it can't hurt. Then, go in and turn on your XMP for your RAM if that was stable before, make sure the PCI_E settings are at "auto" for PCI-E_1 and 2, and bifurcatiin is also set to "auto" as any other setting will assume that you want to force a non-X16 setting on your GPU slot. Anything plugged into the second PCIe slot WILL cut your lanes on the main PCIe to 8, no matter what you set in there.

For now, since your board seems stable at PCIe 5.0, keep the main slot set to "auto".

As an explanation of PCIe slot speeds, PCIe v5 is twice as fast maximum bandwidth as PCIe v4. So v5x8 is the same bandwidth as v4x16 (there's some additional communications overhead there so that's not exact, but close enough for speed estimation). A 5080 (and indeed, a 5090) rarely exceeds the bandwidth available on a v4x16, so yes, you're not going to see much real-world impact on a v5x8 connection.
That being said... the following is an opinion from what I've noticed.
Whenever my system comes up from cold suspend as other than PCIe v5, there's something wrong system-wide that will eventually cause a complete crash and restart. The lower the PCI Gen, the faster it crashes and the nore noticeable OTHER system functions are, like mouse jitter if it's connected through a USB port attached to the chipset (which connects back to the CPU through a PCIe interface) and even transferring files to a USB attached device are more likely to be corrupted with mismatched hashes or outright failures. These instances are eliminated by either doing a warm restart, OR putting my PC back into sleep/suspend and waking it up again. I am now in the habit of just checking GPUz every time I wake the PC.
Another related thought: setting the GPU slot to "auto" allows the PC to come up, even if it can't do the negotiation on PCI higher than 2 or 3 when starting up. Forcing the BIOS to PCI v5 on the GPU slot, if it can't negotiate v5 properly, seems to cause an instant crash. Much faster to sleep/resume and keep my programs running from my last session and let it auto-detect the first time, even if the initial start comes up lower.

For me - I keep Global C-States on as it will save a little power when I'm not needing all cores to be running at full power, and turning on FCH Spread Spectrum seems to assist the PCIe version detection, it does happen less often. I do force PCIe 4.0 on my two SSDs in slots 1 and 3, since they're both PCIe v4 Samsung Pro 990's. PCI_E1 Lanes Configuration is set to "Auto". This is the most stable so far for me. I've been holding off upgrading after A64 only because I don't have time to re-test speeds/etc on my RAM, and the only main problem I am having on my two x870E Carbons is the infrequent PCIe speed detection. But I've never, ever seen it detect fewer lanes.

 

[DarkOmenX]

this is exactly the type of message i wanted, where somebody tells me things i can try and do till the next BIOS comes out, and you also informed me probably about the best, last and most stable AGESA 1.2.X.X BIOS version with A64 i could revert back to, if the next BIOS update still says no x16.
i'm newer to the carbon BIOS so i still have some learning to do, as my brain has been on ASUS for almost 12 years and i have a harder time dealing with good names for the tech, but the descriptions from MSI even when there's so much space to write sooooo much, only really gives you a surface level description of what each change does, AT BEST XD

what i can tell you right now after waking up and looking into BIOS, when i force the 5.0 option, the options for choosing what lane config i want, the highest it tells me i can pick is x8 + x8 aside from the auto option which should probably pick x16 which it doesn't, and according to the manual which i checked, the option x16 should be available. but maybe auto also could just mean *the highest possible/doubles as x16". that one was also not described all that well. also, i ccouldn't find a bifurcationn setting, but i seen the "8/4/4/4" you mentioned when i search for PCIe, and there is an option like that where i also found the "x8 + x8" setting.

as for having it populated, nope, last time i opened up the PC, i just switched the PSU as it was broken, what i didn't do is the whole BIOS procedure you explained, which wasn't necessary with any BIOS yet, but as you said, back on ASUS, pressing that button on the back of the MOBO was sometimes alleviating the issues for a moment, and it was imperial when flashing backwards in versions. next time the oportunity arises, i'm gonna do that. the setting i change are not many anyway. the old ASUS mobo has completly turned me off from learning how to do tighter RAM timings or the whole PBO thing, cause of the immense RAM issues the ASUS BIOS had anyway, where only EXPO tweaked kept it from getting stuck in memory training every time, normal EXPO functions otherwise were FUBAR. and as said, the same RAM is in the new PC now, and those issues never happend aside from now with A95 with the two reboot stucks i had.

i luckily copied the part that i read about the power savings thing and i just found it in my notebook which i meant, which i haven't tried yet which is this quote "In my case the GPU-Z render test uses 65–70W on my 5080, and the PCIe link speed jumps straight to 32.0 GT/s. Of course, i can select “Max Performance” in NVCP and disable the C-states (ASPM) in the BIOS, but then the system constantly consumes unnecessary power (50w+) without any significant benefit."

y'all think that trying to disable c-states is also an idea to try to force my PC into using x16? in HWINFO it says about the PCIe link speed which it only shows "32.0 GT/s" in all slots, but when it would be x16 according to google confusing me hellishly, it says it should say "64.0 GT/s". but in the quote, it sounds like "32.0 GT/s" is correct.

i also googled in how far the x8 would actually slow me down, and it says should be negliable. one oddity regarding my graphics card is, i'm using the nvidia app in beta since 2 days, so that i could've tried the new dynamic frame gen feature, but when i try to change that particular setting there, it says an error occured and it can't change that setting. i'm obviously now OCD'ishly thinking that that doesn't work cause my PCIe is not at x16, cause of the causal chain, but could also just be that the beta doesn't work for me for some other reason which wouldn't be the first time, and nvidia drivers are also not the cream of the crop rn either.

sorry for me being so detailled and trying to stuff as much info density into one message, but i take time to reply cause i try to mention every eventuality, so that somebody can help me which knows the matter better than i do, with more precision, as my knowledge pool thanks to the terrible ASUS board i had, is now moreso not on AM5 skills, but on ASUS mobo troubleshooting, and i have a hard time to let go, cause i subjected myself to that broken mobo for a little over 2 years.
i know, calling it tech-PTSD is a lil much, but thanks to my health, i'm kinda welded to my PC for most of the times, and that is why this whole topic riles me up to the extent that it's on my mind too often, causing me some digital anguish, as the last two "systems" i had, in total, now only didn't have any issues, for like 4 weeks from changing the PSU, to upadating to A95 and i'm back in problem county, population, me, certainly XD
i jsut try to minimize the amount of message that need exchanging, that's why i'm so info dense is what i'm trying to say.

but, your quote is also in my notebook now, and i hope that in the next reply, y'all either give me good info with that c-state thing which may fix the problem for now, or that i had the oportunity to climb behind my PC which is a lil cumbersome cause i'm really bad at "climbing", but that that may do the trick.

thanks for all the help btw.

If neither the PCIE2 nor the M2_2 slot is occupied, your VGA should not run at x8 speed. However, if either or both of these slots are occupied, the switch activates and your VGA will operate at x8. If your VGA is running at x8 speed without anything installed in PCIE2 or M2_2, your motherboard may be faulty and should be considered for RMA. This issue should not occur, and I personally have never experienced it with my X870E Carbon WIFI.
If, for any reason, the SWTICH is activated, a standard CMOS clear should resolve the issue.
For a more thorough CMOS reset, first unplug the board from power, remove the battery, and short the on-board CMOS clear pins. Wait approximately 10 minutes, then reinstall the battery.
This process should restore the board to its original factory settings.
Avoiding references to mythical beings and overly literal expressions can help keep the post focused on the main issue and reduce the size .

 

[Alyred]

If neither the PCIE2 nor the M2_2 slot is occupied, your VGA should not run at x8 speed. However, if either or both of these slots are occupied, the switch activates and your VGA will operate at x8. If your VGA is running at x8 speed without anything installed in PCIE2 or M2_2, your motherboard may be faulty and should be considered for RMA. This issue should not occur, and I personally have never experienced it with my X870E Carbon WIFI.
If, for any reason, the SWTICH is activated, a standard CMOS clear should resolve the issue.
For a more thorough CMOS reset, first unplug the board from power, remove the battery, and short the on-board CMOS clear pins. Wait approximately 10 minutes, then reinstall the battery.
This process should restore the board to its original factory settings.
Avoiding references to mythical beings and overly literal expressions can help keep the post focused on the main issue and reduce the size .

...for someone that hates AI slop, you did a really good impersonation of AI summarization of my post that they were replying to...
There's no need to short pins or pop the BIOS battery on the Carbon. The BIOS Clear/reset button in the back performs the same function (shorting the pins) much more conveniently, so far as I understand it from the documentation.