News Is your Intel Core i9-13900K crashing in games? Your motherboard BIOS settings may be to blame — other high-end Intel CPUs also affected

[JarredWaltonGPU]

You took that current limit and computed voltage using the PL2 power spec, but we know that CPUs can go well beyond PL2, for up to 10 ms. This is known as PL4. For the i9-13900K, PL4 is actually 420W. If you divide 420 W / 307 A, you get 1.368 V.

So, if the current limit works like PL4 (i.e. as an absolute threshold that you cannot go beyond), then that's the power value it should be computed against.

Oh, that's excellent information. I never did really look deeply into this (see: I'm the GPU guy, not CPU), so I was also quite curious about the high current maximum. Because I did the math and couldn't see how the CPU would ever need 500A or whatever the default was. In fact, I was thinking that with a power limit of even 350W, amperage would be well under that level since voltage was above 1.0V.

But that's because I didn't remember the PL4 bit. 1.368V * 307A = 420W. And thus, with my max setting of 325A I'm still hitting higher than 'stock' settings, but probably not exceeding it so badly that my CPU crashes. And of course, limiting the voltage just a bit can pay big dividends as well. 1.3V for example would mean 5% less power, and that might be enough to also prevent a crash.

I still think a lot of this comes down to Intel pushing 'stock' settings much further than it used to, and then motherboard vendors also pushing things just a bit further as well. Back in the day, like 20 years ago, it used to be pretty much a given that Intel had about 20% more headroom in a lot of its chips (for clockspeed). People getting 50% and even 60% overclocks was possible at one point (Celeron 300A being a prime example). These days, it's probably less than 5% on the 13900K and 14900K.

 

[bit_user]

Back in the day, like 20 years ago, it used to be pretty much a given that Intel had about 20% more headroom in a lot of its chips (for clockspeed). People getting 50% and even 60% overclocks was possible at one point (Celeron 300A being a prime example). These days, it's probably less than 5% on the 13900K and 14900K.

I've even heard stories of people overclocking a 386DX-25 by 100% to 50 MHz!

Our first family computer was a PC XT clone, that had a "turbo" mode. The compatible speed was 4.77 MHz, but I think turbo mode would run it at 8 MHz?

It might be cute to put a modern AMD build in a case with a turbo button and somehow use it to control whether the CPU is in PBO or standard mode, though I'm not sure if you can PBO on the fly. Or, since this is an Intel thread, maybe you trigger different power limits, based on the turbo button?

 

[MrWallopy]

How can it not be Intel's fault, if board partners are only doing things Intel considers not to be overclocking or warranty-voiding?

It's Intel's job to place & enforce limits that ensure stable operation of their products. I think we really shouldn't let them off the hook, here.

I both agree and disagree with this.
1) Intel should be holding the board partners accountable. Though it is a free market.
2) If you don't want a board with these features we as consumers could also buy the lower end boards that don't allow OC's OV's etc.

Like another said, Intel does state what their chips can go to. Since it's a free market they can in fact do what ever they want.

On the other hand do we really want Intel to become like Nvidia? Limiting the board partners to keep within their own spec. RIP EVGA
Intel is already on the downslide, and doing something like this would make news and though it would be good for the consumers "news" a.k.a all the youtubers and streamers would put a negative twist on this while receiving a nice little chunk change from AMD. Could potentially be a huge hit to Intel.

In the end there is no true winner. We all lose.

Any who, I just submitted for my RMA as i'm tired of all of my games continuously crashing if i don't set my voltage to 1.6 and turning my pcore ration to x50... And even then it still happens.

 

[MrWallopy]

Its very interesting how these sky-high power limit settings are finally affecting Intel users.

AMD was smart, they put any sort of power limit increases under Precision Boost Overdrive (PBO) and clearly state that it is not stock operation. AMD users have also seen stability problems with PBO, its rare, but it happens. The fix for AMD CPUs is to boost the load line calibration so the vdroop doesn't create instability under higher wattage workloads (even though the voltage is the same).

Hopefully Intel will start putting more clear guidelines on motherboard makers and make sure users understand what is stock operation and what is not.

The problem with AMD is the same, board manufacturers setting no limit out of the box. I buy the higher end boards for more consistent and better vrms and chipsets. Not because i want to push the chip past the stock performance.

My 1700x was alright - Change anything and it was stuck in that state. No power save, stayed solid voltage and clock speed. But gen 1 so ya know yeah.
My two 5800x's, Wanted to offset the voltage to help with temps. NOPE They too get stuck in higher power states, XMP enabled? ooo sorry memory controller issues.
5950x Ran hot, PBO was nice, but got so damn tired of always trying to tweak it every time a new game came out and caused those settings to crash. Yes i could do an overall Offset but anything higher than -.05 would cause a crash on 1 of the cores. FFFSSSS. So now i'm manually tuning each core. 4 of which could hit -.3 amazing! 4 that couldn't get past -.05 and another 8 scattered inbetween.

Even the 7000 series have some form of memory issue.

Why can't we just have chips that work? That you plug in, and they will run at their stock settings. So i don't have to do the work. I bought the higher end crap because it's already higher end. A base clock i9 13900k outperforms the i7 13700k Stock and OC. Good, why do i need to push it further than that? Let me run stock OOB!!!

Let those that want to OC go in and change those settings for themselves.

TLDR; Intel, AMD, and the board partners are all in cahoots to shorten the life span of these chips so we have to keep buying more. Can't be planed obsolescence if it's not the original manufacturer causing your issues, it's the other company's fault.

Sorry for my rant lol. Just so damn tired of spending sometime days when i'm off work trying to play anything but instead testing ram, verifying it's not GPU, ripping apart a water cooling loop.

Maybe it's time to start buying the B series boards again. -_-' Maybe Dell and HP are actually the good guys not giving these options and keeping them limited from the get go. Yuckin sakes...

Also for anyone else, i found this,

https://community.intel.com/t5/Proc...s-Windows-11-with-apps-games-and/td-p/1527297

 

[MrWallopy]

This sounds like a lot of nonsense. I have those errors in almost every UE5 game. That's the common denominator. If the power limit was an issue, I would encounter those errors elsewhere. My PC would never pass a cinebench run or other stress tests where the load is even higher. Maybe downclocking or capping power masks the real issue, but the problem is more likely to be UE5 shader compilation. Perhaps it spins up to many threads on higher end CPUs causing some resource contention.

As noted in the article text, it's not necessarily the power limit itself, but a combination of power limit, current limit, voltage, and/or frequency — one or more of these allows certain CPUs to reach unstable states. Games that use Oodle for data decompression are frequently affected. Oodle has released a statement, basically echoing everything we said, along with some details for Asus, Gigabyte, and MSI motherboards.

It's 100% not a case of simply spinning up too many threads. It's running a workload that happens to push the CPU hard, which can cause overheating, overcurrent, overpower, overvoltage, or some other related problem. Once you limit the clocks, power, voltage, and/or currency to the appropriate levels for your CPU, all issues will go away.

If you look inside of some Intel Forums, One of the "reps" asked them to install the XTU and go to advanced view, Stress Test, and run it for AVX2. It appears to be something wrong with the AVX/AVX2 side of things.
AVX ran for about 2 minutes of the 5 before it failed. AVX2 failed immediately.

They then ask you to go into AITweaker (For aSUS boards) and set SVID to "Intel Fail Safe". And run it agian.

As i did this also it passed. However voltage for me was peaking at 1.68v Heh... NO thanks. I then put in the RMA.

Inside of XTU Advanced Tuning there are options for AVX2 ratio offset and AVX2 Voltage Guardband Scale Factor. As i lowered the Turbo Boost short power max from Unlimited (also inside xtu advanced) my AVX2 Voltage guardband scale factor changed. hmm.

 

[MrWallopy]

Well, i'm not 100% sure on Forum policy for TH here. But i found this forum post on Intels site that started back in sept'23.
They recommended
Download XTU, Go to advanced and Run Stress test for AVX2, if failed go to bios settings and change SVID to Intel Fail Safe, Run again.
I did this failed first time, changed to intel fail safe 1.68v peak, and it passed.

For anyone else wanting to investigate further and tinker with AVX2's ratio offset's or voltage guard band scale factors go on ahead. I have a feeling faulty chip design and something with this instruction set.

Link for more info.
https://community.intel.com/t5/Proc...s-Windows-11-with-apps-games-and/td-p/1527297

 

[bit_user]

2) If you don't want a board with these features we as consumers could also buy the lower end boards that don't allow OC's OV's etc.

...and then you write:

I buy the higher end boards for more consistent and better vrms and chipsets. Not because i want to push the chip past the stock performance

Yes, but now you're contradicting yourself.

I also like to buy a higher-spec board, hoping it'll provide better stability and longevity - not because I want to OC.

 

[TheHerald]

How can it not be Intel's fault, if board partners are only doing things Intel considers not to be overclocking or warranty-voiding?

It's Intel's job to place & enforce limits that ensure stable operation of their products. I think we really shouldn't let them off the hook, here.

Well to be fair, same thing happened with AMD, especially on ASUS (but not just ASUS) motherboards, where they were providing excessive vsoc voltages that led to literally burning both the CPU and the socket.

 

[TheHerald]

Its not the power limits thats just masking the problem forcing a power limit

Its a combination of dumb LLC + AC/DC Load Line + Bad Auto Settings + Non tested SA/ VDDQ / not fully stability tested ram settings
and in many cases people setting a static voltage for something they have no understanding of in many cases

Intel SPEC LLC is 1.1 mOhm which also has to be then reported back to the cpu
messing with this causes staility issues if you dont know what your doing (which most people dont)
AC Loadline depends on Loadline Calibration value and motherboard socket impedance. No one knows what the proper impedance is, and I doubt anyone here has the correct tools to measure it, but it's "probably" between 0.2 to 0.6 milliohms. (So let the vendor set that as its board dependant)

DC Loadline should always be set to the same mohms as LLC (Loadline calibration) value in mOhms however, in order for package power and VID to report close to accurate values (e.g. 1.1 mOhm if LLC Is Standard/Normal/SPEC....)

My 13900K Passes
Karhu + FPU Load 24 hours
Prime 95 Small FFT AVX (440W Load)
y-cruncher
Never Crashes in games
Scores 41300 in Cinebench R23
And has unlimited PL1/PL2
etc etc

and i run a 30mv undervolt on the higher multipliers
I run the cpu
58x (1-3c), 57x (4-5c), 56x (6-7c), 55x (8c)
instead of
58x (1-2c), 55x (8c)
e-cores are also on..

In sumary user/oem error and a nothing burger

With trial and error you can configure it with the sensors from hwinfo. Though I've never managed to on my 14900k, idle voltages were always OFF. On the 12900k the motherboard put the correct ac dc values on it's own, VID and VCORE were always matching both under load and in idle.

 

[Peksha]

It might be cute to put a modern AMD build in a case with a turbo button and somehow use it to control whether the CPU is in PBO or standard mode, though I'm not sure if you can PBO on the fly. Or, since this is an Intel thread, maybe you trigger different power limits, based on the turbo button?

Asus use it as Dynamic OC Switcher (with many triggers).

On topic: I had this problem from day one for 13900ks+gigabyte z790 with default BOIS.
I decided without adjusting the LLC, simply setting:
AC/DC load line=power saving,
LLC=medium,
Adaptive Vcore Legacy auto with offset -0.065V and Power limits=Intel POR (253W). MCE disabled

 

[MrWallopy]

...and then you write:


Yes, but now you're contradicting yourself.

I also like to buy a higher-spec board, hoping it'll provide better stability and longevity - not because I want to OC.

I see both sides of my own argument and will happily contradict myself.... I'm doomed.

Well, i went to Microcenter today and bought a 14900k. CPU temps are already 15* lower than the 13900k were at idle. Also... i noticed this on some of the newer bios that i have not seen before. Or maybe i always overlooked it?? But got a message after installing this new CPU
"New CPU detected"
"If you wish to follow intel guidelines and apply its stock power limits please disable the asus multicore enhancement (MCE) in the bios settings."

 

[JarredWaltonGPU]

I see both sides of my own argument and will happily contradict myself.... I'm doomed.

Well, i went to Microcenter today and bought a 14900k. CPU temps are already 15* lower than the 13900k were at idle. Also... i noticed this on some of the newer bios that i have not seen before. Or maybe i always overlooked it?? But got a message after installing this new CPU
"New CPU detected"
"If you wish to follow intel guidelines and apply its stock power limits please disable the asus multicore enhancement (MCE) in the bios settings."

Yeah, silicon lottery basically. You got a good 14900K, you had a bad 13900K. The latter likely had almost zero headroom left compared to stock settings, so things like MCE created problems.

 

[MrWallopy]

Yeah, silicon lottery basically. You got a good 14900K, you had a bad 13900K. The latter likely had almost zero headroom left compared to stock settings, so things like MCE created problems.

With as many people having this problem overtime, it doesn't seem like lottery, just some bad batch's.
Speaking of, My batch was X246J840 for anyone who wants to compare 😛.

 

[HaninTH]

Are your games crashing left and right? Do you have out of vram errors?

It's not constant, but if i crank the visuals on any game i try, it'll cause all the fans on the system to crank up and crash the system where i have to disconnect power to get it to come back on. It leaves no log entries after doing it either. I just turn down the visual/graphics settings and it will not crash. My GPU is a 3070. I initially thought it might be an underpowered PSU, but i swapped from an i7 7700k on a Z board to the i5-13600k on z790 board with an 800watt seasonic PSU and it does the same thing under the same scenerio. *shrugs*

 

[JarredWaltonGPU]

It's not constant, but if I crank the visuals on any game i try, it'll cause all the fans on the system to crank up and crash the system where I have to disconnect power to get it to come back on. It leaves no log entries after doing it either. I just turn down the visual/graphics settings and it will not crash. My GPU is a 3070. I initially thought it might be an underpowered PSU, but I swapped from an i7 7700k on a Z board to the i5-13600k on z790 board with an 800watt Seasonic PSU and it does the same thing under the same scenario. *shrugs*

That would suggest it's more of a GPU issue than the CPU, if it was happening on both PCs. If you're not the original owner of that 3070, it may also have been pushed hard as a mining GPU for a while, and that can absolutely cause degradation of cooling and other elements to the point where games can crash at more demanding settings.

 

[antych]

There's a new microcode and BIOS for MSI/Asus that according to wccftech affects stability in UE4/5 games and shader compilation. Anyone tried that? https://wccftech.com/intel-700-600-...-disable-cep-gain-cpu-performance-run-cooler/

 

[bit_user]

There's a new microcode and BIOS for MSI/Asus that according to wccftech affects stability in UE4/5 games and shader compilation. Anyone tried that? https://wccftech.com/intel-700-600-...-disable-cep-gain-cpu-performance-run-cooler/

Also covered here:

https://www.tomshardware.com/pc-com...t-losing-performance-recommends-disabling-cep
​

 

[the_vede_]

I'm not sure I really understand either article about the new microcode & BIOS.

The charts show how disabling CEP seems to drop temps and power draw while maintaining or slightly increasing perf under certain heavy workload conditions, like when running Cinebench. That's pretty cool, but does it have the same impact on ALL, MOST, or just SOME heavy workloads? There's nothing in either article that really explains why disabling CEP works or what the potential trade-offs might be.

Here's what I'm left wondering:

• Are there any downsides to disabling CEP? If so, what are they?
• Disabling CEP should be used in concert with what other BIOS power settings? Should we leave BIOS defaults (i.e., unlimited max watts & amps + default AC Load Line) and then disable CEP? Or should we first enforce Intel's power specs and then also disable CEP? Or...?
• If disabling CEP does indeed resolve these types of crashes, how and why is it doing that? And is it THE answer to this riddle or is it just one potential solve?

 

[bit_user]

I'm not sure I really understand either article about the new microcode & BIOS.

The charts show how disabling CEP seems to drop temps and power draw while maintaining or slightly increasing perf under certain heavy workload conditions, like when running Cinebench. That's pretty cool, but does it have the same impact on ALL, MOST, or just SOME heavy workloads? There's nothing in either article that really explains why disabling CEP works or what the potential trade-offs might be.

It might be better to discuss CEP in the comments of that article.

 

[JarredWaltonGPU]

I'm not sure I really understand either article about the new microcode & BIOS.

The charts show how disabling CEP seems to drop temps and power draw while maintaining or slightly increasing perf under certain heavy workload conditions, like when running Cinebench. That's pretty cool, but does it have the same impact on ALL, MOST, or just SOME heavy workloads? There's nothing in either article that really explains why disabling CEP works or what the potential trade-offs might be.

Here's what I'm left wondering:

• Are there any downsides to disabling CEP? If so, what are they?
• Disabling CEP should be used in concert with what other BIOS power settings? Should we leave BIOS defaults (i.e., unlimited max watts & amps + default AC Load Line) and then disable CEP? Or should we first enforce Intel's power specs and then also disable CEP? Or...?
• If disabling CEP does indeed resolve these types of crashes, how and why is it doing that? And is it THE answer to this riddle or is it just one potential solve?

I can't speak to all CPUs or motherboards, but on the MSI MEG Z790 Ace, with a 13900K that's known to have instability problems at "stock" mobo settings, disabling CEP (or setting it to Auto) seems to increase my instability issues with certain games. It's a bit hard to nail down for certain, but I've left it explicitly enabled on my latest testing.

I will also say that, contrary to what I have said elsewhere, my system isn't "100% passes everything and never crashes" stable. Specifically, Far Cry 6 and Watch Dogs Legion both occasionally crash during the initial shader compilation stage — they didn't do that before, and it may be from other stuff that's changed. (I did a full Windows 11 22H2 reinstall, then locked out Windows Update to keep on that release.) Curiously, the games that I had the biggest problems with before (Hogwarts Legacy, The Last of Us, and Metro Exodus) all seem perfectly stable.

 

[ToriWayne]

I have i7-12650H laptop and have these crashes. It's windows 11. Just got it for Christmas and Horizon is the first game I put on it, it stutters a lot, then in major battles it freezes suddenly and the computer forces a long reboot... Got a new comp cuz old one wouldn't run games and now new updated one won't either???

 

[JarredWaltonGPU]

I have i7-12650H laptop and have these crashes. It's windows 11. Just got it for Christmas and Horizon is the first game I put on it, it stutters a lot, then in major battles it freezes suddenly and the computer forces a long reboot... Got a new comp cuz old one wouldn't run games and now new updated one won't either???

It doesn't sound like you're dealing with this particular issue on the laptop. Likely causes for the laptop crashing are that it's either overheating, or you haven't updated all of the drivers, or perhaps it's just a bad laptop. The crashes on the Core i9 desktop chips often give a "GPU out of memory error" and crash to desktop, not a full system crash. Also, they happen during the initial shader compilation, meaning you can't even play the game — the crash happens before the game starts.

For your laptop, I'd update your GPU drivers (if you haven't already), and use Display Driver Uninstaller before installing the new drivers. So run DDU, kill the AMD, Intel, and/or Nvidia GPU drivers (depending on what laptop you have), then reinstall the Intel/AMD integrated graphics drivers, and finally reinstall the AMD/Nvidia dedicated GPU drivers.

If there are still issues with crashing, I'd run HWiNFO64 and log temperatures while playing the game. Have it log every second, and when you get a crash, check the log file. You should see temperatures on the CPU, GPU, and other devices. If the CPU or GPU are breaking 90C, that's a concern. There may be BIOS / configuration options on the laptop to change the fan mode, and you would probably want something with higher RPM fans (and more noise).

 

[Rabbitjxy]

Hello,

how many CPUs are having this issue? i have problems with Horizon Forbidden West crashing constantly with my i7-13700k and 4070TI with a Asus asus rog b760 f gaming. im already in contact with Intel and could get a RMA. But im kinda worried that the new CPU i would get would come with the same problem. is it unlikely that id get the same problem again?

 

[JarredWaltonGPU]

Hello,

how many CPUs are having this issue? i have problems with Horizon Forbidden West crashing constantly with my i7-13700k and 4070TI. im already in contact with Intel and could get a RMA. But im kinda worried that the new CPU i would get would come with the same problem. is it unlikely that id get the same problem again?

Does it crash during the initial loading sequence? Or does it get into the game itself and then crash? Basically, shader compilation happens when you say "load save / continue" and if it's the problem being discussed here, that's when the crash would occur. I haven't had issues with HFW, but then I've already adjusted my power/amperage settings to mitigate the problem.

 

[Rabbitjxy]

Does it crash during the initial loading sequence? Or does it get into the game itself and then crash? Basically, shader compilation happens when you say "load save / continue" and if it's the problem being discussed here, that's when the crash would occur. I haven't had issues with HFW, but then I've already adjusted my power/amperage settings to mitigate the problem.

Sometimes i cant event load in but i do get the "run out of VRAM message most of the time when it crashes. but yeah also when shader compilations. The german Intel support told they never had this problem in their team when i showed them this article, and also blamed it on my GPU/ Driver. I even changed my MSI card with a Asus card. Sometimes im able to play about half an hour before the next crash. Than it randomly closes itself and when restarting and trying to load the safe games it gives me the error message. After restarting the PC im able to go for half an hour again.