News Intel finally announces a solution for CPU crashing errors — claims elevated voltages are the root cause; fix coming by mid-August

[stuff and nonesense]

Unfortunately, Intel CPUs sold to OEMs are warrantied by the OEM. Only retail-boxed CPUs are warrantied directly through Intel. I would imagine each big OEM will probably have some arrangement worked out with Intel, to partially compensate for any products with abnormally high failure rates. But, even then, the OEM will still have some skin in the game.

So, regardless of what Intel says, you're really at the mercy of the OEM and their warranty policies. Dell apparently wants to wait and see how many of its customers the microcode update will take care of, in order to minimize their own cost of having to deal with a flood of RMAs.

It’s sad to say but yes, Dell are trying to reduce their liability even though the manufacturer is admitting there is a fault.

In the UK the vendor is responsible for the liability, because of this I buy all the components for a build from the same place. If there is a fault, little to no problem, I bought them all from “you”, please help. It helps that I have a 15 year relationship with the company and have returned 2 pieces, both bad AIO coolers (cooler master master coolers).

 

[Reality_checker]

Regarding this issue, I can only share my own personal findings.

13700K with a Gigabyte Z790 Aero G: bought last year, updated it to the (then) latest BIOS and have had no issues. Vcore never crosses 1.4V.

13600K with an Asus ProArt B760 Creator WiFi: bought last month. Out of the box, I could see Core VIDs as high as 1.585V. After updating to the latest BIOS, these can still go up to 1.51V.

Vcore maxes out at 1.415V without undervolt. I have also noticed the strange phenomenon of E-cores requesting higher voltage than P-cores (as reflected by their VID readings).

Now, I don't know if this is an Asus issue, or an Intel microcode issue, or just a poorly-binned 13600K, but I am waiting for the BIOS update to come out. I probably won't touch my Gigabyte motherboard since it's behaving well.

I've also noticed that on the Asus motherboard, if you choose the "Asus Performance Enhancement 3.0", it actually sets Tjmax to 90 degrees, sets appropriate power limits and removes the current limit (you should set the current limit back to 200A, or 307A if you are running a higher-end SKU), which is actually better than the 100 Tjmax, unlimited power draw and 400A Iccmax at the "Intel default settings".

Crazy! I am very displeased with the Asus BIOS.

 

[TheHerald]

Regarding this issue, I can only share my own personal findings.

13700K with a Gigabyte Z790 Aero G: bought last year, updated it to the (then) latest BIOS and have had no issues. Vcore never crosses 1.4V.

13600K with an Asus ProArt B760 Creator WiFi: bought last month. Out of the box, I could see Core VIDs as high as 1.585V. After updating to the latest BIOS, these can still go up to 1.51V.

Vcore maxes out at 1.415V without undervolt. I have also noticed the strange phenomenon of E-cores requesting higher voltage than P-cores (as reflected by their VID readings).

Now, I don't know if this is an Asus issue, or an Intel microcode issue, or just a poorly-binned 13600K, but I am waiting for the BIOS update to come out. I probably won't touch my Gigabyte motherboard since it's behaving well.

I've also noticed that on the Asus motherboard, if you choose the "Asus Performance Enhancement 3.0", it actually sets Tjmax to 90 degrees, sets appropriate power limits and removes the current limit (you should set the current limit back to 200A, or 307A if you are running a higher-end SKU), which is actually better than the 100 Tjmax, unlimited power draw and 400A Iccmax at the "Intel default settings".

Crazy! I am very displeased with the Asus BIOS.

There should be an option somewhere that sets a global voltage cap. It should be called IA VR LIMIT or something like that. Set it to let's say 1.35v - and then the CPU will boost like normally as high as it wants but until it hits the voltage cap.

 

[Reality_checker]

There should be an option somewhere that sets a global voltage cap. It should be called IA VR LIMIT or something like that. Set it to let's say 1.35v - and then the CPU will boost like normally as high as it wants but until it hits the voltage cap.

I've updated the BIOS, lowered the loadline from mode 3 to mode 2 and set a slight undervolt. That's working to keep the Vcore voltage where I want it to be, but the VIDs are still high (the voltage that the cores are requesting) so I am curious how this will look after the update.

 

[TheHerald]

I've updated the BIOS, lowered the loadline from mode 3 to mode 2 and set a slight undervolt. That's working to keep the Vcore voltage where I want it to be, but the VIDs are still high (the voltage that the cores are requesting) so I am curious how this will look after the update.

The vids are always going to be high, those are irrelevant. Vcore is what matters

 

[Reality_checker]

Buildzoid just made a video about IA VR Voltage Limit that you mentioned @TheHerald . The setting is called the same on Gigabyte and Asus motherboards at least.

View: https://www.youtube.com/watch?v=2G-Y0yDSfeA


Around 12:40 in the video.

 

[TheHerald]

Buildzoid just made a video about IA VR Voltage Limit that you mentioned @TheHerald . The setting is called the same on Gigabyte and Asus motherboards at least.

View: https://www.youtube.com/watch?v=2G-Y0yDSfeA


Around 12:40 in the video.

He is copying me

 

[Reality_checker]

If I use IA VR Voltage Limit (=1400) on my wife's 13600K, it doesn't seem to draw more than 166W. So it could just be a below average bin.

I'll still update the BIOS and check VIDs then, because the problem could be two-fold (i.e. related to microcode logic as well).

 

[TheHerald]

If I use IA VR Voltage Limit (=1400) on my wife's 13600K, it doesn't seem to draw more than 166W. So it could just be a below average bin.

I'll still update the BIOS and check VIDs then, because the problem could be two-fold (i.e. related to microcode logic as well).

VID will not change, VID is how much the CPU asks from the mobo. Doesn't mean the mobo will provide those values, there is VDROOP involved and the DC light loads. It's very complicated but generally speaking, completely ignore VID.

How much does it score on CBR23 - set it to high priority - with your 1400 vr limit?

What I would do in your position, first of all never update the bioses. Newer doesn't mean better. I would just turn off the ST boost and cap the voltages to 1.3 volts. Disable CEP if it's activated and undervolt the chip ^^.

 

[-Fran-]

More information from Wendell, but from the Linux side and his own investigation.

View: https://www.youtube.com/watch?v=WgFTTP1HxpQ


Totally worth a watch.

Looks like there could be "T" Intel SKUs affected as well.

Regards.

 

[8iosat]

Obiously It was the dilatory nature of the user to provide a proper cooling system. So Intel was forced to take drastic measures and a small error occurred that will be fixed with a bios update, and management engine maybe. Never ever are all this chips defekt. They need a CPUID Code update which is in the biosupdate which ones have to update. There are so many methods to bring an defekt intel Chip back to work. With 99% acid over the contacts so at least 1 millimeter minimum was removed. and even defekt memory controller can be fixed. Intel chips are very robust, much more than AMD chips are, which wouldn´t exist without Intel.

 

[stuff and nonesense]

Obiously It was the dilatory nature of the user to provide a proper cooling system. So Intel was forced to take drastic measures and a small error occurred that will be fixed with a bios update, and management engine maybe. Never ever are all this chips defekt. They need a CPUID Code update which is in the biosupdate which ones have to update. There are so many methods to bring an defekt intel Chip back to work. With 99% acid over the contacts so at least 1 millimeter minimum was removed. and even defekt memory controller can be fixed. Intel chips are very robust, much more than AMD chips are, which wouldn´t exist without Intel.

Got to fix this for you.

When the IBM PC was first built IBM wanted a second source for the CPU, AMD were that second source, Intel got the contract, PCs were built with Intel based chips.. intel took off but without AMD Intel may not have got the gig that made them big.

Intel and AMD are entwined, symbiotic or parasitic, depends on your point of view.

 

[beyondlogic]

intel atm has literally burned to many bridges with consumers lol.

 

[dimar]

I wonder if my 13900K get less chances of damaged if I use Asus high performance BIOS profile along with Noctua 140mm dual fan large cooler. I'm running 13900K with CMK64GX5M2X6800C32 at 6400 CL32 1.45v which runs completely stable.

 

[The Historical Fidelity]

I wonder if my 13900K get less chances of damaged if I use Asus high performance BIOS profile along with Noctua 140mm dual fan large cooler. I'm running 13900K with CMK64GX5M2X6800C32 at 6400 CL32 1.45v which runs completely stable.

Unlikely. Take these following opinions how you will but Wendell at Level1Techs has been doing a deep dive with IT of a number of big Intel customers like investment firms and hedge funds, and based on all the pooled data, he suggests dropping maximum clock multiplier to 53x and maximum DDR clock to 4800 mega transfers. ActuallyHardCoreOverclocking recommends setting VR VID Limit to 1400 which will prevent the CPU from requesting voltages above 1.4V.

 

[thestryker]

The Puget Systems analysis of their own experience is really interesting and it looks like if motherboards were running in line with Intel specifications these issues might have taken longer to be apparent. This seems to be one of those times where Intel's decision to not enforce specifications helped consumers in a very real way.

Puget Systems' Perspective on Intel CPU Instability Issues

You may have heard about instability issues with Intel Core 13th and 14th Gen desktop processors. The issue has been gaining more attention as time goes on. I am posting to share what we've experienced here at Puget Systems, and what we're doing about it.

www.pugetsystems.com

side note: I wish the way they're handling the situation was how all the big players did.

 

[TheHerald]

The Puget Systems analysis of their own experience is really interesting and it looks like if motherboards were running in line with Intel specifications these issues might have taken longer to be apparent. This seems to be one of those times where Intel's decision to not enforce specifications helped consumers in a very real way.

Puget Systems' Perspective on Intel CPU Instability Issues

You may have heard about instability issues with Intel Core 13th and 14th Gen desktop processors. The issue has been gaining more attention as time goes on. I am posting to share what we've experienced here at Puget Systems, and what we're doing about it.

www.pugetsystems.com

side note: I wish the way they're handling the situation was how all the big players did.

Highly ironic, but seems that if you actually care about system stability, Intel is basically your only option, as long as you are not running default mobo settings.

So, I guess I have to keep using Intel until amd gets their instability issues addressed.

 

[Ogotai]

Highly ironic, but seems that if you actually care about system stability, Intel is basically your only option, as long as you are not running default mobo settings.

So, I guess I have to keep using Intel until amd gets their instability issues addressed.

thats a good one, i have 6 comps here, 5 are amd 5k based. and none of them are unstable at all, no matter what i run on them.

because of the instability/crashing issues intel now has, 5 co workers/friends were looking to upgrade, 3 were looking at intel, 2 at amd, now, all of them are looking at AMD, and 2, already upgraded. they dont want to take any risk of their comps being almost unusable, or dealing with intels rma process based on what a few have already gone though cause of the issues with 13th and 14th gen.

 

[bit_user]

Highly ironic, but seems that if you actually care about system stability, Intel is basically your only option, as long as you are not running default mobo settings.

So, I guess I have to keep using Intel until amd gets their instability issues addressed.

Apparently, you didn't look at the data very closely, because Ryzen 7000 has the second-lowest field failure rate, only behind that of Alder Lake.

 

[TheHerald]

Apparently, you didn't look at the data very closely, because Ryzen 7000 has the second-lowest field failure rate, only behind that of Alder Lake.

And zen 3 has the highest, by far. In fact the field failure rates of zen 3 are ~ equal to the total failure rates of the 13900k. That is freaking insane.

But why does that actually matter? Buying a CPU in the DIY market - i will experience the total failure rates, not the field ones. *

*As long as I'm running all chips with their respective recommended settings.

 

[TheHerald]

thats a good one, i have 6 comps here, 5 are amd 5k based. and none of them are unstable at all, no matter what i run on them.

because of the instability/crashing issues intel now has, 5 co workers/friends were looking to upgrade, 3 were looking at intel, 2 at amd, now, all of them are looking at AMD, and 2, already upgraded. they dont want to take any risk of their comps being almost unusable, or dealing with intels rma process based on what a few have already gone though cause of the issues with 13th and 14th gen.

Your own limited experience doesn't mean anything. I have a 13900k and a 14900k, both no issues. Does that mean there is no issue just because I haven't experience any? Of course not.

My 6900hs laptop on the other hand is highly unstable. Does that mean anything? No, not much really. You need a sample size of couple of hundreds to reach any conclusions. Pudget has that sample size and their numbers make it obvious that Intel is way more stable than AMD. Again, that's not my opinion, that's what pudget data suggests.

 

[Ogotai]

maybe not.. but i do have quite a few friends and co workers who havent mentioned any stability issues either... maybe we are all just lucky with ryzen 5k.....

 

[TheHerald]

maybe not.. but i do have quite a few friends and co workers who havent mentioned any stability issues either... maybe we are all just lucky with ryzen 5k.....

You realize that with the failure rate being 4 to 5% only 1 out of 20 will have an instability issue, and then he needs to pinpoint it specifically to the CPU. Yeah, obviously the vast majority of people won't experience any issues with either intel or amd.

I have a 4090 with one of those melting cables. Still going strong. Doesn't mean much, does it?

 

[Reality_checker]

Apparently, you didn't look at the data very closely, because Ryzen 7000 has the second-lowest field failure rate, only behind that of Alder Lake.

What I see is that only Intel's 11th gen was worse than AMD's past and current offerings.

 

[stuff and nonesense]

Maybe titled with a little hyperbole but…

View: https://www.youtube.com/watch?v=b6vQlvefGxk