Intel CPUs causing "GPU crashes"

+

DC9V

Forum veteran
Hello!

Here's a summery of the current drama about the instability of Intel CPUs, posted by Hardware Unboxed (see linked video).

"Owners of some 13th and 14th gen Intel CPUs have been experiencing crashing when performing certain tasks, such as loading games or running game servers. And this would sometimes appear as an out-of-video-memory issue, despite being a CPU problem, not a GPU problem."

Apparently, Intel's recent microcode updates did not solve the issue. Hardware Unboxed says it's expecting another update from Intel within the next couple of weeks.

tags:
13400 13900 K 14900 K 13900K 14900K BIOS update ASUS GIGABYTE micro code stability crash doesn't launch can't start UEFI
 
Hello!

Here's a summery of the current drama about the instability of Intel CPUs, posted by Hardware Unboxed (see linked video).

"Owners of some 13th and 14th gen Intel CPUs have been experiencing crashing when performing certain tasks, such as loading games or running game servers. And this would sometimes appear as an out-of-video-memory issue, despite being a CPU problem, not a GPU problem."

Apparently, Intel's recent microcode updates did not solve the issue. Hardware Unboxed says it's expecting another update from Intel within the next couple of weeks.

tags:
13400 13900 K 14900 K 13900K 14900K BIOS update ASUS GIGABYTE micro code stability crash doesn't launch can't start UEFI
What is your assumption that is going on with these processors?
 
Yeah, it is good that it's uncovered, as to offload the grief to the right place.
It's not the game, it's Intel that effed up.
 
Indeed...I appear to be one of those running afoul of the Intel problem. My CP77 game was getting harder and harder to start...it would crash on the first splashscreen almost instantly after it came up. I tried everything with my game, the mods, graphic settings...nothing worked. I decided it was a Windows problem, and in preparation for doing a full reset of Windows, I updated the BIOS on my MSI MOBO...and that seems to have cured my problem. That BIOS update [issued June 26th, IIRC] included a microcode upgrade...I don't know if it's the one or similar to the one Intel is saying they'll release in mid-August. I just hope my CPU [i9-13900KF] doesn't deteriorate any further. It seems that the CPU problem combines a serious over-volting of the processor, plus a problem [related?] with oxidation inside the chip itself...some kind of fabrication flaw. Intel is dragging its feet about doing anything to remediate the situation...refunds and RMAs are not forthcoming.
 
Afaik, all media say that the microcode update hasn't been released yet, @Tokenn77 . It is expected sometime mid-August.

Before all of this, I was keeping an eye on AMD as my potential next platform. After this, I will think thrice before getting another Intel, and am not even among the damaged folks (I'm with 12-th gen). The biggest pain with AMD is that it's so fussy about RAM. Like... dafuq.
 
Afaik, all media say that the microcode update hasn't been released yet, @Tokenn77 . It is expected sometime mid-August.

Before all of this, I was keeping an eye on AMD as my potential next platform. After this, I will think thrice before getting another Intel, and am not even among the damaged folks (I'm with 12-th gen). The biggest pain with AMD is that it's so fussy about RAM. Like... dafuq.
They have released an interim fix. It dials down the voltage, improving stability and hopefully preventing further damage to the chip, at the cost of some performance. It's something everyone with a 13th or 14th gen Intel chip should install while they wait for the mid-August update.

For those unsure where to get it, go to your Motherboard, or PC manufacturers (if a prebuilt PC), website. Look for the most recent Bios update for your model.
 
What is your assumption that is going on with these processors?

There is no need for assumptions here. The issue is well known and well documented at this point.

In the simplest terms, the issue causes more electricity than supposed to be sent into various sectors of the CPU and essentially burn them out. Which leads to a slew of problems that can appear as if caused by something else than your CPU.

That damage is irreversible.

Afaik, all media say that the microcode update hasn't been released yet, @Tokenn77 . It is expected sometime mid-August.

Before all of this, I was keeping an eye on AMD as my potential next platform. After this, I will think thrice before getting another Intel, and am not even among the damaged folks (I'm with 12-th gen). The biggest pain with AMD is that it's so fussy about RAM. Like... dafuq.

Same. This is absolutely shameful from Intel. They kept it under wraps for two generations. They kept it under wraps until users blew the lid open on this.

I've been building my PCs myself for over 17 years now and it's always been Intel/Nvidia. I'm still on 12th gen too but I'm upgrading as soon as the 5000s series is out but it'll likely be an AMD/Nvidia combo this time around.

Time to give team red a chance.
 
Last edited:
是的,我的CPU是14900KF,我也经常遇到和你一样的问题。虽然我可以通过打开多个程序登录,但游戏在玩了 10-20 分钟后会因 GPU 故障而崩溃。我希望英特尔能在八月中旬之前解决这个问题,因为这让我非常沮丧。
Post automatically merged:

Yes, my CPU is a 14900KF, and I also frequently encounter the same problem as you. Although I can log in by opening multiple programs, the game will crash due to GPU failure after playing for 10-20 minutes. I hope that Intel can fix this problem by mid-August, as it has made me very frustrated.
 
I don't think Intel is going to be able to do anything other than offer a replacement to users that request one, in the end. I do think they might share the load with mobo manufacturers, as microcode is, technically, issued as part of the motherboard BIOS. It comes directly from Microsoft, but mobo manufacturers need to be exhibiting some level of oversight, especially where voltage levels direct to the CPU and RAM are concerned. I actually find it a bit shocking that absolutely no one caught on to the situation before users started exhibiting problems. Goes to show how accurate and rigorous the testing was. (Unless MS directly told motherboard companies, "Oh, don't worry! Those voltage levels spiking up like that are completely normal!" That would make Intel 100% accountable for everything that followed. Hence, I doubt that's what happened. I think it's more like: no one tested them and just assumed everything was good. Sell-sell-sell.)

Whatever the actual situation, it looks like Intel is going to play the waiting game for as long as possible. Absolutely irresponsible and shameful, given how blatant the problem is. Not happy times for them.


_______________


One thing that I noticed on my present rig when I first got it (bought this one off the shelf) is that my performance was all over the place by default. Things that ran smoothly on my old 980 ti system were hitching and stuttering very noticeably on the much more powerful hardware. And it was systemic. It's an AMD system, Ryzen 5600X, and I was finally able to track the issue down to the way the native CPU "boosting" programs were interacting in the BIOS.

There were two: one from AMD directly, and then one that was included by ASRock as part of the motherboard features. Both were on by default, and both would compete with each other, constantly swapping between which technique was being used in a given situation. You should have seen the logs for the voltage. It was like a bloody 8.0 earthquake. I eventually decided, "Enough of this," and completely disabled BOTH of them. System has run buttery smooth every since.

That's a potential workaround for people that have the 13/14 gen chips. The voltage overcharge is likely going to be magnified by these "boost" programs, which attempt to dynamically overclock your CPU in real-time based on demands received from operation. The fluctuations can be pretty needlessly aggressive even when "running correctly". My personal opinion is that these programs are way more trouble than they're worth, and the performance gains they provide are negligible to middling at best. Not a fan of cooking my hardware and having my processor reaching temps of nearly 95*C in order to open a web browser 0.3 seconds faster. In games, I noticed absolutely no practical performance gains whatsoever between the two techniques. Zero. Nothing. Not one frame. (Obviously, everyone's individual mileage will vary, but I doubt anyone will really notice siginificant performance differences with them on or off.)

By disabling anything in your BIOS / UEFI that refers to boost, overdrive, turboboost, boost clock override, etc. (there are a lot of different names for them, depending on manufacturer) you're not underclocking your system. You're essentially telling the PC to run at default clock speeds...which used to...once upon a time...be how computers operated out of the box.

Do note that this is not a "fix" for the situation, but the stable voltage and lack of dynamic management will keep voltage levels more under control and prevent spikes. That will hopefully help to stabilize the PC and increase the longevity of 13th and 14th gen chips.
 

DC9V

Forum veteran
"Oh, don't worry! Those voltage levels spiking up like that are completely normal!"
It's similar to putting your finger in a candle light. All chemical processes take a certain amount of time, and so do chemical reactions like oxidation.

I think voltage spikes usually only last for a couple of micro seconds, whilst VID requests seem to last for several milliseconds. However, I'm not so sure if that alone is enough to cause rapid oxidisation of the CPU.

I was finally able to track the issue down to the way the native CPU "boosting" programs were interacting in the BIOS.
Yeah, those "Game Boosts" are ridiculous. I once accidentally had it enabled via MSI's Dragon Center. It was the only time that I saw CPU temps well above 100°C.
 
It's similar to putting your finger in a candle light. All chemical processes take a certain amount of time, and so do chemical reactions like oxidation.

I think voltage spikes usually only last for a couple of micro seconds, whilst VID requests seem to last for several milliseconds.
Nah. There are videos that show them sitting at well over 1.5v for many seconds on end. In general, it's nothing that one could possibly miss if one is stress testing the machine. Which, of course, would always be done before mass-marketing a brand new generation of CPU. Of course, the motherboard manufacturers would do the same for the various processor and RAM ranges that each of their boards support. And they would, of course, be rigorously tested by professional technicians. I mean, of course.

To be honest, I'm very inclined to say that the microcode has only been designed to "tackle a symptom, rather than the disease", but we'll see how things'll work out.
However, I'm not so sure if that alone is enough to cause rapid oxidisation of the CPU.
I totally agree about there being more to the story, and most likely some sort of manufacturing defect as well as the microcode issue. Intel was definitely feeling the crunch after the success of the Ryzen line. I don't doubt that they "stepped up" the development and manufacturing process to get products out there that would regain the lead and stay ahead.

Yeah, those "Game Boosts" are ridiculous. I once accidentally had it enabled via MSI's Dragon Center. It was the only time that I saw CPU temps well above 100°C.
I reached the, "Okay. I'm done," moment when I saw that the AMD boost had my CPU idling at ~75*C and the hotspot shot to over 95*C opening Firefox and going to Google.com.

Absolutely out of this world.

Brilliant.

A-a-a-mazing.

They be off. And off they be stayin'.
 
Top Bottom