Problem on Intel 13th & 14th Gen CPUs.

[SharpeXB]

I installed the Asus BIOS update for the Z790. I actually get very slightly better Cinebench scores than before. Like 1-5 points better, tiny. But not worse. DCS runs completely fine. It says I’m actually GPU bound at my screen limit of 117Hz.
Even on busy MP. Temps on the CPU during the test were 88 in MT and 65 in ST 

[LucShep]

2 hours ago, SharpeXB said:

I installed the Asus BIOS update for the Z790. I actually get very slightly better Cinebench scores than before. Like 1-5 points better, tiny. But not worse. DCS runs completely fine. It says I’m actually GPU bound at my screen limit of 117Hz.
Even on busy MP. Temps on the CPU during the test were 88 in MT and 65 in ST 

Performance is not a problem with the new microcode. Nor temperatures.
It's the voltages. The frigging 1.5v+ voltage spikes b!tch slapping the poor processor so hard, to the point of slow unpredictable degradation.  They're still there.

That's the price to pay for overambitious boosts on 13th/14th gen => stupic stock high voltages and spikes that slowly eat away your processor's life.

Even the reliable 12th gen i9 and i7, which aren't affected at all by these degradation issues, will also slowly degrade if you also force them stupid high voltages.


People with 13th and 14th Gen 65W+ CPUs, even after installing the new microcode, have two choices:

1.  Keep things stock on clocks, boosts and voltages, and enjoy it as it is. 
   The new microcode at least ensures a slower degradation than before. But it will very likely still occur, sooner or later.  (in six months? five years? ...who knows?)
   If it indeed goes "kaput", well... let's hope it's within the RMA time period and a new replacement is accepted.
   
   or
    
2. Lock the P-Cores (i.e, limiting them all to same max possible clock), so that there is no single/dual-core boost. Reduce the CPU core voltage to about 1.35v (or lower). 
   Worst scenario, you might have to lower the "all P-Core clock" a hundred MHz less than stock. Which won't make any difference to whatever use (unless it's competitive benchmarking?) and the degradation is drastically mitigated.  ....as should've been done imediately when this problem appeared (IMO).

 

Edited August 17, 2024 by LucShep

[SharpeXB]

Eh if it lasts long enough that I’m not replacing yet another CPU/mobo in a year or so I’m good. Right now it looks like DCS doesn’t even find the limits of the 14900 

Hard to imagine what people who aren’t gamers that stay on top of this stuff would do. No normal people ever update their MB BIOS. Of course they probably don’t stress this thing in a way that’s going to cause any trouble and they’ll likely buy a new machine before the chip goes kablooey.

Edited August 17, 2024 by SharpeXB

[Aapje]

I think that people who get a 14900 do stress their system.

Worst off are those with little tech knowledge who get a prebuilt and who might not even know about this, as they don't follow the tech news.

[Devrim]

Do the Intel devs aware of that?

People who know the issue or not, will eventually start to encounter the dyeing indicators due to degradation such as BSOD, CTD, or other crash types, right?. And after they research the cause, they will be surprised and realize that it's Intel's fault for months or years.

How will Intel guys expect customer retention? How will they ensure customer trust? They will lose a huge amount of customer, fans etc. How can they afford this?
Other than releasing micropatches, they need immediately to take action for customers' trust.

At first, for me. 

[SharpeXB]

4 hours ago, Aapje said:

I think that people who get a 14900 do stress their system.

Worst off are those with little tech knowledge who get a prebuilt and who might not even know about this, as they don't follow the tech news.

Well this affects all the 13-14th gen CPUs so certainly users beyond gaming. And I wouldn’t have known about this without the post in this forum. I suppose I might have seen something on social media eventually. 

[Czar66]

11 hours ago, SharpeXB said:

Eh if it lasts long enough that I’m not replacing yet another CPU/mobo in a year or so I’m good. Right now it looks like DCS doesn’t even find the limits of the 14900 

Hard to imagine what people who aren’t gamers that stay on top of this stuff would do. No normal people ever update their MB BIOS. Of course they probably don’t stress this thing in a way that’s going to cause any trouble and they’ll likely buy a new machine before the chip goes kablooey.

 

Once you render a scene on any 3D program or render a video, these CPUs are hitting full throttle until the task gets done. For people who constantly renders videos and people who work with 3D, this issue is major.

5 hours ago, Aapje said:

I think that people who get a 14900 do stress their system.

Worst off are those with little tech knowledge who get a prebuilt and who might not even know about this, as they don't follow the tech news.

The 14900 owner just need to do rendering for either video or 3D to stress their CPUs, and that's very common for high end CPU owners.

These tasks pushes CPUs to 100% no matter what, hence a fairly reliable benchmark tool.

[Devrim]

"According to Intel, a too-high CPU voltage supply can occur not just under load, but also when idle." - PCWorld.

[Nightdare]

On 8/17/2024 at 12:26 PM, Aapje said:

Worst off are those with little tech knowledge who get a prebuilt and who might not even know about this, as they don't follow the tech news.

 

Intel's thoughts on the matter:

https://community.intel.com/t5/Processors/July-2024-Update-on-Instability-Reports-on-Intel-Core-13th-and/m-p/1617113

Quote

For users who purchased 13th/14th Gen-powered desktop systems from OEM/System Integrator - please reach out to your system vendor's customer support team for further assistance.

For users who purchased boxed/tray 13th/14th Gen desktop processors - please reach out to Intel Customer Support for further assistance.

 

I always build my own system and order from reputable shops, but if I wasn't following the YT tech channels I'd not have known about this issue (admitted, I found it here as well)

Taking a big community hub: Steam -Their news only allows searches for games they carry

We would expect intel to 'heads up' their importers and get in contact with the shops to warn the consumers that they need a bios/microcode upgrade, perhaps even contact big news outlets, instead of everyone having to do their own digging

 

when my bike needed a recall, I got a letter to give me notice from the importer and also a call from the dealer, Intel is doing none of that and, if you ask me, is still hoping most CPU's live out their use/warrantee even without the changes
 

[Ramstein]

I have Asus Z790-H ROG Strix Motherboard, with Intel 13900K CPU and updated Bios. so far no issues. I just lucked out, I must have just missed the windows for issues, although Intel refuses to give us the exact dates of the batches that had issues. They just said 2023 in the manufacturing process. My CPU was week 2, March 2023. I don't overclock though. My ram runs at 6600 in XMP. I have it set to Intel settings in the Bios.

 

first time ever I ran Cinebench, no clue if this is good. score = 33293

 

(I left out it's a K model CPU)

Edited August 21, 2024 by Ramstein

[stonewall197922]

9 минут назад, Ramstein сказал:

I have Asus Z790-H ROG Strix Motherboard, with Intel 13900 CPU and updated Bios. so far no issues. I just lucked out, I must have just missed the windows for issues, although Intel refuses to give us the exact dates of the batches that had issues. They just said 2023 in the manufacturing process. My CPU was week 2, March 2023. I don't overclock though. My ram runs at 6600 in XMP.

 

first time ever I ran Cinebench, no clue if this is good. score = 33293

 

Don`t have 13th or 14th gen CPU. But as far as i read and talked to friends in IT department, problem occurs only on "K"  models of CPU.

[Nightdare]

15 hours ago, Ramstein said:

I just lucked out, I must have just missed the windows

15 hours ago, Ramstein said:

I don't overclock though.

 

Only time will tell if running these CPU's without added stress (OCing) will make them last the warranty

For all you know, you've got a ticking timebomb that already has been damaged

 

Without the microcode revision (and probably even then), the Intel settings are still gonna spike your CPU

The Intel Bios Ver is just Intel being wrong about what failed their CPU's (and putting the blame on board partners), they failed because they are faulty production

[Devrim]

I've never seen any problem (BSOD, any error message, CTD or something - nothing) with my CPU.

What does that mean? Can I consider that there's no serious degradation on my CPU, or what?

[LucShep]

13 minutes ago, Devrim said:

I've never seen any problem (BSOD, any error message, CTD or something - nothing) with my CPU.

What does that mean? Can I consider that there's no serious degradation on my CPU, or what?

You can never know for sure, it may or may not happen.

That said, I notice a lot of people overstressing and getting ansiety for this, it's out of proportions, IMO.

You have a device (whatever hardware part), use it and enjoy it - that's what its purpose is after all.
It has a warranty, use it if needed - it breaks and isn't your fault, RMA must be served for a replacement.

Basically what I said in my previous reply, and quoting:

On 8/17/2024 at 4:51 AM, LucShep said:

Performance is not a problem with the new microcode. Nor temperatures.
It's the voltages. The frigging 1.5v+ voltage spikes b!tch slapping the poor processor so hard, to the point of slow unpredictable degradation.  They're still there.

That's the price to pay for overambitious boosts on 13th/14th gen => stupic stock high voltages and spikes that slowly eat away your processor's life.

Even the reliable 12th gen i9 and i7, which aren't affected at all by these degradation issues, will also slowly degrade if you also force them stupid high voltages.


People with 13th and 14th Gen 65W+ CPUs, even after installing the new microcode, have two choices:

1.  Keep things stock on clocks, boosts and voltages, and enjoy it as it is. 
   The new microcode at least ensures a slower degradation than before. But it will very likely still occur, sooner or later.  (in six months? five years? ...who knows?)
   If it indeed goes "kaput", well... let's hope it's within the RMA time period and a new replacement is accepted.
   
   or
    
2. Lock the P-Cores (i.e, limiting them all to same max possible clock), so that there is no single/dual-core boost. Reduce the CPU core voltage to about 1.35v (or lower). 
   Worst scenario, you might have to lower the "all P-Core clock" a hundred MHz less than stock. Which won't make any difference to whatever use (unless it's competitive benchmarking?) and the degradation is drastically mitigated.  ....as should've been done imediately when this problem appeared (IMO).

 

 

 

Edited August 22, 2024 by LucShep

[Nightdare]

6 minutes ago, Devrim said:

What does that mean?

 

No one can say

 

6 minutes ago, Devrim said:

Can I consider that there's no serious degradation on my CPU, or what?

 

Yes, you can consider that, and still be proven wrong over time

 

 

I may be too negative in claiming the new microcode won't prevent affected CPU's from degrading even more, but I still think Intel should have investigated which production period was below standard and RMA those batches

 

 

 

 

[SharpeXB]

32 minutes ago, Devrim said:

I've never seen any problem (BSOD, any error message, CTD or something - nothing) with my CPU.

What does that mean? Can I consider that there's no serious degradation on my CPU, or what?

When my 13900K failed (and I’m just assuming it was from this) the problem presented itself as just severe performance loss (single digit frame rates in DCS), no crashes or anything else though.

Edited August 22, 2024 by SharpeXB

[Ramstein]

intel does not get specific about what part of the zillions of diodes and transistors in the cpu layers have been contaminated... if anyone knows circuits, there are limits to voltages and knee voltages and currents that are gates... if the contamination causes gates to malfunction then instead of a gate drop of a zener diode for voltage at 6.6 volts drops under or over the  that design that that part of the circuit will cause that part to fail. Considering there are zillions of these components in the microscopic circuits, all kinds of trouble in the cpu will occur. Anything from a hiccup to massive failure. Static electricity alone blows silicone chunks out of circuits. I have seen the results of static electricity through an electron microscope. I had a problem in engineering from sodium contamination in die for micro circuits in submarine sonar circuits in Aerospace manufacturing (this was ages ago). It causes the a/d conversion to go bad.

[Aapje]

@Rammstein

The main issue is not contamination, but a design defect, coupled with them pushing the silicon too hard.

[Devrim]

On 8/22/2024 at 6:13 PM, SharpeXB said:

When my 13900K failed (and I’m just assuming it was from this) the problem presented itself as just severe performance loss (single digit frame rates in DCS), no crashes or anything else though.

Ouch. I too encountered 1 FPS performance a few times tho. Did you check the mouse? Even the mouse was moving in 1 FPS.

[SharpeXB]

24 minutes ago, Devrim said:

Ouch. I too encountered 1 FPS performance a few times tho. Did you check the mouse? Even the mouse was moving in 1 FPS.

If you mean the mouse polling issue, that one was rather recent for me. I only discovered that a few days ago, it wasn’t a problem back then. 

[72westy]

I've held off on the x129 update (MSI) given it is a beta. I'm running a 14700F on Pro B760 VC Wifi II

[Ramstein]

On 8/22/2024 at 4:50 PM, Aapje said:

@Rammstein

The main issue is not contamination, but a design defect, coupled with them pushing the silicon too hard.

I didn't say it was all the issues, but oxidation is contamination,...(just saying I have seen similar thing happen) but along similar issues... the high voltage rails being pushed higher than the design for that current...

Edited September 2, 2024 by Ramstein