the54thvoidBut even with these mitigations in place, the guy from Puget says:



It mentions that's currently only 5-7 per month, so not huge numbers for them, though expected to rise.

The point is, Puget aren't representative of the DIY build industry, or other system builders using default mobo (or worse, turned up) settings. Using their data is very much cherry picking the most favourable outcome for Intel.

So wait..11th gen are bad too?? Oyy

fevgatosSo you are basically saying that because zen 4 fail to even make it past the shop into the field, that's what, a good thing? I don't want to put words in your mouth so please, correct me if I'm wrong, but are you suggesting that CPUs that fail in a few days or a week are signs of higher quality control than CPUs that fail after 6 months or a couple of years? Cause that's what's happening according to puget's bench. AMD ships such a high amount of highly problematic zen 4 chips that they don't even make it out of the door. They fail within a few days in their stress test. Intel ships cpus that are working fine for the first couple of years.


Puget does CPU reviews as well. In those reviews they clearly state the settings they are using.

Do you intentionally exclude points from posts? Read up, Zen 4 shop failures from other (much larger) retailers show less than 1%. So does Intel. So, Puget's chart simply is an anomaly, and probably a statistical one because the ship relatively few AMD CPU's. I mean Puget doesn't even ship a thousand damn AMD CPU's per year, whereas the ones I mentioned and you ignored ship thousands a week. Go to Alza.sk, look up CPU's and see the 13900K sticking out (look for complaint rate at the bottom right). Among this <1%, if a CPU is DOA it's much easier to get it replaced and life moves on. If a CPU stops working (with a vague error code too) after work has started, you potentially lose hundreds of thousands in revenue if the CPU is used for work and deployed in farms. This is exactly is continuing to happen and of course it's worse, especially since the percentage is far greater than 1% and >25% in most deployments with 24/7 usage.

Also, shop failures aren't CPU's that failed after a week, they mean DOA because they haven't stated otherwise. And that 2% difference in Puget's chart you seem transfixed on is at best a statistical anomaly. Because Puget just don't have the numbers.

It's funny seeing you ignore the actual issue, which is intel degradation under heavy use which is well documented with a boatload of evidence. You point to a chart, where systems aren't used 24/7 and used with different settings than others (where are the Puget reviews stating BIOS settings?) and go on and on about shop failures of Ryzen CPU's when we have much better data of much larger retailers demonstrating that shop failures are a non-issue when it's less than 1%. Please try to include all sources, because among the recent ones i've seen Puget's is literally the least relevant because even if we ignore the server farms, many larger retailers are saying the same thing - total number of issues are very low for all CPU's except an abnormally high rate for 13/14900k's. There's overwhelming evidence out there and i'm not going to continue this discussion because you've been proven incorrect (or asked to not look too deep into this graph for statistical reasons) but you just refuse to accept reality. Can't be bothered to waste my time on this anymore.

I don't want to jump too much into the Reviewer X is a Y Shills, it's always change depending what's best and it's pointless.

The only things i will say about failure, failure rate etc is it's really different something that just stop working or or you receive it and it's dead.

Having experiencing both, the "it just doesn't work" is soooooo much easier to deal with than the "X load just crash randomly". Retailer and manufacturer can gaslight you soo much before you are able to get a replacement.

The situation would have been easier on Intel owner if the CPU just died. Now, they are just going to freak out at every crash wondering if they should try to RMA their CPU.

I think ppl are being dragged into endless Intel vs AMD again, intentionally by someone, trying to damage control.

Isn't the main point here being Intel 's lack of ability to identify & study & solve the problem ?

When ppl dig deep:
Intel had reports of this problem back in 2022.
Intel finally admitted the problem at May 2024.
- Failed to identify the problem for 2 years.

Intel knew the problem in 2022, they've tried to push microcodes with voltage changes and tests.
And still not able to tell we the customer the exact cause of the problem / affect batches / how to test the problem
- Failed to study, or even just try to setup a way of testing the problem.

Intel's way of solving the problem is extremely shady.
First they denied the problem.
Then they blamed the MB manufacturers.
Then they loosely coop with MB manufacturer to push baseline settings that didn't solve the problem.
Then they went silent again and hope things die out.
When the big news finally went off, now all the sudden full damage control, extend warranty, new microcode, but the problem is still, NOT BEING FIXED.
- Failed, just a big FAILED way of demonstrating its problem solving skills.


None of these is related to AMD.
They are all Intel's own doings.
Their stock price goes down because ppl lost trust in Intel.

CrackongIsn't the main point here being Intel 's lack of ability to identify & study & solve the problem ?

I have feeling the trouble they had with this is - which problem? They have instability issue which has the simple answer of "increase voltages" (or lower frequencies), then they have the issue with degrading (supposedly ring) which has the simple answer of "lower voltages", then there is the min voltage increase issue in firmware in their statement. Plus there is the oxidation problem. Oh, and the excessive power thing that was mostly pushed on to motherboard manufacturers - in part quite correctly - also has a role to play in this. All of these are more than likely related and I really do not envy whoever in Intel technical side had to figure their way out of this.

CrackongI think ppl are being dragged into endless Intel vs AMD again, intentionally by someone, trying to damage control.

Isn't the main point here being Intel 's lack of ability to identify & study & solve the problem ?

When ppl dig deep:
Intel had reports of this problem back in 2022.
Intel finally admitted the problem at May 2024.
- Failed to identify the problem for 2 years.

Intel knew the problem in 2022, they've tried to push microcodes with voltage changes and tests.
And still not able to tell we the customer the exact cause of the problem / affect batches / how to test the problem
- Failed to study, or even just try to setup a way of testing the problem.

Intel's way of solving the problem is extremely shady.
First they denied the problem.
Then they blamed the MB manufacturers.
Then they loosely coop with MB manufacturer to push baseline settings that didn't solve the problem.
Then they went silent again and hope things die out.
When the big news finally went off, now all the sudden full damage control, extend warranty, new microcode, but the problem is still, NOT BEING FIXED.
- Failed, just a big FAILED way of demonstrating its problem solving skills.


None of these is related to AMD.
They are all Intel's own doings.
Their stock price goes down because ppl lost trust in Intel.

I feel the issue is not such much about whether they can figure what went wrong, but the way they are handling the issue. For example,
1. When the issue first started in early 2023, Intel seems to be aware of oxidation issues, but still rejecting customers' RMA. Instead they blamed it on microcode which never fixed the issue.
2. As the issue started growing late last year and early this year, Intel immediately turned on their AIB partners by alleging the motherboards were offering too much power/ voltage. In reality, this practice has been around for a very long time, and Intel simply turned a blind eye even when motherboard makers are pushing the upper limits of their "recommendations", because such practices puts their CPU in a better position to compete.

3. This year, we hear even more customers get RMA rejected and CPU failure rate is very high. So much so that developers went onto social media to all Intel out to get their attention.
4. Intel keeps changing their narratives about the issue either shows their incompetence, or their unwillingness to share the root cause which will impact them further. Given that they have chosen to remain silent, just allowed the issue to grow over time, instead of dying down.

I feel if Intel's microcode fix don't resolve the issue this month, its gonna tank their share prices even more. Its clear their partners and customers are feeling frustrated.

londisteI have feeling the trouble they had with this is - which problem? They have instability issue which has the simple answer of "increase voltages" (or lower frequencies), then they have the issue with degrading (supposedly ring) which has the simple answer of "lower voltages", then there is the min voltage increase issue in firmware in their statement. Plus there is the oxidation problem. Oh, and the excessive power thing that was mostly pushed on to motherboard manufacturers - in part quite correctly - also has a role to play in this. All of these are more than likely related and I really do not envy whoever in Intel technical side had to figure their way out of this.

I feel like all things started at the point they setup the boost clock values.
Super aggressive boost clocks -> aggressive firmware VID -> aggressive voltage -> degrading (some oxidation induced some don't) -> need to lower the voltage -> CPU can't run these frequencies at low voltage -> instability -> more voltage -> LOOP

And they can't lower the boost clocks caz the product not meeting the spec leads to product recall, so the cycle just loop and loop.

watzupken4. Intel keeps changing their narratives about the issue either shows their incompetence, or their unwillingness to share the root cause which will impact them further. Given that they have chosen to remain silent, just allowed the issue to grow over time, instead of dying down.

In the name of intellectual honesty, this is standard damage control for a corporation. It's not incompetence, it is an adroit understanding of human psychology.

The very same media spin tactic you admonish has been successfully used by Apple, Google, EA, Microsoft...

Lets not get lost in the desire to armchair-quarterback what we would reveal to the public if we were a billion dollar corporation, as if billion dollar corporations are run like kingdoms. They are not.

remixedcatSo wait..11th gen are bad too?? Oyy

No, only 13th and 14th family.

TheinsanegamerNI wonder if we are going to see Intel re-embrace this Royal core tech, now that both the alder lake cores and meteor lake are running into problems..

If this article is any indication, the answer is a big fat nope.
Intel reportedly dismantles Jim Keller's revolutionary Royal Core project and cancels Beast Lake - NotebookCheck.net News

Look forward to the next few years of Intel positively having their asses handed to them by AMD because as far as I can see, they have no answer to them.

Then again, this is from Moore's Law is Dead so we kind of have to take this with a grain of salt the size of Texas. So, you know... doubts.

Shareholders are suing - www.reuters.com/legal/intel-is-sued-by-shareholders-alleging-securities-fraud-2024-08-07/

DAPUNISHERShareholders are suing - www.reuters.com/legal/intel-is-sued-by-shareholders-alleging-securities-fraud-2024-08-07/

Not surprised. Shareholders don't like the way Intel is being run and the loss of dividends. The 38% drop in shares over the last 5 days isn't the full story yet. There will still be the costs involved with replacing the defective CPUs and the inevitable lawsuit. The lawyers sharks smell blood and are circling.

64KNot surprised. Shareholders don't like the way Intel is being run and the loss of dividends. The 38% drop in shares over the last 5 days isn't the full story yet. There will still be the costs involved with replacing the defective CPUs and the inevitable lawsuit. The lawyers sharks smell blood and are circling.

I wouldn't be shocked if oems drop em