MPG X870E CARBON WIFI Beta BIOS

Sort by date Sort by votes
Alyred said:
Interesting to know. I'm not THAT hardcore into overclocking, just like to get performance that I paid for at a reasonable heat level. But it does sound that if the motherboard treats all cores at the least core's CO, that still sounds like a BIOS issue to me... otherwise, why even have a per-core setting?
Click to expand...
I really don’t know, if same behavior is happening on other boards or not, I have very little experience with AMD
 
Upvote 0 Downvote
vic123159d02e5 said:
I really don’t know, if same behavior is happening on other boards or not, I have very little experience with AMD
Click to expand...
Well, it didn't happen in previous generations - and it really wouldn't make any sense to have a per-core setting if only the worst core was considered...
 
Upvote 0 Downvote
Alyred said:
Interesting to know. I'm not THAT hardcore into overclocking, just like to get performance that I paid for at a reasonable heat level. But it does sound that if the motherboard treats all cores at the least core's CO, that still sounds like a BIOS issue to me... otherwise, why even have a per-core setting?
Click to expand...
I really don’t know, if same behavior is happening on other boards or not, I have very little experience with AMD
Alyred said:
Well, it didn't happen in previous generations - and it really wouldn't make any sense to have a per-core setting if only the worst core was considered...
Click to expand...
Well, you know, you can set a per-core multiplier with Ryzen Master, but it will only work based on the slowest core speed in an all-core test. There are things like this that just don't make sense but still exist.;)
 
Upvote 0 Downvote
Alyred said:
Think I'll be rolling back to A25 on my test bench.
Click to expand...
I've been running A25 since February 25, 2025.
While more stable than the previous, I did just wake up today to my first reboot with my system at the BIOS and not showing the C drive.
I haven't overclocked yet and didn't touch anything which just means - the BIOS isn't there yet.

Hoping A26 doesn't cause me issues.
 
Upvote 0 Downvote
Anaksagoras said:
A25 works great, I did all kinds of tests, no problems.
Click to expand...
If you're not using a PCIe v5 GPU, A25 works pretty well. A26 gives some more PCIe v5 options but unknown what else it really adds - It would be neat if @Svet could post a short changelog with new updates for us to see what to look for and focus our testing on, but I'm glad for the betas either way.
I've had a few stability issues with A26, but most of them seem to have resolved since pushing hard on the memory training and getting it to finally "reset" and clear out anything still in there. RAM seems to be running stable now again, though with a bit less extreme timings than A25 afforded. Haven't had it drop a RAM socket in a while and just passed 2 Memtest86 4-pass runs as well as a full 6.5 hour gold OCCT stability run with A26 and fairly tight timings starting with the Hynix 24GB optimized timings as base and going from there. Shutting down that test bench for a few hours to let it cool and then will run another one of those Gold (no GPU) OCCT tests.
For most folks, A26 should be a significant improvement over A23 either way, and should be posted in the "official" section for now, I've sent a LOT of folks to the A26 Beta posts from Reddit after having problems with A23 from the website.
 
Upvote 0 Downvote
Alyred said:
If you're not using a PCIe v5 GPU, A25 works pretty well. A26 gives some more PCIe v5 options but unknown what else it really adds - It would be neat if @Svet could post a short changelog with new updates for us to see what to look for and focus our testing on, but I'm glad for the betas either way.
I've had a few stability issues with A26, but most of them seem to have resolved since pushing hard on the memory training and getting it to finally "reset" and clear out anything still in there. RAM seems to be running stable now again, though with a bit less extreme timings than A25 afforded. Haven't had it drop a RAM socket in a while and just passed 2 Memtest86 4-pass runs as well as a full 6.5 hour gold OCCT stability run with A26 and fairly tight timings starting with the Hynix 24GB optimized timings as base and going from there. Shutting down that test bench for a few hours to let it cool and then will run another one of those Gold (no GPU) OCCT tests.
For most folks, A26 should be a significant improvement over A23 either way, and should be posted in the "official" section for now, I've sent a LOT of folks to the A26 Beta posts from Reddit after having problems with A23 from the website.
Click to expand...
What are the usual problems with the official A23?
I'm running it and looks okay for me. Sometimes I can't boot on the first try then I remove some usb from the I/O and it boots fine
 
Upvote 0 Downvote
Im having random shutdowns. System just powers right off, no BSOD. On BIOS A25 right now. No overheating as im watching those temps on my Y70 screen as i game. Only thing i can note is that the BIOS code on the mobo is steady 48. On earlier versions of the BIOS i noticed the Dominator was flashing white RGB which indicates a problem. Will Try A26.
 
Upvote 0 Downvote
kkrott87 said:
What are the usual problems with the official A23?
I'm running it and looks okay for me. Sometimes I can't boot on the first try then I remove some usb from the I/O and it boots fine
Click to expand...
Mostly the boot loops with USB plugged in, but a few other stability issues. M.2_1 random write speed is lower than it should be. A25 improves the M2_1 speeds a bit and A26 improves them more, but still not what they should be.
mrdecrypte15a802e0 said:
Im having random shutdowns. System just powers right off, no BSOD. On BIOS A25 right now. No overheating as im watching those temps on my Y70 screen as i game. Only thing i can note is that the BIOS code on the mobo is steady 48. On earlier versions of the BIOS i noticed the Dominator was flashing white RGB which indicates a problem. Will Try A26.
Click to expand...
XMP settings on this board are rough. Disable XMP and try using the "Optimized Hynix 48GB" setting (right by the XMP enabled in Advanced BIOS) in the BIOS and raise the Mhz to 6200 which is much more stable, and you /might/ be able to train it up to 6400 if you get up to it. You can probably bring your CAS latency down to 30 with those chips with 1.4v, but any 2-rank modules are harder to train than 1-rank. Look up some of Buildazoid's guides on manually tuning your RAM. But at 6600 you're either sacrificing stability running the MEMCLK=UCLK or you're sacrificing latency running MEMCLK/2=UCLK. Those memory kits are going to run HOT with dual-rank, so make sure you have enough air circulation. You might be shutting down due to overheating, check your iCue settings on the RAM to see if you have any shutdown set on the alerts - I think they've removed this in recent iCue versions but there were some having this problem because of an iCUe setting. But if it's getting too hot, it might just plain be that the memory is shutting off and taking the computer with it.
Basically, even though your RAM is 6600 if you're managing to hit that, there's likely so many errors being corrected you're sacrificing speed and the crashes are probably related to that. Try the hynix optimized 6000 with everything at auto, and enable Memory Context Restore so it only re-trains when you change your RAM settings. Training can take a LONG time, 2-3 minutes stuck on code 15.
Also, unless one of your RAID0 SSDs is running on the M2_1 slot, it's not likely you're seeing any real-world speed increases on that as both of those go through the chipset, so any read or write operation has to wait on the single 4x gen 4 bandwidth between the CPU and the M2 slots, then wait for the other M2, each time.
 
Upvote 0 Downvote
Man, going from A23 > A26 was rough, had to redo everything! Settled on something similar, bit less performance but I'll wait till more official versions to tweak more.
Mostly because there seems to be less weird [***CENSORED***] going on, like double reboots/boots etc. Middle USB ports are still a complete bust though, just moved some stuff around and with two plugged in there, my keyboard (connected at top) connects/reconnects till I remove those.
Not a great experience so far for a 500+ board tbh.
 
Upvote 0 Downvote
rand_1152702d9 said:
Man, going from A23 > A26 was rough, had to redo everything! Settled on something similar, bit less performance but I'll wait till more official versions to tweak more.
Mostly because there seems to be less weird [***CENSORED***] going on, like double reboots/boots etc. Middle USB ports are still a complete bust though, just moved some stuff around and with two plugged in there, my keyboard (connected at top) connects/reconnects till I remove those.
Not a great experience so far for a 500+ board tbh.
Click to expand...
Yeah, I hear you. Most involved and difficult build I've ever done.
Always plan on re-doing your settings and timings each BIOS update as you should clear it out and find the new timings for stability, using your old ones as a starting point.
 
Upvote 0 Downvote
Alyred said:
Yeah, I hear you. Most involved and difficult build I've ever done.
Always plan on re-doing your settings and timings each BIOS update as you should clear it out and find the new timings for stability, using your old ones as a starting point.
Click to expand...
Yeah I know you can be forced to do that. Was just surprised on the extent needed here. Especially since the board has been out for a while now.
Also quite used to things just working I guess, coming from an older asus Z690-A intel build, just reload settings each new bios version and completely stable :beerchug:
 
Upvote 0 Downvote
rand_1152702d9 said:
Yeah I know you can be forced to do that. Was just surprised on the extent needed here. Especially since the board has been out for a while now.
Also quite used to things just working I guess, coming from an older asus Z690-A intel build, just reload settings each new bios version and completely stable :beerchug:
Click to expand...
Indeed. I came from an AM4 platform from 2019 and things just mostly worked. Can't rely on XMP at all in AMD's platform right now.
 
Upvote 0 Downvote
Alyred said:
Mostly the boot loops with USB plugged in, but a few other stability issues. M.2_1 random write speed is lower than it should be. A25 improves the M2_1 speeds a bit and A26 improves them more, but still not what they should be.

XMP settings on this board are rough. Disable XMP and try using the "Optimized Hynix 48GB" setting (right by the XMP enabled in Advanced BIOS) in the BIOS and raise the Mhz to 6200 which is much more stable, and you /might/ be able to train it up to 6400 if you get up to it. You can probably bring your CAS latency down to 30 with those chips with 1.4v, but any 2-rank modules are harder to train than 1-rank. Look up some of Buildazoid's guides on manually tuning your RAM. But at 6600 you're either sacrificing stability running the MEMCLK=UCLK or you're sacrificing latency running MEMCLK/2=UCLK. Those memory kits are going to run HOT with dual-rank, so make sure you have enough air circulation. You might be shutting down due to overheating, check your iCue settings on the RAM to see if you have any shutdown set on the alerts - I think they've removed this in recent iCue versions but there were some having this problem because of an iCUe setting. But if it's getting too hot, it might just plain be that the memory is shutting off and taking the computer with it.
Basically, even though your RAM is 6600 if you're managing to hit that, there's likely so many errors being corrected you're sacrificing speed and the crashes are probably related to that. Try the hynix optimized 6000 with everything at auto, and enable Memory Context Restore so it only re-trains when you change your RAM settings. Training can take a LONG time, 2-3 minutes stuck on code 15.
Also, unless one of your RAID0 SSDs is running on the M2_1 slot, it's not likely you're seeing any real-world speed increases on that as both of those go through the chipset, so any read or write operation has to wait on the single 4x gen 4 bandwidth between the CPU and the M2 slots, then wait for the other M2, each time.
Click to expand...
Thanks for the tips will look into the RAM tuning. I have flashed to A26 to see how things go. As for the RAID its just a windows RAID and those old drives are just for secondary off-storage. I dont keep anything on it that is used in processing like Games etc. Purely for photos and that sort of thing.
 
Upvote 0 Downvote
Hello @Svet ,

I tried the 1A26 BIOS and stuttering like hell on the third boot of the week-end like I had with 1A21 and 1A23 BIOS. (BIOS 1A1A OK)
I took a GPU-Z screen to show that PCI express is stuck on 1.1 mode even in the render test and the maximum is 3.0 instead of 4.0.
A26StutteringGPU-Z.png

I reboot --> OK.

Thanks.
 
Last edited:
Upvote 0 Downvote
Guys, sorry to hear that you're experiencing so many issues. I have version 1A26 installed and I haven't encountered any problems with it.
Games run smoothly (except for MSFS 2024) Cyberpunk, Elite Dangerous, The Witcher, nothing lags or stutters - Windows 10 Pro is also working fine.
 
Upvote 0 Downvote
Back
Top