MSI PRO Z790-P Wifi won't boot after BIOS update

[soul.cai157202db]

Hi guys, how are you?

I'm having a huge problema with my MSI motherboard after the update of bios. I got the last version of it directly from the MSI website, all the process runned with no problems. When the flash bios finished and computer restarted, a lot os bsod's popped, with different stop code messages. I counted at least 5 or 6.

I tried a lot of things, even searching in here. Another bios version, clear cmos, switched the RAM to another slot, i even tried another video card and nothing.

The Windows don't boot and when a try to boot with a pen drive for a fresh windows installation, the bsod shows up. I simply cannot pass beyond the MSI logo screen. In the rare times that happen, few seconds later the bsod shows up.

You guys have any idea of what i can do?

Follow my setup below:

- Core i7 14700K

- MSI PRO Z790-P Wifi

- 2 x 16gb Kingston Fury

- Corsair HX1000w

- GeForce RTX 3070 ti

 

[soul.cai157202db]

It would appear so. And like i said, the possible scenario is, you ran an old BIOS/microcode (causing voltage spikes) for months, you also might have (inadvertently?) turned on some overclocking, which made the CPU run at 1.6V and very high temperatures (even with low load). But despite all the downsides this caused (high power draw, high temperatures, thermal throttling...), it effectively could have masked any stability problems from CPU deterioration. I guess with 1.6V, you can make almost anything stable again. But of course this is no way for a CPU to run, even the performance will suffer because it will start to enter thermal throttling very quickly.

Now in the new BIOS, for the first time the CPU has to run with default settings. Despite MSI having cranked up the default mode for CPU Lite Load now, this still might not be nearly enough for your degraded CPU anymore. If the above assumptions aren't totally off, then your CPU might be ripe for a replacement by Intel.

By the way, did you build this PC yourself, or did someone / some company build it for you? Maybe they have implemented some weird overclocking when they initially set this up. Which already didn't work well before (from the voltages and temperatures you mentioned), and a 14700K is already a monster as it is, if you try to overclock it, it will usually only make everything worse. But now, coupled with the late BIOS update (running too long with voltage spikes), it might have caused extreme deterioration in the end.

To be clear, the late BIOS update is not your fault, they can't expect every Intel 13th/14th gen user to follow all the news about any issues. This is Intel's fault, and if it's really as assumed here, then they have to replace the CPU for you.

Of course, ideally you would test with a different CPU (any 12th-14th gen, even a Celeron) to confirm that the CPU is the culprit.

It was me, yeah!

This mobo with the processor was sended to us by intel via partnership with my channel on youtube.

I mounted everything and don't changed nothing in the bios. Since the first momento i turn the computer the temperatures was too high. After sometime, i did a undervolt, the way i explained before until now, when i updated the bios. I have this setup for about a year.

With all this, and your explanation, i think the motherboard isn't the problem here. I wish i had another processor LGA1700 to try, but i have only a core i5 10gen here. I will contact them during this week to see what we can do.

If they send to me another 14700k, with this mobo and his new bios, probably (if the mobo is not damaged) it will run properly, right?

 

[citay]

This mobo with the processor was sended to us by intel via partnership with my channel on youtube.

I mounted everything and don't changed nothing in the bios. Since the first momento i turn the computer the temperatures was too high.

Ok, interesting. There should've been no problem with your cooler, because with the new BIOS, the CPU temperature in the BIOS is normal. If there was a problem with the cooler (not mounted correctly, air bubbles or whatever else), then right now the temperature in the BIOS should already be 50, 60°C or more.

The board came brand new, unopened bag inside the box, nothing set already? Did you do a BIOS update right away (it's usually the first thing i do when i get a new board, because the factory BIOS will already be a couple months outdated)? I guess not, eh.

While the 14700K can be too much for a lot of systems even when it runs at stock settings, they would never result in a near-constant 1.6V and high temperatures when barely doing anything. So already something must've been wrong from the beginning. I'm just not sure what.

If they send to me another 14700k, with this mobo and his new bios, probably (if the mobo is not damaged) it will run properly, right?

Yep. There will be no voltage spikes on the latest BIOS/microcode version, it will be fully stable once more (if your current CPU was degraded), and you should even be able to lower the CPU Lite Load mode for some nice benefits in how it runs.

 

[soul.cai157202db]

It would appear so. And like i said, the possible scenario is, you ran an old BIOS/microcode (causing voltage spikes) for months, you also might have (inadvertently?) turned on some overclocking, which made the CPU run at 1.6V and very high temperatures (even with low load). But despite all the downsides this caused (high power draw, high temperatures, thermal throttling...), it effectively could have masked any stability problems from CPU deterioration. I guess with 1.6V, you can make almost anything stable again. But of course this is no way for a CPU to run, even the performance will suffer because it will start to enter thermal throttling very quickly.

Now in the new BIOS, for the first time the CPU has to run with default settings. Despite MSI having cranked up the default mode for CPU Lite Load now, this still might not be nearly enough for your degraded CPU anymore. If the above assumptions aren't totally off, then your CPU might be ripe for a replacement by Intel.

By the way, did you build this PC yourself, or did someone / some company build it for you? Maybe they have implemented some weird overclocking when they initially set this up. Which already didn't work well before (from the voltages and temperatures you mentioned), and a 14700K is already a monster as it is, if you try to overclock it, it will usually only make everything worse. But now, coupled with the late BIOS update (running too long with voltage spikes), it might have caused extreme deterioration in the end.

To be clear, the late BIOS update is not your fault, they can't expect every Intel 13th/14th gen user to follow all the news about any issues. This is Intel's fault, and if it's really as assumed here, then they have to replace the CPU for you.

Of course, ideally you would test with a different CPU (any 12th-14th gen, even a Celeron) to confirm that the CPU is the culprit.

Yo man, how are you?
I got a Core i5 14400F processor from the 14th generation and was able to test it using the MSI motherboard's BIOS update.

It worked perfectly. I was able to install Windows, and everything is fine. In fact, my Core i7 14700k has degraded.

Now I would like some advice from you. I work with video editing, rendering, recording and live streaming in 4K, as well as games, of course. I want to get a processor that is at the same level as my previous one or better, and I was thinking about getting a 14700k again, or a Core i9 14900k. What do you recommend?

The other configurations of my PC are in the first post.

 

[citay]

It worked perfectly. I was able to install Windows, and everything is fine. In fact, my Core i7 14700k has degraded.

Interesting. But yeah, it was starting to be likely at this point.

and I was thinking about getting a 14700k again, or a Core i9 14900k. What do you recommend?

For me, the i7 has historically always been the better overall choice, the better allrounder. The 14900K in particular is just a bit too crazy. They should've never gone to such an extreme in the first place. At a bit lower frequencies (and thus voltages), it would've been better, but they always had to push the limits to look ok against AMD in the launch reviews. The i7 is easier to tame, the i9 will also have to be tamed (by using my guide for example), but it's crazier. So you have to decide if shaving off a couple minutes from your rendering time is worth it.

 

[soul.cai157202db]

Hey guys, good news! I managed to get a i9 14900KS from intel that its working now in my MSI motherboard.

But i have a question. It is normal for that kind of processor to have high peaks of temperature, right?

For exemple, here now, navigating at chrome, or in my computer, the temperature remains stable at 34, even 40. I'm using the water cooler Lian Li Ga II Trinity SL-INF 360.

But, when i open some programs, as steam, for example, he imediatly do a jump for 50, 60, or even 70, just for a few seconds and stabelize again.

I did some tests with programs like OBS, Adobe Premiere, or games and some cases the processor reached 90, 95 or 100 degrees for a few seconds again, and stabelize in 70, 75 with a few peaks of higher, but never constant.

Just to add! I just did a new test running resident evil 4 on steam and he remain stable at 62, 65, with some peaks of 85, 90 for about 2 seconds, returning to 60.

Just when i force the processor somehow, opening things, rendering, processing games and stuff, that he managed to get this high peaks for a few seconds. Even now, writing this post, he is stable at 35, 36, but for a second or two, the temperature jumps for 50, 52 and go back.

Is that normal, right? Because of the number of threads and cores that processor have. What you guys think?

 

[citay]

Hey guys, good news! I managed to get a i9 14900KS from intel that its working now in my MSI motherboard.

But i have a question. It is normal for that kind of processor to have high peaks of temperature, right?

The 14900KS is completely bonkers. It's the same CPU as a 14900K, but with 3.5% = 200 MHz higher peak turbo clocks than the -K, usually sold at a 20-25% higher price. This bump in frequencies makes the -KS the "least efficient CPU model on the market", because they had to use even higher voltages for this, when the 14900K was already crazy. This goes back to what i posted here.

So for every kind of load, this will tend to have a higher power draw than any other CPU. The temperature is only the outcome of the heat, which is the outcome of the power draw, which is the outcome of VCore and clocks. If you check in HWinfo, under "CPU Package Power" (if set to English), you will see the power draw of the CPU. This will get very high with any kind of load, compared to other CPUs.

Just to add! I just did a new test running resident evil 4 on steam and he remain stable at 62, 65, with some peaks of 85, 90 for about 2 seconds, returning to 60.

In games, the CPU is around half load (sometimes more, sometimes much less, depending on the game), only the GPU is at full load. You will stress the CPU much more with things like rendering.

What you see is expected of a 14900KS, especially when you haven't tried to tune it according to my guide yet. This CPU probably should have never existed, like i said in the last paragraph here. The 14900KS is close to uncoolable with maxed out power limits, maybe unless you have a very powerful custom loop. If it's even just a regular 360mm AIO, you will have to work with power limits already (not to mention an air cooler, however nice it is).

 

[soul.cai157202db]

The 14900KS is completely bonkers. It's the same CPU as a 14900K, but with 3.5% = 200 MHz higher peak turbo clocks than the -K, usually sold at a 20-25% higher price. This bump in frequencies makes the -KS the "least efficient CPU model on the market", because they had to use even higher voltages for this, when the 14900K was already crazy. This goes back to what i posted here.

So for every kind of load, this will tend to have a higher power draw than any other CPU. The temperature is only the outcome of the heat, which is the outcome of the power draw, which is the outcome of VCore and clocks. If you check in HWinfo, under "CPU Package Power" (if set to English), you will see the power draw of the CPU. This will get very high with any kind of load, compared to other CPUs.



In games, the CPU is around half load (sometimes more, sometimes much less, depending on the game), only the GPU is at full load. You will stress the CPU much more with things like rendering.

What you see is expected of a 14900KS, especially when you haven't tried to tune it according to my guide yet. This CPU probably should have never existed, like i said in the last paragraph here. The 14900KS is close to uncoolable with maxed out power limits, maybe unless you have a very powerful custom loop. If it's even just a regular 360mm AIO, you will have to work with power limits already (not to mention an air cooler, however nice it is).

Thanks for the response! I was surprised when Intel sent to me this one. At least i didn't spend money on it, it was a donation. It is very expensive and don't worth the money, like you said. hehe

And yeah, i use HWInfo for monitor the temperature. I installed a new game of the steam a few minutes ago, and the processor reached those highest peaks again, see the image:

Right now its 31 degrees, but when i was installing the game, he peaked 90 and 100 degrees, and stabelized on 70, 75 until the game was finished. Steam app consume this much of the processor?

I will get your guide again and collect insights to made adjustments here. In my case, as long as the processor doesn't stay at constantly high temperatures, everything is fine for this model, right?

 

[citay]

Thanks for the response! I was surprised when Intel sent to me this one. At least i didn't spend money on it, it was a donation. It is very expensive and don't worth the money, like you said. hehe

Yes, it's surprising they sent you a 14900KS in return, but since it's a monster, you have to put it on a leash.

And yeah, i use HWInfo for monitor the temperature. I installed a new game of the steam a few minutes ago, and the processor reached those highest peaks again, see the image:

The temperature, as i said, it's only the outcome of the heat which comes from the power draw. So, set your HWinfo to English, and then look further down, for "CPU Package Power" in Watts. This will explain better what it going on, when looked at together with the temperature.

You basically have to go by my guide again. Step 1), set the power limits (using Cinebench for testing) so the CPU never goes above 90°C in HWinfo. Step 2), try to lower the mode for "CPU Lite Load" and do stability testing with OCCT, Prime95 or so. Find a lower mode, one step above the lowest stable one. Usually, on a 14900KS, you may not be able to lower the mode as much as for "lesser" CPU models, because that CPU uses such high frequencies, so they had to push it to the extreme from factory.

In some cases, it's actually better to turn the 14900KS into a 14900K again (frequency-wise), because that will take a lot of stress out of the CPU. With 200 MHz less for all cores (per-core underclock), the CPU will still perform excellent, but not need such high voltages anymore (so you can lower the CPU Lite Load mode better). If you want, i can tell you how to do that, this is probably how i would do it if i had this CPU, because the KS was pushed too far. But you can also just try to go by my guide first, who knows, maybe you can already find a mode that's low enough and still stable. But you also have to set power limits, no doubt about this.

 

[soul.cai157202db]

Yes, it's surprising they sent you a 14900KS in return, but since it's a monster, you have to put it on a leash.



The temperature, as i said, it's only the outcome of the heat which comes from the power draw. So, set your HWinfo to English, and then look further down, for "CPU Package Power" in Watts. This will explain better what it going on, when looked at together with the temperature.

You basically have to go by my guide again. Step 1), set the power limits (using Cinebench for testing) so the CPU never goes above 90°C in HWinfo. Step 2), try to lower the mode for "CPU Lite Load" and do stability testing with OCCT, Prime95 or so. Find a lower mode, one step above the lowest stable one. Usually, on a 14900KS, you may not be able to lower the mode as much as for "lesser" CPU models, because that CPU uses such high frequencies, so they had to push it to the extreme from factory.

In some cases, it's actually better to turn the 14900KS into a 14900K again (frequency-wise), because that will take a lot of stress out of the CPU. With 200 MHz less for all cores (per-core underclock), the CPU will still perform excellent, but not need such high voltages anymore (so you can lower the CPU Lite Load mode better). If you want, i can tell you how to do that, this is probably how i would do it if i had this CPU, because the KS was pushed too far. But you can also just try to go by my guide first, who knows, maybe you can already find a mode that's low enough and still stable. But you also have to set power limits, no doubt about this.

If it's not too much to ask, since you've been helping me a lot since my first post, could you tell me what you have in mind to stabilize the use of KS specifically?

Turning it into a K, as you said.

That will already help a lot. hehe

 

[citay]

First, try to see what's going on with the power draw, see what kind of power limits you have set now, and how you can adjust them to where the CPU stays below 90°C. This we need anyway. Because if your cooling is set up properly already, the cooling capabilities won't really change, so you set good power limits once, and then they stay like that. This should normally be the first step.

You can show me a screenshot of HWinfo (in English) after a 10-minute run of Cinebench R23, when you have already adjusted the power limits. If you are not sure what limits to set, then choose something random like 230W and see how high the temperatures get with full load from Cinebench. Then we can see about the next steps.

 

[soul.cai157202db]

First, try to see what's going on with the power draw, see what kind of power limits you have set now, and how you can adjust them to where the CPU stays below 90°C. This we need anyway. Because if your cooling is set up properly already, the cooling capabilities won't really change, so you set good power limits once, and then they stay like that. This should normally be the first step.

You can show me a screenshot of HWinfo (in English) after a 10-minute run of Cinebench R23, when you have already adjusted the power limits. If you are not sure what limits to set, then choose something random like 230W and see how high the temperatures get with full load from Cinebench. Then we can see about the next steps.

While you responded here, i took some pictures of my bios configurations. The long and short power are set to 253w, because of the cooler tuning, that you explain in the guide.

Shoud i set something like 230w for long and short, then test??

 

[soul.cai157202db]

Complementing the last post! I get the image of the test i did earlier installing a steam game and the result was this.

I edited the text putting them in english after i changed here on the hwinfo.

The processor reached highest peaks of temperatures for a short time, when he reached 184w of power, then he stabelized at 65, 70 until the process finish, returning to 40. So we need to test some short and long values below that max, right?

 

[citay]

Please set HWinfo to English like i asked. Most things are in English there anyway.

For the CPU, don't run a Steam game, run Cinebench like i describe in my guide. If the CPU really starts to reach 100°C temperature at 185W already, something is not right with your cooling. If you are lucky, the fan curves are simply not set properly yet. Or maybe it's how it's placed inside your system. Because a 360mm AIO should handle at least up to 220W, otherwise, what's the point. If it's really good, it should even be able to just about handle the full 253W from the default power limits.

That's why i need to see the entire HWinfo screenshot, like i show in my guide:

Everything is interacting with each other. I need to see the power draw, the temperatures, what the fans are doing, and under proper full CPU load (Cinebench). If it goes too 100°C immediately in Cinebench, then you don't have to wait for the full 10 minutes to complete, just take a screenshow of HWinfo, in English, with the window maximized and the sensors expanded to several columns.

 

[soul.cai157202db]

Please set HWinfo to English like i asked. Most things are in English there anyway.

For the CPU, don't run a Steam game, run Cinebench like i describe in my guide. If the CPU really starts to reach 100°C temperature at 185W already, something is not right with your cooling. If you are lucky, the fan curves are simply not set properly yet. Or maybe it's how it's placed inside your system. Because a 360mm AIO should handle at least up to 220W, otherwise, what's the point. If it's really good, it should even be able to just about handle the full 253W from the default power limits.

That's why i need to see the entire HWinfo screenshot, like i show in my guide:

Everything is interacting with each other. I need to see the power draw, the temperatures, what the fans are doing, and under proper full CPU load (Cinebench). If it goes too 100°C immediately in Cinebench, then you don't have to wait for the full 10 minutes to complete, just take a screenshow of HWinfo, in English, with the window maximized and the sensors expanded to several columns.

I just runned the Cinebench R23, follow the HWinfo below. Started stable, but after 5 minutes running the processor reached 95, 100 degrees, and after that time, he remained there. I stoped the execution after 7 to 8 minutes.

 

[citay]

Ok, so 230W power limits are still not enough with the current setup. But we have to look at your fan curves now. See the three fan headers in the middle? On PUMP1 aka PUMP_FAN, that should be your water pump, it's normal for that to run at a steady high RPM. But then, on System1 and System2 (SYS_FAN1 and 2), we have two fans which are also on a fixed RPM. Meaning, no matter if the CPU is at 30°C or 100°C, they don't change their speed. Does this make any sense? No. So you have to set up a fan curve for them.

So before lowering the power limits further, we have to improve the way the fans behave, take care of better airflow and cooling that way. The three fans on the AIO's radiator, where are they connected? Are they controlled in some Windows software? I assume that SYS_FAN1 and 2 are your case fans?

If you have a photo of your system with the side panel off, that might also help, so i see where all the fans are.

 

[soul.cai157202db]

Ok, so 230W power limits are still not enough with the current setup. But we have to look at your fan curves now. See the three fan headers in the middle? On PUMP1 aka PUMP_FAN, that should be your water pump, it's normal for that to run at a steady high RPM. But then, on System1 and System2 (SYS_FAN1 and 2), we have two fans which are also on a fixed RPM. Meaning, no matter if the CPU is at 30°C or 100°C, they don't change their speed. Does this make any sense? No. So you have to set up a fan curve for them.

So before lowering the power limits further, we have to improve the way the fans behave, take care of better airflow and cooling that way. The three fans on the AIO's radiator, where are they connected? Are they controlled in some Windows software? I assume that SYS_FAN1 and 2 are your case fans?

If you have a photo of your system with the side panel off, that might also help, so i see where all the fans are.

Here the photos. All the six fans of my case are connect on this little board:

The fans of the water cooler are plugged in here:

And in the attach are the photos of my desktop, where the fans and the pump are plugged in the motherboard too. I putted some photos of my fan system in the bios too and how the curves are.

About your question, i don't use any software for control the rotation of them.

Just for note, i was searching for the watercooler lian li galahad II trinity performance earlier, that are more efficient for cooling, and i saw a comment of a user saiyng that his 14700k don't pass 70 degrees with it, and i thought, what?

 

[citay]

Here the photos. All the six fans of my case are connect on this little board:

Ok, that just seems to bundle all the case fans into one connector or something. What's your case model?

The fans of the water cooler are plugged in here:

That's Lian Li's fan controller, it should plug in via USB to the motherboard. In order to control the radiator fans, you have to install the Lian Li software: https://lian-li.com/l-connect3/

About your question, i don't use any software for control the rotation of them.

Then they are either on a factory fan curve, or on some steady speed (which would be bad for cooling performance).

And in the attach are the photos of my desktop, where the fans and the pump are plugged in the motherboard too.

No fan curves are being used for the system fans. Put the BIOS in advanced view (F7), enter Hardware Monitor, and create fan curves according to my guide. I explain everything in there, so i won't bother going into detail here.

You have tons of fans and cooling ability, but at the moment it's not utilized very well at all, that's why your cooler is underperforming so badly. Once this is sorted out, it should perform as expected, hopefully.

 

[soul.cai157202db]

Ok, that just seems to bundle all the case fans into one connector or something. What's your case model?



That's Lian Li's fan controller, it should plug in via USB to the motherboard. In order to control the radiator fans, you have to install the Lian Li software: https://lian-li.com/l-connect3/



Then they are either on a factory fan curve, or on some steady speed (which would be bad for cooling performance).



No fan curves are being used for the system fans. Put the BIOS in advanced view (F7), enter Hardware Monitor, and create fan curves according to my guide. I explain everything in there, so i won't bother going into detail here.

You have tons of fans and cooling ability, but at the moment it's not utilized very well at all, that's why your cooler is underperforming so badly. Once this is sorted out, it should perform as expected, hopefully.

My case is the Lian Li O11 Dynamic XL, Full Tower.

I will install the lian li software here and set the curves on the bios. I will sent the images here when i'm done.

 

[soul.cai157202db]

Ok, that just seems to bundle all the case fans into one connector or something. What's your case model?



That's Lian Li's fan controller, it should plug in via USB to the motherboard. In order to control the radiator fans, you have to install the Lian Li software: https://lian-li.com/l-connect3/



Then they are either on a factory fan curve, or on some steady speed (which would be bad for cooling performance).



No fan curves are being used for the system fans. Put the BIOS in advanced view (F7), enter Hardware Monitor, and create fan curves according to my guide. I explain everything in there, so i won't bother going into detail here.

You have tons of fans and cooling ability, but at the moment it's not utilized very well at all, that's why your cooler is underperforming so badly. Once this is sorted out, it should perform as expected, hopefully.

I just made the first adjustments to the curves of my fans and the water cooler pump and the results were very promising, compared to the first time I ran Cinebench.

The settings I made in the BIOS were as follows:

After running Cinebench, these were the results in HWInfo. The most interesting thing is that it only reached 90 degrees after 7 minutes of execution, just for a few seconds, and then went back to 82, 84, until it finished.

Regarding the Lian Li software, I didn't quite understand where to change the fans that are on the radiator. hehehe
Oh, i get it now, is the port1 in the image, sorry! The RPM is couting in there.

 

[citay]

I just made the first adjustments to the curves of my fans and the water cooler pump and the results were very promising, compared to the first time I ran Cinebench.

Yes, you had a lot of unused potential in your cooling, since it was running with the handbrake on. Now it looks much better, only very mild thermal throttling with up to 267W of CPU power draw. I would say, maybe set 230W power limits, then you can totally stay away from >90°C.

About your fan curves, for the low temperatures, you can set even lower percentages, for example at 35°C i have my fans at 20%, 24% and 26% (different fan models). The idea is to get around 400-500 RPM in idle and in the BIOS, if your fans allow for that low speed. This way, it's nice and quiet when the CPU is cool enough. They only need to spin faster when the CPU heats up.

As the next step, you should lower "CPU Lite Load", Mode 16 is too high, i don't like the high voltages you are getting with that (pretty typical of a -KS though). And we can even look at turning it into a -K model. But let's see how far you can get reducing the mode. For this, you need to do stability testing to find the lowest stable mode. Then test the next higher up mode, and if that's fully stable, raise by another step to have stability headroom. As a random example: Mode 8 crashes/errors in stress-tests, so test Mode 9, and if it's stable, set Mode 10. But it doesn't have to be like that, your CPU can still be stable at Mode 5, or it might need Mode 11, who knows. This CPU model is unpredictable.