Intel 13900k and 14900k chip failures

[LunarMist]

I'm not seeing the nexus between those industries, but I'm not much concerned about aircraft. I use many CPus every day yet fly only a few segments per monthly nowadays. Even at most it was not more than 72 a year. When the planes fail it is usually detectable.
CPUs are far more problematic since failures can be hidden. Are the XeNON family afflicted like the consumer parts?

 

[Handruin]

My understanding from dd's point was less of a nexus and more of a parallel in outcomes. Both major industries seem to suffer from poor sr. management and also a shareholder shift to seek total profit in ethically gray ways (and some criminal) at the large expense of product quality and R&D that has lead to their downfall in poor product delivery and corporate outcome.

 

[ddrueding]

Asus released a beta BIOS for my board that included the 0x129 microcode update that theoretically stops my chip from cooking itself. After installation and resetting to factory defaults system was far from stable. After disabling all but 4 of the efficiency cores (part of my old spec) it seems stable?

 

[ddrueding]

They did extend the warranty to 5 years, so that is good?

 

[Mercutio]

I think in this case the greater concern comes with what happens to the CPU, not the motherboard. Asus can't really make Intel do anything on that front.

 

[ddrueding]

True, but the motherboard BIOS is the delivery vehicle for the CPU microcode. Reportedly Intel released the "fixed" microcode called "0x129" to motherboard manufacturers a while ago, and beta BIOSes with the change are now becoming available. While eventually I think a BIOS with the new microcode is a no-brainer, I'm tempted to let my CPU slowly break a little bit longer until a stable BIOS/microcode can be sorted out.

 

[Mercutio]

I barely trust Asus production BIOSes let alone beta versions. I can see why you'd want to wait.

 

[ddrueding]

Yup, kicking myself for going ASUS in the first place. Just as we were talking about Noctua in the fans space, there are companies that charge a premium for the luxury of not worrying about it (Seasonic for power supplies, for example). Some part of my brain still puts ASUS in that camp, though they haven't deserved it in years.

 

[Mercutio]

It's a tough call but I think I still have a generally lower opinion of MSI than Asus. For as much as I hate the company that hasn't ever successfully completed an RMA with me or sold motherboards that don't let end users change the shutdown temperature, it's still not the company that two different times deployed motherboards with capacitors so bad they went off like a bullet and dented the sides of computer cases.

It's really, truly gratifying to see the world catch up to the "Asus has criminally bad warranty support" issue, though.

 

[LunarMist]

I used Asus many years ago until they started dying, then started using the Gigabit, then MSI. The last six have been MSI with no problems yet. They have been the practical choice for me now on AM4/AM5 due to the inadequate number of slots/lanes/ports on the others. I suppose the GB brand OC's a little better, but I'm not pushing it that much.

 

[Mercutio]

I feel like there's a relationship between seeing big tiddy bondage goth demon babes on product packaging and things I don't care to buy and that is definitely also a factor in avoiding MSI.

Asrock seems to hit just right as the vanilla and inexpensive option for almost everything and so far as I know, nobody ever seems to have a bad thing to say about 'em. Gigabyte is my usual go-to. It might not be perfect but it hasn't really let me down, either.

 

[LunarMist]

I feel like there's a relationship between seeing big tiddy bondage goth demon babes on product packaging and things I don't care to buy and that is definitely also a factor in avoiding MSI.

Is your defect rate high with the recent AM4 and AM5 boards or some older technology? You have demonstrated an unnatural dislike for the WD HDDs based mostly on ancient history. Is this a similar scenario?

 

[jtr1962]

You have demonstrated an unnatural dislike for the WD HDDs based mostly on ancient history.

I would hope their SSDs are no worse than any other brand. Whatever issues Merc had with WD HDDs, it's (hopefully) safe to say their SSDs aren't affected given that they're a completely different technology.

 

[Mercutio]

Is your defect rate high with the recent AM4 and AM5 boards or some older technology? You have demonstrated an unnatural dislike for the WD HDDs based mostly on ancient history. Is this a similar scenario?

Every time I've given WD a chance, EVERY SINGLE TIME, I get either a justifiably suspect product or an outright bad product. It ran out of chances long ago. I'll buy something in Ultrastar trade dress. I'll buy Sandisk memory cards, but WD is less a data storage option than an expensive replacement for concrete.

I have not purchased an MSI board in the AM4 era, although I have used a couple Zx90 Intel boards from sometime before Intel started doing the E-core nonsense.

 

[LunarMist]

Yup, kicking myself for going ASUS in the first place. Just as we were talking about Noctua in the fans space, there are companies that charge a premium for the luxury of not worrying about it (Seasonic for power supplies, for example). Some part of my brain still puts ASUS in that camp, though they haven't deserved it in years.

True, but the motherboard BIOS is the delivery vehicle for the CPU microcode. Reportedly Intel released the "fixed" microcode called "0x129" to motherboard manufacturers a while ago, and beta BIOSes with the change are now becoming available. While eventually I think a BIOS with the new microcode is a no-brainer, I'm tempted to let my CPU slowly break a little bit longer until a stable BIOS/microcode can be sorted out.

Is your CPU damaged such that the lower, safer settings from the new BIOS are not adequate to drive it now?

 

[jtr1962]

I'll buy Sandisk memory cards, but WD is less a data storage option than an expensive replacement for concrete.

Just wondering if you had similar outcomes with their SSDs, assuming you even tried them.

 

[LunarMist]

Every time I've given WD a chance, EVERY SINGLE TIME, I get either a justifiably suspect product or an outright bad product. It ran out of chances long ago. I'll buy something in Ultrastar trade dress. I'll buy Sandisk memory cards, but WD is less a data storage option than an expensive replacement for concrete.

I have not purchased an MSI board in the AM4 era, although I have used a couple Zx90 Intel boards from sometime before Intel started doing the E-core nonsense.

I don't have your quantities, but over 50 WD hard drives since the He era and no outright failures. I'm only using >8TB He-sealed drives, mostly 8-14TB from new retail Elementals and lately 18-20TB WD DC drives. I don't count the slow and hot, <12TB 2020s air-powered drives or older Blue and Green drives from ancient times. Those slow WD drives were not the greatest.

 

[LunarMist]

Just wondering if you had similar outcomes with their SSDs, assuming you even tried them.

WD bought Sandisk in 2016, so many WD and Sandisk SSDs are now the same if they sold them in the same configuration. There are fewer SanDisk SSDs nowadays - mostly some SATA III and a few NVMe. Due to the silly, inconsistent SanDisk brand nomenclature (Extreme, Extreme Pro, Ultra, etc.) it's often difficult to determine the equivalent WD model number, so it's better to just get the numbered WDs.

 

[Mercutio]

Just wondering if you had similar outcomes with their SSDs, assuming you even tried them.

I've never deliberately used a WD-branded SSD. A few WD Green drives have passed through my hands in Lenovo or HP consumer laptops. Weirdly, the same PCs are just as likely to have what I believe are very good Kyoxia, Samsung or Micron drives as crappy (no write cache) Silicon Power or Teamgroup branded ones. I can only guess that procurement really is "whatever comes first on the truck."

 

[fb]

I had a pretty "fun" bug on my work laptop (1255U CPU) when it was almost new: All cores rän on 400 MHz, all the time. It got fixed in a BIOS update but it was unusable for a few days. (Teams + Outlook starts automatically..)

Regarding ASUS, I haven't kept up with the motherboard market in the last few years, so I still bought one last year. I had to RMA a ASUS graphics card around 13-14 years ago and it was a really painful process.
Maybe I'll give Gigabyte a try next time!

 

[Mercutio]

I'm absolutely shocked you got an RMA back from Asus at all. I've sent them things. Best they've ever done as far returning anything back to me was sending me back my same box, unopened.

 

[LunarMist]

My policy is not to RAM computer parts, especially mainboards. Perhaps I would with the defecative iNtel CPUs. I wonder if people have been returning them yet.

 

[Handruin]

According to some reports out there, people have been having issues with RMA attempts for their Intel CPUs with Intel denying their requests.

Everything I've read over the years about ASUS being a huge PITA with support and RMA, I've been avoiding their products going forward.

 

[LunarMist]

I would not touch an Intel for another generation.

If you buy a DELL then they will deal with it, or is a DELL so slow it doesn't cause failure?

 

[Handruin]

I don't know about the situations with Dell. I agree, I'd assume they would handle the warranty issues in those cases. I'm not familiar with how Dell would manage the settings/timings. I would expect their Alienware lineup would run at advertised speeds which means they're susceptible but I don't have any experience with their consumer grade systems.

The stories I've read were about individual system builders like us who have had denied claims.

 

[ddrueding]

Is your CPU damaged such that the lower, safer settings from the new BIOS are not adequate to drive it now?

I hope not? I didn't want it to be damaged further, so I updated preventatively. I now have it fairly stable, but only after disabling all but 4 of the e-cores and hyperthreading. Performance is still fine for all the things I do, but mentally my off-ramp is the 9950X3D whenever that is available.

 

[LunarMist]

I have not had a CPU in over a decade that was not completely stable. Are you running some crazy OC speeds or voltages?

 

[ddrueding]

I normally experiment with overclocking in the first weeks of building a new computer. Just to see what kind of headroom it has. Not all samples are created equal, and in the past I've had chips that could handle on OC without any voltage increase, or handle stock speeds at a lower voltage. But stability is more important than all of this. If the machine won't remain stable under stress for days then I don't run it.

Currently I'm running my chip at stock voltage, stock clocks, with hyperthreading and a bunch of efficiency cores disabled and it seems fine. Not ideal.