• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

Intel still has not fixed the crashes with 13th and 14th gen CPUs

Leonidas

AMD's Dogma: ARyzen (No Intel inside)
Interesting that this is your first post in the thread. What is your take on the topic?
I've had my i5 for ~20 months with no issues, so its not of my concern. I also never recommended anyone buy the CPUs that ended up having the issue. Glad I went with the i5 in 2022, still more MT and better than AMD gaming performance for the money.

I will be upgrading this fall, and will buy whatever CPU is the best for me, just like I did in 2022. Issues from an old product will not stop me from buying one in the future. Both CPU manufacturers have had issues recently.
 
Last edited:

JohnnyFootball

GerAlt-Right. Ciriously.
I was looking forward to Arrow Lake, but this makes me far less interested.

You can't fuck up this bad and have my trust. Arrow Lake could still be good, but right now there is no reason to switch from AMD.

I might...unlikely though....switch from my 7800X3D to a 9700X if there is no performance differemce, but I can take advantage of higher speed memory and lower temps.

My 7800X3D runs hotter...even with undervolting....than my 7700X did with PBO -30 offset and an 85W power limit.
 
I was looking forward to Arrow Lake, but this makes me far less interested.

You can't fuck up this bad and have my trust. Arrow Lake could still be good, but right now there is no reason to switch from AMD.

I might...unlikely though....switch from my 7800X3D to a 9700X if there is no performance differemce, but I can take advantage of higher speed memory and lower temps.

My 7800X3D runs hotter...even with undervolting....than my 7700X did with PBO -30 offset and an 85W power limit.
The 7800X3D is designed to run at 89°C at 100% usage to push the CPU cooler to the max and maintain the highest possible frequency for a limited time.

Personally, I'm not worried about CPUs getting too hot. In 2007, people told me not to OC my core 2 Q6600 because it would degrade faster, but I did it anyway. My Core 2 Q6600 OC'ed to 3.2GHz has been running at 85-90°C since 2007 until now (my niece played on that CPU until recently) and still works. In 2012 I bought Core i7 3770K and OC'd it to 4.5GHz and this CPU also still works despite running at 80-90°C for all these years.

I think the reason why modern i7 / i9 degrade so fast (on a monthly basis according to some users) is just because of too high voltage and maybe design flaw, not high temperatures.
 
Last edited:

Allandor

Member
Currently I don't think Intel can fix this problem without heavily reducing voltage even more than they have now or without a new hardware revisio. It still degrades fast it seems and the damage is already done, so expect almost every i9 & i7 k processor to fail in near future.
So if they give you a new CPU it will still degrade. So the problem isn't that easy to solve.
 

JohnnyFootball

GerAlt-Right. Ciriously.
I've had my i5 for ~20 months with no issues, so its not of my concern. I also never recommended anyone buy the CPUs that ended up having the issue. Glad I went with the i5 in 2022, still more MT and better than AMD gaming performance for the money.

I will be upgrading this fall, and will buy whatever CPU is the best for me, just like I did in 2022. Issues from an old product will not stop me from buying one in the future. Both CPU manufacturers have had issues recently.
Don't you have a 12600K? I mean...like what would you upgrade to? It would be festively stupid to upgrade to upgrade to a 13600K or 14600K since those CPUs are also in the mix of having these issues as well (although at a much lower rate)
 

marquimvfs

Member
Is there anyway to test for this particular issue? My 13900k seems fine in gaming and benchmarks but better to be aware of the issue before it starts causing issues.
In the 2 clients I had with the issue, I could observe the following:

CPUs were extremely hot, both were constantly trotting in any highter-than-light operation, all were watwecooled and it happened in matter of seconds executting such tasks. One of the client had severe flaws with a i9 13900K, it throttled with the CPU limited in bios to 85W, even with a corsair 360 watercooler.

Long file copies from any CPU controled PCIe ssd to any disk (ssd, hd, sata, pcie) on the machine caused BSOD when the copy took more than 5 minutes, like moving a game folder from CPU controlled SSD to a secondary sata disk.

Intermittend BSOD on CPU bound gaming.

Sometimes, even installing games made the pc show a BSOD, namely on intensive decompression tasks.

It started slowl, not immediately. I've built the i9 machine in 03/18/2023 and the client started to notice some problems only 04/13. The first crash occured 05/11, with COD Warzone. I beginned the troubleshooting by updating bios, then capping the processor to lower wattages then proceded to lower the rams to jedec defaults. Sometimes the problem happened right before the restart, other times it worked as intended, its completely random, but the failures become more frequent in a matter of weeks, and the i9 client ended with a corrupted OS. Both cases the CPU were replaced. i9 owner sold the machine after the new CPU arrived.

I think that, if you haven't experienced anything so far, no suspicious crashes, BSODs and no excessive heat, you are in an allright zone. Just pay attention to temperature and random crashes in CPU intensive tasks. Try to use your pc more, to see if it manifests some failure within the warranty period.
 
Last edited:
At least now you are being specific, because users before you did not include the monitor and speakers in their measurements, and review sites do not include them either. If your PC pulls only 100 watts, I have no idea why you were so upset about the high power consumption of AMD CPUs. You wrote that your 7950x3D sucks power like crazy at idle (those are your own words), and I dont think 100W idle on a 16C32T CPU is that bad. Dude, your old i7 was only 4C8T, so no wonder it was even more power efficient. If you bought a modern i7 like 13700K / 14700K, not to mention i9, your PC would use even more power compared to your AMD PC.

I checked again how much youtube draws and made sure to run a 4K stream). It's between 96-98W (6-8W more compared to IDLE), and from my perspective, that's nothing to cry about. You bought a high-end PC and are worried about such low power consumption, your behavior makes no sense whatsoever.
I'm upset because 1) my Intel PC pulled only 75-78w sitting idle on the desktop, but more importantly 2) it also drawed this same amount WHILE using the PC for light loads. The thing didn't boost instantly to max the second you touch the mouse or play a web browser stream like the AMD rig does. Suddenly that 20-25w gap jumps to nearly 50w when the PC goes from 98w to 125w. That is unacceptable and I'm sick of people defending it. AMD needs to get their shit together for low workloads. You didn't even acknowledge what I said about how their mobile parts are all monolithic dies for low power consumption. It's real, just because you don't want to recognize that doesn't change this fact.
 
I'm upset because 1) my Intel PC pulled only 75-78w sitting idle on the desktop, but more importantly 2) it also drawed this same amount WHILE using the PC for light loads. The thing didn't boost instantly to max the second you touch the mouse or play a web browser stream like the AMD rig does. Suddenly that 20-25w gap jumps to nearly 50w when the PC goes from 98w to 125w. That is unacceptable and I'm sick of people defending it. You didn't even acknowledge what I said about how their mobile parts are all monolithic dies for low power consumption. It's real, just because you don't want to recognize that doesn't change this fact.
Dude, you said that your PC's power consumption goes from 198W to 232W just by moving the mouse, so you asked me to use a digital wattmeter to see for myself this unacceptable power consumption under light workloads. I did exactly what you asked, but I only saw a 6W difference (from 90W idle to 96W during mouse movement). Other light tasks like YT, MP3, or Bluray playback also only draw 6-8W more. My measurements did not confirm your findings, sorry dude.

AMD needs to get their shit together for low workloads.
And you need to realize that modern i7s / i9s PC consume significantly more power than your old i7 7700 PC, including light workloads. Do you really think something like i9 14900K PC (including monitor and 5.1 sound system) pulls only 150W IDLE? A modern Intel PC would consume even more power than your AMD PC as confirmed by every tech site, so if you want to criticize AMD for unacceptable power consumption under light workloads, you should criticize both, AMD and Intel.
 

bender

What time is it?
That guy is a hack…

Canadian too...

shudder-nerea-rodriguez-tremble-mp3cpl8d1yon6x5f.gif
 

Leonidas

AMD's Dogma: ARyzen (No Intel inside)
Don't you have a 12600K? I mean...like what would you upgrade to? It would be festively stupid to upgrade to upgrade to a 13600K or 14600K since those CPUs are also in the mix of having these issues as well (although at a much lower rate)
Never had a 12600K. I've had 13600K for ~20 months now, overclocked past 14600K in all that time with no issues.

I don't make minor CPU upgrades.

I will definately be upgrading when Arrow Lake comes out to either an Arrow Lake CPU or a Zen5 CPU, depending on which best meets my needs.

I will most likely end up getting an i5 Arrow Lake (or higher) as I care about great gaming and great MT performance per dollar, something which the i5/i7 excelled in the past few generations... and I like overclocking, something which I've had more fun on Intel platforms than AMD recently.

The only Zen5 CPU that interests me is the 9950X, since 6-8 cores is too low and I don't want to deal with the 2x6 core issues AMD 12 cores have...
 
Last edited:

Leonidas

AMD's Dogma: ARyzen (No Intel inside)
All I know is that I switched to AMD when Ryzen came out (1600) and never regretted it. Step into the light, Intel chuds.
I've used Zen1-Zen3 and went with 13600K back in 2022, because it was the better product for my needs vs. Zen4 at the time.

And this fall, I will again choose whichever CPU is the best option for me. Only a fool would buy a CPU because it was AMD or Intel, I only buy which CPU is the best for my needs.
 
Last edited:

MarkyG

Member
RMA'd my 13900K this week. I'm seriously thinking about a platform change to AMD as their 9000 series CPUs are supposed to release at the end of this month. It's annoying as I was really pleased with the 13900K, bought it Dec 2022.

However, since May this year, it's been increasingly unstable. Shit like any Chromium based browsers just erroring on (seemingly) random webpages, typing comments into YT videos would exhibit really weird lagging when typing. Facebook Messenger would error, Discord crashing also. Really frustrating. Glad I didn't overclock it.

The amount of hoops I had to jump through with Intel was pretty insane too, but I got there eventually. Sad.
 
The only Zen5 CPU that interests me is the 9950X, since 6-8 cores is too low and I don't want to deal with the 2x6 core issues AMD 12 cores have...
DSOG is always testing how modern games run on different CPU configurations, and there are not many games that can use more than 6C12C CPUs in any meaningful way. Even your 13600K CPU has no problem running the latest games despite having 6 P cores (I'm not counting E cores because they have no use in gaming and can actually degrade performance), so it makes me wonder why you even want 12 or more CPU cores? Are you a content creator who needs an insane number of CPU cores for your work?
 
All I know is that I switched to AMD when Ryzen came out (1600) and never regretted it. Step into the light, Intel chuds.

Same. At the time I was doing some rendering on the side and the i5 4690 wasn't cutting it. Instead of going with an i7 I decided to just do a new build when Ryzen released. The R5 1600 cut render time from 90 minutes to 20, and I could still use the PC for YouTube or something during that 20 minutes. Was a huge jump for the meagre price. The original 1600 I bought is still going today in my brothers missus PC, works well enough with a 2060 for a cheap 1080/60 build.
 

Xellos

Member
I have a i7-13700 (non-k) and could never get it stable beyond 4800 cl40 ddr5, but I've still been happy with the performance. Always thought it was a bit of bad luck since Intel says it supports up to 5600 ddr5, but maybe it's related to this? I've been conservative with power limits (105/150) so hopefully that helps. Been over a year without a crash.
 

Leonidas

AMD's Dogma: ARyzen (No Intel inside)
DSOG is always testing how modern games run on different CPU configurations, and there are not many games that can use more than 6C12C CPUs in any meaningful way.
6 modern cores is enough for the majority of games, but I don't just game on my PC. I didn't build a PC worth ~$2000 to only play video games on it.

Even your 13600K CPU has no problem running the latest games despite having 6 P cores
And I will still upgrade because my rig will be 2 years old in several months, its time for an upgrade... I'm not sticking around holding onto last gen stuff when the new stuff is going to be a lot better. As soon as Arrow Lake launches I will buy either a 9950X or one of the Arrow Lake K-SKUs and at least one game I played recently could see a huge benefit in areas (where I saw CPU limited drops to ~80 FPS).

it makes me wonder why you even want 12 or more CPU cores? Are you a content creator who needs an insane number of CPU cores for your work?
I don't need an insane number of cores, but I am absolutely not going backwards or making a side-grade in terms of MT. My next CPU will have to be at least 20% better in rendering compared to what I have today.

And I wouldn't even get a 12 core from AMD because I'm not dealing with 2x6 core clusters, which could impact certain scenarios. If I get Zen5, it'd have to be 9950X. Intel doesn't have the same issue with its CPU design.
 
Last edited:

Bojji

Member
Digital Foundry weighs in: (20:12)
They've experienced stability issues themselves


And yet there are Intel fanboys that think this is bullshit.

I'm on AMD since 2019 (12 years on Intel before that) and have 0 problems with stability.
 

qbxwhi

Member
Is there anyway to test for this particular issue? My 13900k seems fine in gaming and benchmarks but better to be aware of the issue before it starts causing issues.

try playing a UE5 game.
if it crashes ,then you are affected.
so, something like Fornite.
 

EekTheKat

Member
Apparently - something is happening on the mobile front as well.


Intel claims it's not the same bug, but the fact that the relatively new Asus G16 will be migrating from Intel to an announced up and coming AMD cpu is concerning.
 


Silicon lottery turned into a silicon Russian roulette.

In all seriousness, this is unacceptable and goes beyond the Intel vs. AMD fanboy wars (really, it's only one person who's engaging in it). Intel slow walked its response to address the chip failures and the solution it has provided is more akin to kicking the can down the road than actually fixing the (multifactorial) problem.
 

kruis

Exposing the sinister cartel of retailers who allow companies to pay for advertising space.
Is there anyway to test for this particular issue? My 13900k seems fine in gaming and benchmarks but better to be aware of the issue before it starts causing issues.

One way to test the problem is to do a full install of the Nvidia Gefore driver (inc. Geforce Experience). If you're able to install the drivers ten times without an error screen, you should be good (for now).
 

kruis

Exposing the sinister cartel of retailers who allow companies to pay for advertising space.
RIP intel



This was obviously recorded before Intel came out with the news that they'd found the root cause of the instability and it would be fixed with a microcore update in August. Linus last comment was about the EVB bug that Intel had found earlier.
 
This was obviously recorded before Intel came out with the news that they'd found the root cause of the instability and it would be fixed with a microcore update in August. Linus last comment was about the EVB bug that Intel had found earlier.
It seems that the root cause is a design flaw that cannot be fixed with an update. I think Intel just wants people to think there's hope for their CPUs. Maybe this patch will improve stability and slow down the degradation, but it will not stop it or reverse the degradation. If they succeed in slowing down this degradation, new CPUs will come out soon and people will forget about the problems with 13 / 14 gen.
 

simpatico

Member
This was obviously recorded before Intel came out with the news that they'd found the root cause of the instability and it would be fixed with a microcore update in August. Linus last comment was about the EVB bug that Intel had found earlier.
Do you think the micro code updates are going to completely fix the problem?
 

kruis

Exposing the sinister cartel of retailers who allow companies to pay for advertising space.
Do you think the micro code updates are going to completely fix the problem?

Hopefully, but the fact that Intel isn't sure they have now found the root cause of the instability issues is worrying.


– Intel observes a significant increase to the minimum operating voltage (Vmin) across multiple cores on returned affected processors from customers.
– This increase is similar in outcome to parts subjected to elevated voltage and temperature conditions for reliability testing.
– Factors contributing to this Vmin increase include elevated voltage, high frequency, and elevated temperature.
– Even under idle conditions at relatively cool temperatures, sporadic elevated voltages are observed when the processor is resumed from low power states in order to service background operations before entering a low power state again.
– At a sufficiently high voltage, these short-duration events can accumulate over time, contributing to the increase in Vmin.
– Intel analysis indicates a need to reduce the maximum voltage requested by the processor in order to reduce or eliminate accumulated exposure to voltages which may result in an increase to Vmin.
While Intel has confirmed elevated voltages impact the increase in Vmin, investigation continues in order to fully understand root cause and address other potential aspects of this issue.
– Intel is validating a microcode update to limit VID requests above 1.55V as a potential future corrective action, targeted for production release in mid-August to NDA customers.
Early testing by Intel on a small number of benchmarks indicates minimal performance impact due to this microcode change.
– While this microcode update addresses the elevated voltage aspect of this issue, further analysis is required to understand if this proposed mitigation addresses all scenarios.
This microcode update, once validated and released, may not address existing systems in the field with instability symptoms.
Systems which continue to exhibit symptoms associated with this issue should have the processor returned to Intel for RMA.
––Intel—
 
Last edited:


The failure rate is so bad in Minecraft servers, that 30% of Intel CPUs would die in 1 to 2 months.



What the fuck?! That is NOT acceptable. Holy shit. I didn't think it was THIS bad.
 

simpatico

Member
Hopefully, but the fact that Intel isn't sure they have now found the root cause of the instability issues is worrying.


– Intel observes a significant increase to the minimum operating voltage (Vmin) across multiple cores on returned affected processors from customers.
– This increase is similar in outcome to parts subjected to elevated voltage and temperature conditions for reliability testing.
– Factors contributing to this Vmin increase include elevated voltage, high frequency, and elevated temperature.
– Even under idle conditions at relatively cool temperatures, sporadic elevated voltages are observed when the processor is resumed from low power states in order to service background operations before entering a low power state again.
– At a sufficiently high voltage, these short-duration events can accumulate over time, contributing to the increase in Vmin.
– Intel analysis indicates a need to reduce the maximum voltage requested by the processor in order to reduce or eliminate accumulated exposure to voltages which may result in an increase to Vmin.
While Intel has confirmed elevated voltages impact the increase in Vmin, investigation continues in order to fully understand root cause and address other potential aspects of this issue.
– Intel is validating a microcode update to limit VID requests above 1.55V as a potential future corrective action, targeted for production release in mid-August to NDA customers.
Early testing by Intel on a small number of benchmarks indicates minimal performance impact due to this microcode change.
– While this microcode update addresses the elevated voltage aspect of this issue, further analysis is required to understand if this proposed mitigation addresses all scenarios.
This microcode update, once validated and released, may not address existing systems in the field with instability symptoms.
Systems which continue to exhibit symptoms associated with this issue should have the processor returned to Intel for RMA.
––Intel—
Reads like something I would tell my boss to buy time. I know full well the swirling desperation the engineer felt when he ran the micro code, lowered voltage and still got a crash. It hurts.
 
Top Bottom