MSI BIOS Update to address Intel Raptor Lake Instability - enforce Intel standard power limits?

jlkoras12df02b8

New member
Joined
Apr 23, 2024
Messages
4
Will MSI be issuing a BIOS update yo address the ongoing Raptor Lake instability issues. ASUS has already issued a beta update allowing for a enforcement to standard Raptor Lake power limits... Will MSI be following suit? I am currently sitting with a Z790 Godlike with Intel 14900K and experiencing issues regardless of what I do to tweak the BIOS. While some tweaks make situation better it never seems to resolve without drastic power reductions. Can we just get a BIOS with enforced standards and I'll determine my overclocks, is update in the works? If there is already a current tread on this then I apologize and ask moderator to move.
 
So basicall Intel was just waiting for the community to find a solution and then gave an official response. Good job. But why is it taking so long? Is it so difficult to correct a value in a microcode? We can literally do it by hand and just set the default 512A to the recommended 307A
 
So basicall Intel was just waiting for the community to find a solution and then gave an official response. Good job. But why is it taking so long? Is it so difficult to correct a value in a microcode? We can literally do it by hand and just set the default 512A to the recommended 307A
It seems surprisingly the voltage issue is with near-idle voltages

Meaning that its not like its just a bug that needs to be fixed, it sounds like they have to re-evaluate and re-write the power specs for the chip
 
I posted my response to Intel’s ”announcement“ in this thread https://forum-en.msi.com/index.php?...wifi-undervolt-experience.399899/post-2271948 if anyone is interested in my thoughts. I think there are actionable items in that thread that relate to this whole instability issue. I doubt this new microcode update will fix things without some kind of action up by the user, like undoing settings that had previously worked.
 
Last edited:
I have a MPG Z790 EDGE WIFI and a 14900ks and tried the 7D91vHC3(Beta version) ran great but had my ecores on my 14900ks running 1.45v anytime they were @ 4.5ghz. Seemed a little high. Rolled back to 7D91vHB things are better there for the time being (not riding the 1.45v when @ 4.5ghz anymore. Was hoping MSI would have released an update that is non beta by now but we wait...
 
I had installed the new BIOS with the new micro code (125) for the Z790 GAMING PLUS WIFI with the i7-14700k on board.
After a couple of days I reverted back to the previous BIOS version “7E06vH52(Beta version)”, because even with the same BIOS settings, the CPU temperatures had increased about 10° under heavy load.
In my case, I think it is better to wait for the “final fix” from Intel before upgrading to a new BIOS.
 

Attachments

  • Cinebench_07-24-2024_01- New 125  BIOS.jpg
    Cinebench_07-24-2024_01- New 125 BIOS.jpg
    929.4 KB · Views: 116
  • Cinebench_07-25-2024_01-beta bios h52.jpg
    Cinebench_07-25-2024_01-beta bios h52.jpg
    931 KB · Views: 103
Does anyone know what at stock voltage settings the 14900ks E-Core voltage "should be around @ 4.5ghz"? I have been trying to find something but everything I seem to find is related to P-Cores. My chip is stable just trying to keep it that way :).
 
I had installed the new BIOS with the new micro code (125) for the Z790 GAMING PLUS WIFI with the i7-14700k on board.
After a couple of days I reverted back to the previous BIOS version “7E06vH52(Beta version)”, because even with the same BIOS settings, the CPU temperatures had increased about 10° under heavy load.
In my case, I think it is better to wait for the “final fix” from Intel before upgrading to a new BIOS.
Pretty significant diff there. I may try this one myself.
 
I had installed the new BIOS with the new micro code (125) for the Z790 GAMING PLUS WIFI with the i7-14700k on board.
After a couple of days I reverted back to the previous BIOS version “7E06vH52(Beta version)”, because even with the same BIOS settings, the CPU temperatures had increased about 10° under heavy load.
In my case, I think it is better to wait for the “final fix” from Intel before upgrading to a new BIOS.
Wow, this actually reveals my high temps aftter the new bios install with my 14700K... I thought it was one of the reasons but couldn't check... Thanx
 
I thought it was one of the reasons but couldn't check
Pretty significant diff there. I may try this one myself.
I'm glad if it can help in some way.
Maybe we had the same problem, however, we also have to consider that we have different motherboards, as well as other components.
Anyway...
Personally, after installing and setting up the new BIOS, I did a test with Cinebench r23 and repeated it three or four times.
For each test, I noticed the same thing, after a few minutes the temperatures were around 76c (as with the old BIOS), but this time towards the end of the test, Cinebench r23 would freeze for a brief moment, the mouse cursor spinning, and suddenly the temperatures jumped to 86c, even 90c in one test.
So I went back to the previous BIOS, repeated the test, and this time there was no abnormal behavior.
Maybe it was just a coincidence? some program open in the background that interfered with the test? that I can't know, but for now, to be more relaxed, I'm staying with the old BIOS.
 
Last edited:
I had installed the new BIOS with the new micro code (125) for the Z790 GAMING PLUS WIFI with the i7-14700k on board.
After a couple of days I reverted back to the previous BIOS version “7E06vH52(Beta version)”, because even with the same BIOS settings, the CPU temperatures had increased about 10° under heavy load.
In my case, I think it is better to wait for the “final fix” from Intel before upgrading to a new BIOS.

You should have checked what "CPU Lite Load" mode the BIOS versions default to.

1) PRO Z790-A MAX WIFI on BIOS 7E07vM52(Beta version): 10-20° hotter under full load
2) Z790 TOMAHAWK MAX WIFI on BIOS 7E25vA62(Beta version): "CPU Lite Load" on Mode 22 by default

Now, these are the BIOS versions with the changelog "Replaced MSI's own tweaked system power settings with added Intel Default Settings, and users can still optimize system performance with alternatives from MSI." and no explicit mention of "μCode version 0x125", so this part of the equasion doesn't quite go together.

However, it does seem that they're currently trying to pull a bit of a Gigabyte move, meaning, increasing the V/F curve in an attempt to counteract any instability (as we know, the two main ways to gain more stability are to lower the frequencies and to increase the voltage). Only problem is, we all know Intel finally came out of the woodwork two days ago, explaining "exposure to elevated voltages" is a "key element" in all this. So if MSI really tried to raise the voltages a bit as some kind of improvement for this situation, it's might be entirely the wrong thing to do.

In the past, especially some Gigabyte BIOSes set the AC loadline too low, which led to instability at stock settings, something mentioned here after the third quote. But if they really crank up CPU Lite Load more by default now (i don't know this for sure, only observing a couple cases where it might have happened), then the pendulum swings too much in the other direction. They need to find a happy medium. Those mid-August BIOS updates can't come soon enough.
 
citay, This thing looks really complicated, it looks a bit like a short blanket, if you pull it to one side you leave that other side uncovered.
Anyway, right after installing the new BIOS I had checked the default CPU Lite Load settings.
With both the new BIOS and the old one, after applying Intel's profile and rebooting the PC, BIOS automatically sets The CPU Lite Load to 16.
However, I set this manually to 7 as I had it on old BIOS. All other settings are also the same in both BIOS versions.
I attach the settings I set in the BIOS as a screenshot.
 

Attachments

  • MSI_SnapShot.jpg
    MSI_SnapShot.jpg
    655.7 KB · Views: 156
Back on the last non beta version7D91vHB (for now) here with my MPG Z790 EDGE WIFI - LL defaults to 9 with that version for me. I manually set it to 8.
Hoping someone can chime in with more info on the e-core voltage for the the 14900ks on this board and what would be considered a normal range when not overclocking at all. I cant seem to find anything on it.
 
Last edited:
citay, This thing looks really complicated, it looks a bit like a short blanket, if you pull it to one side you leave that other side uncovered.
Anyway, right after installing the new BIOS I had checked the default CPU Lite Load settings.
With both the new BIOS and the old one, after applying Intel's profile and rebooting the PC, BIOS automatically sets The CPU Lite Load to 16.
However, I set this manually to 7 as I had it on old BIOS. All other settings are also the same in both BIOS versions.
I attach the settings I set in the BIOS as a screenshot.
"The CPU Lite Load to 16." same for me on the new beta bios. Wich = excessive voltage and heat for sure in my case with the 14900KS. I am back on the latest non beta bios -0.050 adaptive offset from a voltage perspective and Lite Load set to Auto (which is 11 greyed out beside the setting) (these are my only changes tuning wise). Seeing high 38k in cinebench r23 and 32452'ish in timespy with my TUF gaming 4090. Max voltage on any core is between 1.42 and 1.44 during a timespy run (system specs below). WIth LL set to 10 I see 1.32 to 1.42 during a timespy run. I want to run the 0x125 code but just not comfortable with the additional heat... I only game and run 3dmark here and there and cinebench to make sure things look good when I change a bios setting so this is fine as long as it is stable this way :).
CPU 14900KS (Stock clocks) (Stock power limits 253/253/307/56s and MCE off)
CPU Cooler - EK Nucleus AIO CR360 Lux D-RGB - Top mounted Radiator (fans set to exhaust)
MPG Z790 EDGE WIFI (Bios rev 7D91vHB)
Gskill DDR 5 F5-6000J3636F16 x2 32 GiB Kit XMP Enabled - performance extension selected
ASUS Tuff Gaming OC 4090
Seasonic Vertex GX-1000 (1000watt) PSU
Case O11 EVO RGB
Fan config
-3 120mm exhaust up top (mounted to AIO)
-3 120mm side intake
-3 140mm bottom intake
-1 120mm rear exhaust
 
Last edited:
This is what continues to confuse me. Why do some people have more issues with the 0x125 microcode versus others? BuildZoid demonstrated in his latest video that average voltages dropped with the new BIOS release (albeit on a Gigabyte motherboard) and that was with a 14900K, similar to your CPU. But you are not the first person to see higher temps with the new code. Now, there are obviously many BIOS settings to tweak and twiddle that affect voltages and therefore temperature. In fact, it looks to me like there are certain combinations that work well together and others that do not. And unfortunately, unlike in the past, those harmonious settings appear to be significantly influenced by which BIOS you are on, and what defaults the motherboard manufacturer chose for you. In the MSI world, it is very frustrating that a critical setting like Load Line Calibration would not show you the current value corresponding to the Auto setting. Many other settings show you this info. It has become obvious that the two most important settings that affect voltages (and temps) are CPU Lite Load (AC_LL/DC_LL) and LLC. If they are tuned properly to work together, you get a good base to work from; manual voltage offsets are then the cherry on the top. And then that just leaves you the dialing in of PL1/PL2 to match your cooling solution and your personal appetite for higher temps. This shouldn’t be so mysterious, but both Intel and motherboard manufacturers have muddied the waters so much that it’s easy to screw up the settings. There’s also anecdotal evidence that you get inconsistent results when making BIOS changes, almost like the microcode has bugs in it. That’s the best summation of where my mind is at present. It’s hard to give generalized advice when everyone has a different experience. We‘re all now pioneers and pathfinders.
 
Last edited:
As a complete novice to this sort of things, even though I am a 40 year IT veteran as a DEV, I am still so baffled as to what I should be looking for to see if I have, or will have, any issues.

I have an i5-14600K on a PRO B760M-A WIFI DDR4. When the system first arrived, I ran Cinebench R23 and it thermal throttled. Using help from these forums it turned out that that board's PL1 was being auto-set to 185w, whereas Intel's docs say 125w. I changed the setting and ran the same tests and this time things were a little bit more controlled. Also changed the NZXT air-cooler for a Corsair 240 AIO.

Watching Jay's videos about checking BIOS updates, info is showing E7D99IMS.1A0 with a build date of 03/21/2024 and looking at MSI download centre, this appears to be the latest.

So, my question is, can someone explain, in layman's terms, what should I be looking at, in whatever various monitoring tools, to check that my system is behaving.

Thanks.
 
So, my question is, can someone explain, in layman's terms, what should I be looking at, in whatever various monitoring tools, to check that my system is behaving.

BIOS update tends to be a good idea at the moment. I'd do that first. I then recommend setting power limits, mainly to protect your cooling and to prevent thermal throttling, and later also lowering the voltage to what your individual CPU really needs for full stability. I have described the entire procedure here. In that thread, the user has a high-end AIO, so he can set high power limits, but for a 240mm AIO, you could try 200W as a starting point.

First you'd set power limits roughly according to what you think your cooling can get rid of - even for long periods of full load - while still keeping the CPU temperature within the 80°C range. Once you set those "test" power limits in the BIOS (explained in the link), you check the resulting temperatures with Cinebench, which creates fully multithreaded conventional load (the highest load that does not come from a stress testing tool like Prime95). If they're mid-80°C, perfect. Above 90°C, you should reduce the limits, below 80°C you can raise the limits if you want, of course it also depends on the noise you want to tolerate, that has to do with the fan curves. I recommend the highest point of the curve (almost full or full fan speed) at 85° or 90°C. I would try to stay away from the 90°C range CPU temperatures under full load, because that slowly enters thermal throttling territory. It is not good to rely on that, plus you want to have some headroom for higher ambient temperatures.

Then, to reduce the CPU's power draw in all load states, you could look at lowering the setting "CPU Lite Load" to your CPU's needs. Again, i explain that in the thread, also follow the links in there. No need to change the power limits from what you determined though, because those limits only have to do with your cooling capabilities, they don't change once you determined what your cooling can handle. But when you make the CPU draw a bit less power, it will also power-limit-throttle less under full load (if the CPU model natively wants to draw more power than where you set the limits at), and it can boost the clocks a bit higher. This is the beauty of optimizing CPU Lite Load: When done correctly, it will not lower performance, it will actually increase it. Because within the power limits you set, when there's less voltage used for a given frequency, it can then boost higher again. But it has to be checked for stability, and i recommend raising the CPU Lite Load mode one higher than the lowest stable mode for extra stability headroom.
 
I have all my curves (including the AiO) set to the below.
45%/28c 55%/45c 75%/60c 100%/70c
Pump 100% 24/7

For me a cinebench r23 multicore run I see a max temp of 68c (max wattage pull of 220'ish) last run score was 38720, in 3dmark the gathering system info I see 90'ish but only then (during the run Low 50's to low to mid 60's) less when I am more aggressive with the Lite Load settings..
 
You could allow a bit higher temps than 70°C before putting the fan curves to 100%, unless you don't mind the noise. For the CPU fan and maybe the rear exhaust fan (with a conventional fan setup) i may put them to 100% speed at 90°C, but for other case fans, depending on their maximum RPM, if you put them all to 100% at 90°C (let alone below that temperature) and your CPU actually gets that hot, then it can become too much of a hurricane sound. Then i'd rather set the power limits a bit lower until it stays more manageable.

That CB23 score is impressive, and indeed, with a 14900KS that you have... how much lower can you set CPU Lite Load before there's any instability, have you tested it?

Looking at your system from here, it would make sense that you can really make the CPU cooler struggle once that GeForce 4090 goes to work as well. That's an additonal 450W worth of heat dumped into the case.
 
You could allow a bit higher temps than 70°C before putting the fan curves to 100%, unless you don't mind the noise. For the CPU fan and maybe the rear exhaust fan (with a conventional fan setup) i may put them to 100% speed at 90°C, but for other case fans, depending on their maximum RPM, if you put them all to 100% at 90°C (let alone below that temperature) and your CPU actually gets that hot, then it can become too much of a hurricane sound. Then i'd rather set the power limits a bit lower until it stays more manageable.

That CB23 score is impressive, and indeed, with a 14900KS that you have... how much lower can you set CPU Lite Load before there's any instability, have you tested it?

Looking at your system from here, it would make sense that you can really make the CPU cooler struggle once that GeForce 4090 goes to work as well. That's an additonal 450W worth of heat dumped into the case.
Trying to battle that inital bump in temps running 3dmark lol its so quick though by the time the fans ramp up its back to normal #smh. Finding if I go lower than 9 I start getting a whea error here and there and the score drops as well. I just loaded back up the latest beta bios and even with a -0.050 adaptive offset on the cpu when I enable intel default for lite load I saw 1.54v (I think) (never see that on the last final rev), if I set LL to 9 my cinebench r23 mc score was cut almost in half. I started stepped up stopping at 12 and did a -0.065uv and saw a high 37k in cinebench r23 and just ran a Timespy run which was 32411voltage dropped to a max of 1.457 that run from the 1.54'ish I saw with LL set to intel default (I think this is 16 btw)

CPU hovers at 5.9 alot with it setup like this
1722116617554.png

HWMON info from that run
1722116646038.png
 
Last edited:
Back
Top