Intel processors crashing Unreal engine games (and others)

[DAPUNISHER]

As long as Intel keeps honoring RMAs and issuing refunds when requested, it's a manageable PR mess. Despite the infamous Xbox 360 RROD debacle, the console was wildly popular and led the PS3 until near the end of the generation. MS took over a billion dollar hit on that in unadjusted dollars to service affected owners. That's how you retain customers; makes things right as fast and easy as possible. This won't get anywhere close to that level of losses.

 

[igor_kavinski]

https://www.reddit.com/r/intel/comments/199dyud/i914900k_stock_vs_undervolted_peak_power

I like this quote:

It's a monster chip but getting them stable or just running at sensible settings can be an absolute fustercluck with some of the default motherboard settings. To me it seems like intel is at the very edge of "out of the box" stability with some batches+motherboard combinations.

Intel has a lovely Blender Blunder thread on their forums too!

Blender 4.0.2 CPU rendering is causing multiple 14900k users to BSOD

Hi, about a week ago someone posted a thread about blender crashing and blue screening on their 14900k, (https://community.intel.com/t5/Processors/Critical-Overheating-Issues-with-i9-14900K-PC-Crashes/m-p/1563723/thread-id/68588) I'm posting because this is not an isolated issue, over the past...

community.intel.com

Spoiler: Quotes quotes quotes!

As for wert5465, from the symptoms you mention, it could be an overheating issue. I highly recommend creating a new thread so we can properly address the issue that you are experiencing.

Dear moderator, why should I create a separate topic, judging by most of the messages here, everyone is overheating!

1. In BIOS, select "OC," select "CPU Core Voltage Mode," select "Offset Mode," select "+(by PWM)", .Then, adjust the voltage until the system is stable. We recommend not exceeding 0.025V for a single increase.
2. In BIOS, select "OC," select "DigitALL Power," change "CPU Loadline Calibration Control," . We recommend starting from "Mode 7" to a lower value until the system is stable.

Note: After trying out 1 or 2, run the XTU test again and see if the AVX2 test can pass. You may download Intel XTU at this link: https://www.intel.com/content/www/us/en/download/17881/intel-extreme-tuning-utility-intel-XTU.html. After that, try running Blender 4.0.2 to see if the crash will occur again.

I9 14900k Seems to be my last intel procesor i would EVER buy. such a Scam of a proccesor. not only blender is affected, there are tons of issues in gaming too. Dx12 crashes, fornite is unable to run and so many. intel needs to pull back 14900k.

This is happening to me and it's extremely frustrating. It is affecting my profession Have invested 5000$ and facing this hardware issue. I tested my PC and it is definitely blender (any version) and its incompatibility with the 14900k that's causing the issue. I really hope it gets looked into asap. Please intel look after it. Thanks

Hello guys. I got the solution. I am facing the same issue with my i9 14th gen and I fix this.
Open Device Manager -> disable Intel UHD Graphics 770 and run System on my Nvidia graphics card. and it working very well. now my blender not crashing while rendering.

Could this be THE solution???

 

[Mopetar]

Sounds like a temporary bandaid if the GPU is what is pushing it over the edge.

 

[coercitiv]

I read through the thread @igor_kavinski posted and... damn: they really are leaving their best customers out in the cold.

The golden post from Intel support is this one though:

We do have some troubleshooting steps here that you can try. Please try the following steps, as I can see you are using an MSI motherboard:

1. In BIOS, select "OC," select "CPU Core Voltage Mode," select "Offset Mode," select "+(by PWM)", .Then, adjust the voltage until the system is stable. We recommend not exceeding 0.025V for a single increase.
2. In BIOS, select "OC," select "DigitALL Power," change "CPU Loadline Calibration Control," . We recommend starting from "Mode 7" to a lower value until the system is stable.

Note: After trying out 1 or 2, run the XTU test again and see if the AVX2 test can pass. You may download Intel XTU at this link: https://www.intel.com/content/www/us/en/download/17881/intel-extreme-tuning-utility-intel-XTU.html. After that, try running Blender 4.0.2 to see if the crash will occur again.

They are advising their users to blindly bump up Vcore until system seems stable, and I say blindly because they might not even know point (1) and (2) will influence each other. I guess that's the premium i9 experience.

 

[poke01]

Is this only effecting the 13th and 14th gen K SKUs? The non-K variants are fine?

 

[coercitiv]

Is this only effecting the 13th and 14th gen K SKUs? The non-K variants are fine?

For 14th gen I think we've had isolated reports that 14900 non-K is also having issues, at this point if I had to draw a line, I think everything above 13600K/14600K might be affected. Think of it as a gradient though, the lower you go in the SKU line... the fewer CPUs with problems you'll get, probably an exponential decrease too.

 

[igor_kavinski]

My friend with Core i5-12400 was saying that his mouse pointer was freezing frequently during normal usage. Not sure if it's related to the contact frame or this degradation issue. He does have an ASUS mobo and he did not upgrade the BIOS from what came with the mobo so it could point to voltage related degradation. Unfortunately, he's not taking it seriously and all my attempts to get him to pinpoint the cause have been in vain so far.

 

[coercitiv]

My friend with Core i5-12400 was saying that his mouse pointer was freezing frequently during normal usage. Not sure if it's related to the contact frame or this degradation issue. He does have an ASUS mobo and he did not upgrade the BIOS from what came with the mobo so it could point to voltage related degradation. Unfortunately, he's not taking it seriously and all my attempts to get him to pinpoint the cause have been in vain so far.

Instead of a contact frame, did he try uninstalling Armory Crate instead?

 

[igor_kavinski]

Instead of a contact frame, did he try uninstalling Armory Crate instead?

Will ask him to do that too.

 

[DrMrLordX]

They are advising their users to blindly bump up Vcore until system seems stable, and I say blindly because they might not even know point (1) and (2) will influence each other. I guess that's the premium i9 experience.

Isn't that going to lead to faster degradation?

 

[DAPUNISHER]

Isn't that going to lead to faster degradation?

That seems like a reasonable deduction.

As a user, I would not do any of that. I'd respond that I did not pay good money to have to troubleshoot and test a product that I expected to work as advertised. TS and testing I may not have done before, and may even find complicated and consequently time consuming. And as no definitive solution has as of yet been offered I'd like a refund, since I may otherwise end up facing the same issues at any time in the future.

 

[coercitiv]

Isn't that going to lead to faster degradation?

It depends, but making things worse is indeed on the list of possible outcomes.

The weird thing about this conundrum Intel is caught within right now is the CPUs are both overvolted and undervolted at the same time. I know it sounds crazy, but when you break down what some of the boosting & power management settings do, you realize that before the UEFI updates a CPU like the 14900K was likely overvolted during light/medium loads, and undervolted during heavy loads.

Taking things one step further, in this particular case undervolting can also ironically be an indirect cause of degradation. With current limits disabled, lowering the voltage curve will allow the CPU to use more current before it reaches the power/thermal limit. From this PoV adding a positive voltage offset may actually protect the CPU from harm by reaching the power limit at lower clocks, with lower current consumption stress. Wrong solution, somewhat happier outcome.

All of this is speculation though, we might even be witnessing the first CPUs that are being slowly fried during light loads and also slowly boiled during heavy loads.

 

[darkswordsman17]

Please someone tell me that Userbenchmark or whatever that one trash site that kept "adjusting" their scoring tried to trash AMD over that situation, only for it to degrade in their face?

As long as Intel keeps honoring RMAs and issuing refunds when requested, it's a manageable PR mess. Despite the infamous Xbox 360 RROD debacle, the console was wildly popular and led the PS3 until near the end of the generation. MS took over a billion dollar hit on that in unadjusted dollars to service affected owners. That's how you retain customers; makes things right as fast and easy as possible. This won't get anywhere close to that level of losses.

What a weird post, I'm not sure what you're even responding to? Certainly Intel's weathered bigger issues than this. And more recently than the Xbox 360 even (no clue why you even brought that up). Intel wrote off a billion on just their Atom line alone (granted that was because they couldn't even pay people to use it in phones and tablets). And they took a bigger loss than that on the Sandy Bridge mobo issue.

I actually wouldn't be surprised if this costs Intel more than a billion, especially if it means they have to adjust their future products because something tells me Intel pushing their chips to the absolute limits just to be competitive is not going to stop all of a sudden. It seems to be getting worse, if anything.

 

[DAPUNISHER]

What a weird post,

It certainly is when it is not understood. Did you read the whole thread?

I'm not sure what you're even responding to?

How Intel is handling this mess. I thought it obvious, but here we are.

Certainly Intel's weathered bigger issues than this. And more recently than the Xbox 360 even (no clue why you even brought that up).

Again I thought it was obvious. I explicitly used it as a good example of how to handle defective products with your customers. I have also wrote about Intel's impressive legal teams, when it comes to court cases, multiple times in this thread. However, I did not address the contra revenue or other losses in the past. They are no longer in the position they were in back then either, so it would hurt more now than then. But again, if you have read the whole thread you know I don't think this is going to be all that bad for them financially. Any expense like this is unwelcome of course.

I actually wouldn't be surprised if this costs Intel more than a billion,

I speculated it won't get anywhere near what it cost MS, not even in unadjusted dollars. Which an inflation calculator says a billion is over 1.5 billion now. My reasoning is that if they had to replace a million CPUs, which is probably far too high an estimate, it would not cost nearly a billion. And my contention, as I have written pages back, is that they will never payout anywhere near whatever judgement might be made against them. They are still stringing out a penalty from 20yrs ago, links are again, pages back.

specially if it means they have to adjust their future products because something tells me Intel pushing their chips to the absolute limits just to be competitive is not going to stop all of a sudden. It seems to be getting worse, if anything.

I posited along similar lines pages back. If they need to dial it up to 11 again to compete, they will do it without hesitation. Only a large judgement where they actually had to payout might give them pause. And as I wrote pages back, that is highly unlikely to happen.

 

[DrMrLordX]

All of this is speculation though, we might even be witnessing the first CPUs that are being slowly fried during light loads and also slowly boiled during heavy loads.

That's odd. At low current levels, high voltage isn't supposed to make that big of a difference. Or at least it hasn't made any difference on AMD CPUs for years.

 

[Panino Manino]

Wendell from L1 did an investigation.
This is costing people a lot of money.

 

[Hitman928]

Wendell from L1 did an investigation.
This is costing people a lot of money.

The processors crashing even on W series motherboards should put the final nail in the coffin that this is a motherboard manufacturer caused issue.

 

[Markfw]

The processors crashing even on W series motherboards should put the final nail in the coffin that this is a motherboard manufacturer caused issue.

But Intel has no sayso or approval of bios settings ?

I mean its been proven that if you "declock" the CPUs correctly, they are stable.

 

[gdansk]

Wendell from L1 did an investigation.
This is costing people a lot of money.

He does ask to take this data with a grain of salt but this video makes it sound much worse than I thought.

 

[Ranulf]

Wow, the data center Wendell talked to was reporting problems with the i9 13th and 14th gen and thus charging more for support contracts.

 

[coercitiv]

But Intel has no sayso or approval of bios settings ?

I mean its been proven that if you "declock" the CPUs correctly, they are stable.

I have yet to watch the video, but I think @Hitman928 meant that W boards crashing as well pretty much exonerates Mobo makers or at least points to a bigger problem than BIOS settings alone.

 

[RnR_au]

Wendell from L1 did an investigation.
This is costing people a lot of money.

Offtopic, but one of the comments on the video said this;

I can't recall an issue like this since Sun accidentally shipped radioactive chips that were unstable.

Wait... what!?

After some googling;

When Sun folks get together and bullshit about their theories of why Sun died, the one that comes up most often is another one of these supplier disasters. Towards the end of the DotCom bubble, we introduced the UltraSPARC-II. Total killer product for large datacenters. We sold lots. But then reports started coming in of odd failures. Systems would crash strangely. We'd get crashes in applications. All applications. Crashes in the kernel. Not very often, but often enough to be problems for customers. Sun customers were used to uptimes of years. The US-II was giving uptimes of weeks. We couldn't even figure out if it was a hardware problem or a software problem - Solaris had to be updated for the new machine, so it could have been a kernel problem. But nothing was reproducible. We'd get core dumps and spend hours pouring over them. Some were just crazy, showing values in registers that were simply impossible given the preceeding instructions. We tried everything. Replacing processor boards. Replacing backplanes. It was deeply random. It's very randomness suggested that maybe it was a physics problem: maybe it was alpha particles or cosmic rays. Maybe it was machines close to nuclear power plants. One site experiencing problems was near Fermilab. We actually mapped out failures geographically to see if they correlated to such particle sources. Nope. In desperation, a bright hardware engineer decided to measure the radioactivity of the systems themselves. Bingo! Particles! But from where? Much detailed scanning and it turned out that the packaging of the cache ram chips we were using was noticeably radioactive. We switched suppliers and the problem totally went away. After two years of tearing out hair out, we had a solution.

From https://news.ycombinator.com/item?id=10639755

Those poor engineers...

 

[coercitiv]

There's a lot to unpack from the L1 video, but the info from large system integrators alone should be sobering news: they said around 10-25% of the CPUs have a problem or are marginal in some way (stability wise). Wendel thinks this is the lower floor for the real-world numbers because the telemetry data made him think the number is much higher and systems are more likely to develop problems as time passes by (obviously his telemetry data was for gaming population). 13900K/14900K game servers having major issues even with pro hardware and the ease they had to find unstable systems in the data center also seem to confirm this.

I think it's safe to say now that stability issues on 13900K/14900K are not isolated events, but rather likely events. Up until a month ago I would have had no problem buying one since I thought that a correct mobo setup would keep everything in check. However, in the light of this new info and the fact that Intel has yet to come up with a definitive statement, I would stay away from any 13th/14th gen CPU for now.

 

[RnR_au]

Are laptop 13th and 14th gen suspect as well? Haven't seen any discussion on this.

 

[SteinFG]

Are laptop 13th and 14th gen suspect as well? Haven't seen any discussion on this.

Only HX i7/i9 series uses Raptor lake silicon (and I'm not sure about i5 HX) . All H and U chips are alderlake copies. It will be hard to gauge if HX chips degrade cause I think there isn't a lot of them being sold in the first place.