Intel processors crashing Unreal engine games (and others)

[igor_kavinski]

I am at the point where I'm about 95% certain I'm going to return my motherboard and i9-14900KF for a Ryzen 9 7950X3D.

Aren't you reading the news? The 9950X is coming on 31st July.

 

[Futuremotion]

Aren't you reading the news? The 9950X is coming on 31st July.

Admittedly, I totally missed that. Like I said, I've always gone Intel and never really looked at AMD's offering in detail.

After now coming across this thread and reading all the various articles about 13-14th gen Intel instability I am just now starting to research AMD's product roadmap.

Are there adequate motherboards available that will fully support the 9950X on day one?

 

[poke01]

Are there adequate motherboards available that will fully support the 9950X on day one?

Yep, plenty. Like this one: https://www.asrock.com/mb/AMD/X670E Taichi/index.asp

Just type in Google x670/x670 AM5 motherboards.

 

[Futuremotion]

Yep, plenty. Like this one: https://www.asrock.com/mb/AMD/X670E Taichi/index.asp

Just type in Google x670/x670 AM5 motherboards.

Alright. I'm officially sold... time to switch teams baby. Farewell Intel!

Going with the Ryzen 9 9950X when It's released and that Taichi board looks wicked. Fortunately I already have every component I need other than the motherboard and CPU. So glad I stumbled across this thread, it definitely helped me avoid a huge headache with the i9.

 

[Markfw]

Alright. I'm officially sold... time to switch teams baby. Farewell Intel!

Going with the Ryzen 9 9950X when It's released and that Taichi board looks wicked. Fortunately I already have every component I need other than the motherboard and CPU. So glad I stumbled across this thread, it definitely helped me avoid a huge headache with the i9.

Notice the small heatsink with a fan. Thats actually for the NVME !!!!! (on the asrockx670e)

I have 2 of those, and they are incredible.

But back to the topic.

 

[poke01]

https://twitter.com/x/status/1811808666565091574

More game devs are reporting the crashing

 

[RnR_au]

Intel can avoid repeating 1.13G Coppermine just fine

Some history for any of the young ones here. Intel Pentium III 'Coppermine';

A 1.13 GHz version (S-Spec SL4HH) was released in mid-2000 but famously recalled after a collaboration between HardOCP and Tom's Hardware[8] discovered various instabilities with the operation of the new CPU speed grade. The Coppermine core was unable to reliably reach the 1.13 GHz speed without various tweaks to the processor's microcode, effective cooling, higher voltage (1.75 V vs. 1.65 V), and specifically validated platforms.[8] Intel only officially supported the processor on its own VC820 i820-based motherboard, but even this motherboard displayed instability in the independent tests of the hardware review sites. In benchmarks that were stable, performance was shown to be sub-par, with the 1.13 GHz CPU equalling a 1.0 GHz model. Tom's Hardware attributed this performance deficit to relaxed tuning of the CPU and motherboard to improve stability.[9] Intel needed at least six months to resolve the problems using a new cD0 stepping and re-released 1.1 GHz and 1.13 GHz versions in 2001.

Source

I don't think Intel has the cashflow for a recall.

 

[adroc_thurston]

I wonder if RPL-HX crashes too.
That might just be even funnier.

 

[DrMrLordX]

If it is a degradation issue from trying to run too high of frequencies, it could be that those CPUs were already too badly damaged that even down clocking will no longer work.

In the case of the game server Raptor Lake CPUs running in the W680 boards, I don't think they ever ran at "too high" frequencies in those setups. Certainly not compared to similar CPUs in desktop boards.

 

[DAPUNISHER]

Retail is already seeing the effects as the news spreads and confidence drops. - https://hardwaretimes.com/pc-gamers-amd-ryzen-intel-13900k-14900k-crash-fail/

 

[CakeMonster]

Kind of a lucky timing for AMD given the imminent Z5 release...

 

[ErichZannMusician]

Anecdotes on reddit point to usage of a contact frame being associated with vastly lower incidence of instability.

Might be worth exploring.

Rumors and leaks already point to different mechanical specifications for LGA1851 that eliminates the need for contact frames.

Would be very surprised if this issue persists in Arrow Lake.

 

[RnR_au]

So hypothetically the cpu is being literally crushed to death?

 

[coercitiv]

Anecdotes on reddit point to usage of a contact frame being associated with vastly lower incidence of instability.

ADL is on the same socket, so mechanical cause is unlikely.

However, contact frame lowers temps and hot spots, so thermal cause is possible.

People using contact frames are also arguably more inclined to optimize cooling further, so the incidence of contact frames on more stable systems might also be explained by user behavior patterns that limit or delay degradation.

 

[igor_kavinski]

So hypothetically the cpu is being literally crushed to death?

It's more the warping of the IHS creating an airgap. I had used a graphite pad before on my 12700K in winter (had a contact frame but was too squeamish to install it). Come summer (after more than a year of being unused), I was getting thermal throttling in XTU. Took off the graphite pad, applied paste and this time did two things different:

1) Put the casing on its side so the heavy heatsink (https://www.anandtech.com/show/17608/the-iceberg-thermal-icesleet-g6-stealth-review) is pressing hard on the CPU with its weight.

2) Screwed it down a bit too hard (was afraid that I had crushed it and it wouldn't boot up).

The result?

Page 545 - Discussion - Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)

Page 545 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

Spoiler: Crush the damn thing!

Faster on air than a 14600K in ST and MT, possibly because the contact between the IHS and heatsink base has no gap.

 

[DAPUNISHER]

While we are doing Reddit anecdotes -

Commenter 1

Been building PCs for a while and this is the first time I have had so many bad i9 CPUs this year and last. On one build we went through 3.

Commenter 2

I work in a computer shop as a tech. The amount of 13th & 14th gen we've sent back to RMA is insane. It's not just the i9s. I really wish I had access to the data and could publish it... It's i5s, i7s and i9. Haven't seen i3s yet.

Commenter 3

I'm reading this as I'm helping a friend as we speak diagnose his 13600KF that's throwing thousands of errors in OCCT. It's not just the i9s.

Commenter 4

This is purely anecdotal, but for what it's worth I was seeing a ton of instability from day 1 with the stock motherboard settings on my Asus z790-p with a 14900k. Obviously there was the usual "idiot OEM running the cpu faaaaar beyond spec for voltage and current", so dialling it back to Intel spec helped, but didn't fully resolve the issue. Lowering the PL1 from 253 to 225W helped even more, but I was still seeing periodic soft errors (WHEA) in the event log. Dropping to 200W, however, has completely 100% fixed the issue on my system, but this comes at the cost of losing several hundred mhz on the boost clocks, which naturally isn't ideal.

While not a fix, if someone is struggling a lot with their system and they're not in a position to RMA the chip or wanting to roll the dice at this time while investigations are still ongoing and there's a good chance of simply getting yet another faulty chip, it may be worth dialling down the power limits to even more conservative levels than the official spec as it did resolve 100% of instability that I was seeing, even under sustained overnight stress testing.

Commenter 5

I've seen the same problem exhibit on most mid range models of the 13th and 14th gen as well, but no where near as frequent as the 13900 and 14900. Seen it a couple of times on 12th Gen but telling users to update their BIOS and make sure they have the latest Intel Windows updates seems to resolve the issue 12th gen.

At first I thought it was a memory controller cause all the problems seemed to point back to something to do with storage or memory. Manually setting XMP values on RAM or slightly lower clocks and timings seemed to create better stability too. But appears these are only temporary fixes for many people.

I have also been seeing this problem since the release of 13th gen. Have been repeatedly reminding end users I've helped troubleshoot and tech support with to flood Intel with the complaints. Seems that pushback of customers to Intel lead them to telling everyone it's a developer issue to which many developers are now coming out and saying "no, there's definitely something off with the CPUs"

Commenter 6

My program at college just bought 50 systems for two labs. Asus prime b760 plus boards with i7-14700K CPUs. Almost all of the CPUs have failed (5 did not) as the boards came shipped with a bios that tried pushing 1.71V upon first boot. This is not a fun RMA :/

Commenter 7

I got 2 13900k cpus refunded by Intel because they were broken. AMD is a must at the high end right now sadly

Commenter 8

I had an issue with MW3 zombies crashing, never multiplayer. It would go a few days sometimes without crashing and sometimes crashing twice a day. My friend had a R9 5950X and never crashed while we played together, not once. I finally got some new BIOS updates from MSI, and the latest BIOS, I haven’t crashed a single time. Knock on wood. It’s been 3 months since updating the BIOS and all is well finally. I have an Intel 13700KF

Commenter 9

My 13900k just degraded after a year of daily usage. Had to replace it with AMD Ryzen 7950x3d, not a single issue since

Commenter 10

I had to set the 13900KF pcore ratio to 52x after a while or it would BSOD loop. When I play The First Descendant I also have to switch to PC to power saver mode so it doesn't crash loading shaders. Most DX12 games are too unstable and crash so I have to downgrade to the DX11 versions if there is any.

Commenter 11

I too have a 14900k and was experiencing blue screens when i first got it put together. Kept reading and reading about what to do and someone told me to go into bios and choose box cooler rather than water cooled which helped but underclocked my CPU i didn't pay top dollar to have all these issues and need to underclock my CPU. Wish i would have done more research and gotten a Ryzen CPU instead now. Before i switched to box cooler mode i was getting 90c temps idle on desktop. Craps ridiculous

Commenter 12

Built a system with a 13600k for a friend about a month ago, worked fine for a couple of weeks and then suddenly half his games were glitching out and lightroom would constantly crash. Took me 2 days to figure out it was a weird CPU fault, it worked fine with my 12700k, waiting on a return now.

Would it maybe be better to just switch to a 12600k or 12700k instead?

My favorite comment was that someone needs to do a wellness check on the Userbenchmark guy. 🤣

 

[igor_kavinski]

@BoomerD , everything fine with your 13600K?

 

[NTMBK]

These frequencies are not a result of a true CPU capability, but a result of desparate Intel wanting to beat/match AMD in review charts. They are really CRAZY. Do not use them.

The CPUs are fine. They break if you let them run at crazy frequencies.

Intel sells these processors at a specific advertised performance level, and people pay a premium for that performance. If the product cannot achieve the advertised specification without breaking down, then it is defective and Intel should issue refunds.

 

[BTRY B 529th FA BN]

Didn't realize this was still going on. I've heard so many people say they don't have an issue with their 13 and 14th gen chips, however I am seeing a lot of them for sale in secondary markets on other forums.

 

[igor_kavinski]

however I am seeing a lot of them for sale in secondary markets on other forums.

And the uninformed buyers of those will have no chance of knowing that the CPU is bad unless they push it hard in specific instances.

In-person meetup sales will go something like,

Seller: So you can see Geekbench, Cinebench, 3DMark etc. all working fine. You are getting a good price. However, you know there is no warranty for used stuff so I won't be responsible if you have any issues afterwards.

Buyer: Can you show me running this and this and this?

After a few demos of stuff that the buyer wants to see, he's satisfied and seals the deal. And he's happy as could be, until a few days or weeks later when he isn't...

Nothing to do when that happens other than cursing the seller and maybe Intel and maybe himself for not researching things more. And trying to pass that crap onto some other shmuck down the line.

 

[moinmoin]

Warframe devs post some info on their forums:

Instability on recent Intel Processors

While investigating crashes in Warframe we came across a particular series that were not crashing in our code (they were crashing in nvgpucomp64.dll, a component of Nvidia drivers). After aggregating hundreds of reports from helpful players we discovered a pattern: almost all were coming from sys...

forums.warframe.com

Following image needs to be quoted big:

Note: Those are crashes by all CPU manufactures and chip generations.

And addendum: That forum post is one example where updating to the latest BIOS seems to have helped resolve the issue (though I'd personally not hold my breath due to Intel's lacklustre messaging).

 

[DAPUNISHER]

That pie chart shuts down the claims I have read from some users that all you need to do is disable the iGPU and it fixes things. At least for Warframe. Only time will tell if the bios fix holds up.

I am of 2 minds about the claims the contact frame is an effective high percentage solution. My gut brain says it's astroturfing or guerrilla marketing. My smooth brain says there are reasons to think the claims have some legitimacy. And if so, the notification to end users would have to be on the ARC rebar beat you over the head with it level. It would have to be provided with every new CPU or mobo sold. They'd have to be willing to service systems where the user is not competent or capable of installing the CF. Ship them for free to current owners. Legally it might even open them up to liability there, since it would be an admission there is a design flaw in how the CPU is installed on the board. Without even getting into the liability for defective CPUS. And ultimately it won't be THE solution just part of it.

If it is astroturfing or marketing? On the bright side, at least you will be contributing to Roman's retirement fund.

 

[DAPUNISHER]

Matt from Alderon Games posted this on reddit a couple of hours ago -

Thanks for posting about this:

We had bulk purchased lots of 13900k and 14900ks to use for servers pre-paid for a year. This was a very expensive massive failure for us and spent months trying to get the right BIOS settings and patches to work.

We even have developers who are on laptop chips having defects and crashes, One MSI Laptop dev who got a really expensive 13th gen HX now has a paperweight and cant' RMA it.

This doesn't include the time we wasted pulling the game apart, thinking it was our fault and realising, oh no, this is a hardware issue at massive scale.

It's getting to the point where we may have to ban 13th and 14th gen chips being used by any developer with Unreal Engine at our company because of the productivity loss and file corruption, which can get checked into version control and cause problems for other devs.

Uh oh, the HX are not immune either.

His feed is juicy - https://www.reddit.com/user/Matt_AlderonGames/

 

[IEC]

1) Contact frames are unlikely to significantly mitigate the problem. 12th gen is not affected despite same ILM
2) Better cooling? Wendell reports hot spot temps in high 60s and low 70s at maximum, meaning exceptional cooling in a DC environment. Despite this, they had repeated failures
3) Until Intel makes a blanket "we'll make it right" statement and offers more information, the internet rumor mill is going to go into overdrive because recent news finally validates that it is not PEBKAC and it is not a "developer issue" - it is likely a hardware issue given how widespread it is and how over-represented Raptor Lake is in the crash data

 

[DAPUNISHER]

Not like the contact frame is going to work on those laptops either.