Intel Investigating Games Crashing On 13th and 14th Gen Core i9 Processors (theverge.com) 35
An anonymous reader quotes a report from Ars Technica: Owners of Intel's latest 13th and 14th Gen Core i9 desktop processors have been noticing an increase in game crashes in recent months. It's happening in games like The Finals, Fortnite, and Tekken 8, and has even led Epic Games to issue a support notice to encourage Intel Core i9 13900K and 14900K owners to adjust BIOS settings. Now, Intel says it's investigating the reports. "Intel is aware of problems that occur when executing certain tasks on 13th and 14th generation core processors for desktop PCs, and is analyzing them with major affiliates," says an Intel spokesperson in a statement to ZDNet Korea.
The crashes vary in severity depending on the game, with some titles producing an "out of memory" error, others simply exiting out to the desktop, and some locking up a machine entirely. Most of the games affected seem to be based on the Unreal Engine, which could point to a stability issue that Intel needs to address. The only workarounds that seem to improve stability involve manually downclocking or undervolting Intel's processors. Epic Games has suggested changing the SVID behavior to Intel Fail Safe in the BIOS settings of Asus, Gigabyte, or MSI motherboards. Custom PC builders Power GPU recommend reducing the performance core ratio limit, which seems to help with stability in certain games.
The crashes vary in severity depending on the game, with some titles producing an "out of memory" error, others simply exiting out to the desktop, and some locking up a machine entirely. Most of the games affected seem to be based on the Unreal Engine, which could point to a stability issue that Intel needs to address. The only workarounds that seem to improve stability involve manually downclocking or undervolting Intel's processors. Epic Games has suggested changing the SVID behavior to Intel Fail Safe in the BIOS settings of Asus, Gigabyte, or MSI motherboards. Custom PC builders Power GPU recommend reducing the performance core ratio limit, which seems to help with stability in certain games.
Yep, had this issue (Score:5, Informative)
Setting it in the bios still causes the issue to sometimes appear, albeit less so.
I think the board makers are setting some default bios settings too aggressively, by using the Intel XTU program it sets a number of those to safe intel defaults overriding those bios settings.
This caused me months of hair pulling why a number of my games like Spider-Man, Miles Morales, Ratchet & Clank and others would suddenly bomb out at random.
Re: (Score:3)
Re: (Score:2)
The issue a year ago was motherboard manufacturers pushing cpus past their limits, in search of an edge in performance.
The issue now seems to be the completely different issue of motherboard manufacturers pushing cpus past their limits, in search of an edge in performance?
Yup, completely different.
Re: Yep, had this issue (Score:3)
Instead of being condescending, why not explain (especially for those of us who do not keep up on hardware)?
Re: Yep, had this issue (Score:4, Informative)
I'm not being condescending in the least. Take this attitude into a discussion about some of the particle physics or astronomy submissions we get and see what happens when dealing with other posters here.
It was discovered that certain EXPO memory kits were pushing Raphael-X CPUs into dangerous vSoC territory (1.2v+) on certain x670E motherboards. Users were able to reduce vSoC manually to get CPUs into spec without compromising on performance. It was a combination of unclear specs from AMD and foolish motherboard OEMs erring on the side of too much vSoC to ensure compatibility with popular EXPO memory kits. Also some of the mobo OEMs were setting up their power delivery system to misreport vSoC to the CPU's sensors.
AMD pushed some AGESA updates to fix the problem.
It turns out that the issue wasn't exclusive to Raphael-X either. It had been subtley affecting older Raphael CPUs and may have killed or shortened the lives of at least a few 7950X CPUs.
Re: (Score:2, Insightful)
I'm not being condescending in the least.
Yes you were, but I'm glad you learnt to talk like a normal person finally and contribute something positive to Slashdot including adding some neat technical details. Keep it up. This is a good look. Your most recent post is what Slashdot should be like (and used to be like). The rest of your posts are modded down for the toxic attitude you brought.
Please continue contributing to making Slashdot a better place, rather than what you were doing.
Re: Yep, had this issue (Score:4, Funny)
Re: (Score:2)
Slashdot isnâ(TM)t a place for feelings.
Congrats on missing the point of my post. You see I don't give a shit about feelings. But your post isn't as bad as the ones I criticised and let me explain why. You postulated and argument, and gave reasoning. Sure that reasoning means you missed the point, but you formulated a reasoned argument rather than just being a dismissive arsehat, and I applaud you for it. Non reasoned one line insults is what belong on X.
Now youâ(TM)re the one being condescending.
I don't think you understand what that word means. But hey, not everyone has English as a fir
Re: (Score:3)
Re: (Score:3)
Not exactly.
The problem is the third party manufacturers being "too flexible" in their optimized settings instead of capping "optimized defaults" to not exceed the CPU manufacturer's data and only leaving that as advanced settings.
Flying Too Close to the Sun? (Score:5, Interesting)
This seems very similar to a setting JayzTwoCents complained about [youtube.com] over a month ago.
TLDW: Some motherboard manufacturers are setting BIOS defaults that push the hardware too far.
Re: (Score:3)
That's Intel's new recommended conservative settings. :)
That's why they call it "the bleeding edge" (Score:5, Insightful)
Buying the newest, spendiest stuff gets you bragging rights, but you also get to experience all the bugs that haven't been squashed yet. Just ask some of the Cybertruck owners who are having issues.
I've personally always been satisfied with gaming on hardware that is a little behind the curve. Gotta save money where you can in this economy. Besides, I have no delusions about my middle-aged reflexes being anywhere close to what they were when I was a teenager, so I'm not gonna blame the rig when I get pwned.
Re: (Score:3)
Bullshit, Intel has been selling the same basic Golden Cove/Raptor Cove design since they launched the 12900k. They're pushing higher power levels and giving the mobo OEMs too much leeway to overclock CPUs out of the box.
Re:That's why they call it "the bleeding edge" (Score:5, Interesting)
The whole point of buying a K series processor is for the overclockability. If Intel is just doing it on a wink-wink, nudge-nudge basis now "Yeah, you technically can overclock it, but it ain't gonna be stable.", that's kind of shady. You might as well just opt for the non-K version, save a few bucks, and have a more stable system. Win-win.
Re: (Score:2)
Just because it has a k at the end and uses too much power doesn't make it bleeding edge.
Re: That's why they call it "the bleeding edge" (Score:4, Insightful)
Re: (Score:3)
If Intel is just doing it on a wink-wink, nudge-nudge basis now "Yeah, you technically can overclock it, but it ain't gonna be stable.", that's kind of shady.
No. Not a single Intel processor has ever promised to function stably when overclocked. CPU binning is done by vendors to provide the fastest grade silicon within their specs. Intel provide K series processors unlocked so users have the option to attempt to push hardware beyond limits set by Intel themselves. Additionally modern CPUs will dynamically adjust their clock speed to take take advantage of spare overhead (thermal, power, etc) to reach the performance limit of the system. There is not now, and nev
Re:That's why they call it "the bleeding edge" (Score:4, Interesting)
Re: (Score:2)
This is the way.
Re: (Score:3)
For me, I just have no time and resources to play them. That's fine since I have been gaming during the early rad 80s starting with Atari 2600 and arcades.
Motherboard vendors (Score:5, Interesting)
Re: (Score:2)
It'll be to Intel's recommendations. That's just where Intel is these days. It's the only way to get any comparative performance.
Intel or Epic? (Score:3)
I've tried to trace down these kinds of bugs, and they're very nasty. But I wouldn't necessarily jump to blame Intel, sometimes it's the software is making an invalid assumption about the state of the hardware. Compiler bugs can be very hard to find.
Re:Intel or Epic? (Score:5, Interesting)
It's been investigated and found to be related to how modern game engines (notably ue5) are putting more cores to use and exposing problems with how Intel's boost algo responds while under elevated power/clockspeed levels.
It really lies at the feet of OEMs and their willingness to push the 14900ks beyond Intel's advertised power specs.
Re: (Score:2)
Or entirely Intel's problem, alone.
Overclock (Score:1)
The first thing I always do is turn off motherboard vendors overclocking bullshit. Also have ECC and huge margins in current limits and temps with everything running flat out and want to keep it that way. Pointless to push ones luck for marginal gains.
Not the first time. (Score:4, Interesting)
Anyone else remember the Intel chips failing on encryption with 10th gen cores? Most online games would crash at startup as they initialized OpenSSL, used by almost everything. Intel still has part of the workaround, disable Intel extensions, [intel.com] which is still sometimes needed for older games.
It's the stupid BIOS defaults (Score:5, Interesting)
Neither ASUS, Gigabyte, MSI, or ASRock ship their z-series boards with Intel stock defaults. They all do dumb shit to the default settings, which for some can just cause high temps or slow downs, but for others can cause the machine to crash after 10 minutes, or immediately when you start calling AVX.
I had to deal with this firsthand. I thought I had a faulty motherboard from Gigabyte, but in the end it wound up being bad defaults. I wasted 3 days and had to become familiar with overclocking in order to fix their fucked defaults.
Intel needs to get put pressure on their partners to only ship boards with Intel's specifications as the default. They can instead include other profiles that have what the motherboard manufacturer wants.
Re: (Score:2)
Oh, right, it'd never be a single point of failure. Must be everyone else other than Intel getting it wrong.
Re: (Score:2)
Oh Intel is chiefly to blame here. Motherboard vendors are only doing this because they're allowed to do so. Intel has the power to tell motherboard vendors to knock it off, but they do not do so.
Data point (Score:4, Informative)
I've had two occurrences where I found a single core on a multi-core 13th generation chip to be bad. The user reports intermitted crashes in high cpu workloads, e.g. video rendering, and stressing the CPU with Prime95 shows that it really is the culprit.
I've replaced 3 defective CPUs in 20+ years, and 2 of them were this year.