MPG X870E CARBON WIFI Beta BIOS

Svet

Svet

Well-known member
Administrator
Joined
Nov 20, 2003
Messages
85,347
>>> 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 <<<
 
Last edited by a moderator:
Sort by date Sort by votes
marceluss19615d002e6 said:
„the neighbor's grass is greener” my first AMD was x570 Aorus Master, mobo was great for me, so I bought X670E Aorus Master. This mobo was absolute nightmare, so replaced this garbage with x670E Taichi Carrara. This mobo was great so I bought X870E Taichi for 9950X3D build. Returned it after reading about cpu drama. Bought MSI x870E Carbon, after reading on Reddit that Asus Croshair is POS mobo with tons of problems. There is no good solution this days. Shitshow everywhere. Fingers crossed, my Carbon is good so far. Btw. F..k A33 lottery good/bad for my build.
Click to expand...
Yeah, there's always the greener grass, but while there's still problems out there with all of the x870 lines other manufacturers seem to have worked out most of their issues. There's still some problems in the MSI x870 boards that have been there since day 1, and none of the other manufacturer's boards seem to be quite so "random" in things like PCI speed lockups, core voltage spikes, and overall stability.
 
Upvote 0 Downvote
Indabox said:
I checked the CPU, no bent pins and it's aligned correctly. Also reseated Ram.

Settings right now: (everything else is default)
AMD Overclocking:
Expo 6000 (CL30-36-36-80) with Kingston Furry 32gb kit Hynix A-die
PBO: Advanced
PBO Limits: Motherboard
Curve Optimizer: All core, negative, 20

CPU will idle at 42c in BIOS, in windows Hwinfo64 shows 39c.
CPU Core voltage showing 1.166 V in bios at idle
I do see 1 critical error in windows system log when it fails to boot "The system has rebooted without cleanly shutting down" so its like windows booted, but i never made it to the desktop. just freezes at bios screen with the spinning circle showing windows loading but circle also stops moving..

Cinebench r23 will run 30 min with no issues, memtest no issues, and Aida64 no issues..

Screenshot shows voltage with those settings idle in BIOS.

Just stumped here at this point.. Anything else I should check?
Click to expand...
As @vic123159d02e5 was saying, try the typical current option. I don't know if you've tried a fresh Windows install, but it might be something partially corrupted (driver? sys file?) during boot. Try running the sfc updates and scans, and maybe DDU your chipset and video drivers and re-install them fresh. Since once you DO get into windows it sounds stable... this is usually something trying to load at boot.
 
Upvote 0 Downvote
EDIT: found the setting and changed it to typical.. will test

Where in the BIOS is the "typical current" option? I looked and am not seeing it. I will try that and if no luck, will pull my GPU and try running just the integrated graphics to rule out a GPU issue / GPU driver issue. I'm not getting any driver related errors in windows event viewer, but wont hurt.
 
Last edited:
Upvote 0 Downvote
Indabox said:
EDIT: found the setting and changed it to typical.. will test

Where in the BIOS is the "typical current" option? I looked and am not seeing it. I will try that and if no luck, will pull my GPU and try running just the integrated graphics to rule out a GPU issue / GPU driver issue. I'm not getting any driver related errors in windows event viewer, but wont hurt.
Click to expand...
(CBS overclocking part) "typical current Idle" instead of "auto"
 
Upvote 0 Downvote
A33 is one of the most stable versions for me with 9950x3d.
Personally, the only thing I still miss is a bit more fine-tuning options, similar to what another brand (the one starting with "A") provides.
Would it be possible for MSI to add more advanced settings directly in the BIOS? Things like manual disabling of SATA and USB ports, extended voltage tuning, PMIC adjustments, and more.
Alternatively, it would be awesome to have a hidden "Expert Mode" for experienced users — just like it's done on some laptops.
I'm sure many enthusiasts would appreciate this. No need to make it public — let it stay hidden for those who know how to find it (a secret expert mode).
I believe many would support me on this.
@Svet if possible, could you please pass this along to the BIOS team?
I feel like opening a ticket for a feature request might not be very effective.
Of course, I understand MSI's concerns about burning CPUs, but I'm sure some functions could be made accessible without posing any risk to the hardware.
 
Last edited:
Upvote 0 Downvote
@Alyred If you have problems with the PCIe bus after a cold boot, try turning on Spread Spectrum. It helped me — the bus doesn't get stuck at Gen 1.1 as often, but it doesn't always fix it. I’m using a 9800X3D, 5080, and MPG B850 EDGE TI WIFI. Until they fix it with a BIOS update, we just have to wait. Nvidia says the issue is entirely the motherboard manufacturer's fault.
 
Last edited:
Upvote 0 Downvote
Alyred said:
M2_3 actually lost sequential read/write but gained in Rnadom Read/Write IOPS
Click to expand...
M.2 on the chipset controller may have up to 5% to 10% reduction in performance compared to ones on the CPU
 
Upvote 0 Downvote
L E G I T said:
@Alyred If you have problems with the PCIe bus after a cold boot, try turning on Spread Spectrum. It helped me — the bus doesn't get stuck at Gen 1.1 as often, but it doesn't always fix it. I’m using a 9800X3D, 5080, and MPG B850 EDGE TI WIFI. Until they fix it with a BIOS update, we just have to wait. Nvidia says the issue is entirely the motherboard manufacturer's fault.
Click to expand...
I'm sure Nvidia points finger on motherboard vendors and you ask msi and other vendors they point right back
just wondering how do you diagnostic on which gen the m,2 is running? what softer or method do you use?
I have 3 M.2s in my system 2x4GEN and 1x5 GEN they all are running at respective speeds, generations and lane allocations I never touched any setting in BIOS regarding M.2 maybe some controller on your bord is faulty in some way?
 
Upvote 0 Downvote
pogdmm said:
A33 is one of the most stable versions for me with 9950x3d.
Personally, the only thing I still miss is a bit more fine-tuning options, similar to what another brand (the one starting with "A") provides.
Would it be possible for MSI to add more advanced settings directly in the BIOS? Things like manual disabling of SATA and USB ports, extended voltage tuning, PMIC adjustments, and more.
Alternatively, it would be awesome to have a hidden "Expert Mode" for experienced users — just like it's done on some laptops.
I'm sure many enthusiasts would appreciate this. No need to make it public — let it stay hidden for those who know how to find it (a secret expert mode).
I believe many would support me on this.
@Svet if possible, could you please pass this along to the BIOS team?
I feel like opening a ticket for a feature request might not be very effective.
Of course, I understand MSI's concerns about burning CPUs, but I'm sure some functions could be made accessible without posing any risk to the hardware.
Click to expand...
It's stable, but is it the best BIOS they've delivered?
Depends on what you're after. If it's pure performance, maybe not. I'm currently running 105 BCLK, passing all benches,OCCT, AIDA64, CPU-Z, Y-cruncher, Cinebench R23, with performance similar to 104 BCLK on A25, but needing much more vcore.
I was thrilled at first by A33, not so much now, it's somewhere between a "meh" and a "m...k".

As for the BIOS settings, rather than keeping them vague or acting like it's some secret, I wish they'd just explain what the settings actually do. Take the CPPC settings in the AMD CBS Overclocking section, values : Frequency, cache, Driver .... I have a rough idea what they mean, but in practice, I’ve seen little to no real-world impact from tweaking them.
  • Not to mention that the appearance of that setting coincide with loss of performance in single thread and multi-thread for every bios version( from A25 and on), wrong or right do not know but the timeline says so .
Now, about missing settings, the one I really want is Fmax. ASUS and Gigabyte include it, but MSI apparently thinks we're too dumb to use it, especially critical when you're overclocking via BCLK and there's no ECLK support on that board... That’s exactly when an Fmax setting becomes useful. Way more so than the CPU Watchdog, which I disabled the second I saw it.
 
Last edited:
Upvote 0 Downvote
Sethos812d002b2 said:
And now the line about adjusting CPU core voltage has been removed on the support page for A33
Click to expand...
Typical MSI , I'm long enough alive to know what MSI once was , unfortunately they degraded massively through out the years as a company . But then again so is ASUS as one of the worse out there and Gigabyte not far behind , then there is ASROCk with their motherboards roasting the x3d and at the end you left with Fking nothing . But all of them has something in common rising and scalping prices while offering very little or nothing in return .
 
Upvote 0 Downvote
Alan J T said:
M.2 on the chipset controller may have up to 5% to 10% reduction in performance compared to ones on the CPU
Click to expand...
Yeah, understand this. However, the performance shouldn't decrease between bios versions on the same port, which it has on several different BIOS revisions. Performance hit from going through the chipset is usually in the 3-5% range so long as you aren't doing anything to make those PCIe lanes be used which causes the drive to wait.
In A33's case, M2_3's performance decreased from what M2_3's performance was in A25.
 
Upvote 0 Downvote
L E G I T said:
@Alyred If you have problems with the PCIe bus after a cold boot, try turning on Spread Spectrum. It helped me — the bus doesn't get stuck at Gen 1.1 as often, but it doesn't always fix it. I’m using a 9800X3D, 5080, and MPG B850 EDGE TI WIFI. Until they fix it with a BIOS update, we just have to wait. Nvidia says the issue is entirely the motherboard manufacturer's fault.
Click to expand...
Spread Spectrum reduces interference with EMI but doesn't eliminate it entirely. In my case, I hadn't had a single PCIe lock until a couple of days ago, with nothing else in my environment having changed. I upgraded to A33 and had the same problem. Switching the PCIe port from forced PCIe v5 to "auto" seems to have fixed the problem, for now. We'll see how it behaves in the next few days, but so far it's been properly scaling down to 2.0 and immediately jumping back up to 5.0 whenever I run a render test or push on 3d acceleration at all.
If that's the working mitigation, I suspect it's still something that MSI hasn't locked down and fixed with how the PCIe bus interacts with OSes.
For me, it wasn't the cold boot - that always worked to clear up the issue. It was waking from sleep that kept causing the PCIe to sometimes come up at 1.1, noticable immediately since the mouse would start chopping during movement every 3 seconds or so.
 
Upvote 0 Downvote
vic123159d02e5 said:
Thanks for the info. Oddly enough, I found out myself last night that using Auto Voltage instead of AMD Overclocking mode reduced the voltage slightly.
Click to expand...
Interesting, you set VCore (or were you talking vSoC?) to auto rather than setting a fixed voltage, and it is MORE stable than the fixed number? Similar to my PCIe setting I wonder if MSI isn't properly honoring values set in those fields.
 
Upvote 0 Downvote
Alyred said:
Interesting, you set VCore (or were you talking vSoC?) to auto rather than setting a fixed voltage, and it is MORE stable than the fixed number? Similar to my PCIe setting I wonder if MSI isn't properly honoring values set in those fields.
Click to expand...
I never stick to a fixed vcore. Using a set number is the death of my BCLK overclocking. I let the motherboard handle the juice and tweak it with the CO curve and CO curve shaper (yep, I use both, actually -45 all core and -5 on curve shaper).
When it comes to AMD overclocking, I always keep things on auto, unless I’m experimenting with an all-core fixed overclock. That only works well with SMT off though. Fun fact: that was actually my first try at overclocking with this board and CPU. It felt more like an Intel-style overclock. Sure, it delivers, but you end up losing a lot of multi-threaded performance.
For Vsoc, I use a fixed voltage set to 1.3 V.
I’ve been running BCLK at 105 recently, hitting around 5696 MHz on max boost. The downside is that it delivers about the same performance as I had with A25 at 104 BCLK. ( except memory benches where it is really really faster)
1745947478144.png
 
Last edited:
Upvote 0 Downvote
Sethos812d002b2 said:
And now the line about adjusting CPU core voltage has been removed on the support page for A33
Click to expand...
The reason they recommend it for that bios version is if you use the non AMD Voltage setting, you may experience a problem of the CPU getting locked as a LOW Speed, lets say, ie 1200MHz or lower instead of the base 3600MHz

And this is only if you change the CPU voltage for overclocking or undervolting
SOC is not affected
 
Upvote 0 Downvote
Back
Top