News Intel Arrow Lake and Lunar Lake CPUs are unaffected by crashing issues — Vmin Shift Instability issue only impacts 13th and 14th Gen CPUs

[YSCCC]

What do you mean "have I seen any more burned x3ds"? I haven't seen any burned x3ds or any degraded RPLs.

Of course the issue isn't fixed, cause the issue isn't the CPU degrading. The issue is the CPU has no protections leading it to immolating with over 1.25 vsoc, a thing that should never happen and it cannot be fixed by software updates. Putting a limit on the vsoc provided automatically isn't a fix to the underlying problem, the problem being the CPU has no protections and it catches fire and burns itself and your motherboard. It literally freaking melts, which is absurd don't you think?

It's like having a car blowing up when going over 100km/h. A software update that caps it to 99km/h doesn't really fix the issue, does it?

That is a fix. every freaking car have a rev limiter for precisely that. it is designed to not run over 1.25v Soc, and thus a bios update is implemented for all board vendors to not allow anything higher than that, and that is a fix. you can try put 2v in bios into Vcore of your CPU and it will freaking burn something in your board also.

As this is a fix so no more complains after using the new bios of more X3D ryzens burnt after that.

And remeber, intel up to this point, saying their new CPUs thanks to the new architecture, will not have the degradation issue, havn't come out and say "this is definitely the root cause of the problem, and we fixed the issue by 0x129 microcode", they just said it's a mitigation and stopping the degradation due to excessive VID. Could it be the ring bus? could it be the "extreme" profile? they still didn't confirm or deny any of those. If you didn't find and publish the root casue of it, it will be a hard time to convince those who stay aware of the current gen issues from buying new CPUs from you in the coming 1-2 generations. It's not like the CPU from past 4-5 generations won't serve 90% of users out there. Ppl might not switch to AMD in a heartbeat, but would definitely wait until something confirms they are trust worthy again. Intel needing to give such a press is already indication of significant trend of that is happening

 

[TheHerald]

That is a fix. every freaking car have a rev limiter for precisely that. it is designed to not run over 1.25v Soc, and thus a bios update is implemented for all board vendors to not allow anything higher than that, and that is a fix. you can try put 2v in bios into Vcore of your CPU and it will freaking burn something in your board also.

As this is a fix so no more complains after using the new bios of more X3D ryzens burnt after that.

And remeber, intel up to this point, saying their new CPUs thanks to the new architecture, will not have the degradation issue, havn't come out and say "this is definitely the root cause of the problem, and we fixed the issue by 0x129 microcode", they just said it's a mitigation and stopping the degradation due to excessive VID. Could it be the ring bus? could it be the "extreme" profile? they still didn't confirm or deny any of those. If you didn't find and publish the root casue of it, it will be a hard time to convince those who stay aware of the current gen issues from buying new CPUs from you in the coming 1-2 generations. It's not like the CPU from past 4-5 generations won't serve 90% of users out there. Ppl might not switch to AMD in a heartbeat, but would definitely wait until something confirms they are trust worthy again. Intel needing to give such a press is already indication of significant trend of that is happening

No it won't. If I put excessive voltage on any other CPU it will either thermal throttle to protect itself, or if I have insanely good cooling that can keep from thermal throttling it will slowly (or quickly) degrade. It won't immolate itself, melt the pins, melt the mobo and itself. That's a flaw in the design and it hasn't been fixed.

AMD hasn't identified the root cause of the immolation either, nobody cares, people are still buying amd chips. So why would it stop anyone from buying Intel?

 

[rluker5]

Higher clocks leads to higher voltages, sure, but consumers and coporate clients rely on Intel to spec it correctly so that it don't self destruct or degrade in reasonble time

I didn't know corporate used a lot of unlocked CPUs

It is not only asus and MSI, it happens across the board, MSI isn't immune, so does Asrock, and other like even uses server boards and clock+use ram speed lower than what consumer does. Yet the degradation is still there.

Per your lower statement the level of degradation is unknown.

Why it will not be able to change the bios outhere is a puzzle to me, do you forget when we do in socket upgrade, we need to wait for a UPDATED BIOS release before the board can even recognize the new CPU? If intel did correctly assess the chip, and that it is unsafe to run at the aggressive LLC and power limit which essentially everyone and their dog is using, I can't understand why they do not advise the board partners about that and at least advise them not to do so, if they don't they should release a note/ news that some bios maynot be compliant and may need to do certain tunings for safety, at least for the biggest board vendors. They did none of the above, and yet they don't even found out/admit it's voltage and VID issues or faulty eTVB code without enough safeguard.

Are you aware of what Intel is contractually allowed to demand from motherboard manufacturers? I haven't seen the contracts so I'm not going to assume they can demand anything they want. I don't even know if the motherboard manufacturers can be held liable by causing the degradation and separate instability with their poor power delivery. The motherboard manufacturers may even be able to sue Intel if Intel refuses to sell them chipsets.

Also the LLC setting is not aggressive in the traditional sense. It is failing to deliver stable voltage under load(from too much vdroop) so everything has to be tuned to use more volts to compensate for this. The LLC isn't causing as much droop under higher boost clock single and double core use so the raised volts go even higher.

Given advise to ppl isn't that hard, just give two instructions, cap the vcore supply to 1.55v max, AND set sustained vcore below 1.4v is not difficult isn't it? if you gave such warnings that old boards without safe bios might harm it in the future, I don't see why the boarad partners won't rush out a new, safe bios with correct LLC and voltage settings implemented. It's all excuses.

Doesn't the 129 microcode, combined with heat dissipation limits already do this?
Personally, I don't have a problem keeping even my spikes below 1.4v with my 13900kf. But that is with a max 6GHz single core and 5.5 all p-core. Many have different priorities and chips that are stable at higher frequencies. And are the instructions equally applicable to somebody running a minecraft server 24/7 as someone who games in a GPU limited scenario 8 hours a week? How about what sensors do you determine the voltage from on a motherboard, are they all the same? I know the VRM voltage spikes aren't.

And I'm guessing that Intel is cracking down on motherboard manufacturers power delivery abuses with the upcoming CPUs. That may be the reason behind the statement this article was based on.

As for the degree of disaster compared to Zen 2 or other CPU gens, none have any numbers yet. But this is the only time, from both Intel and AMD, we have seen a widespread frustration, even with coporate customers. individual DIYers are more likely to complain out of their arrogance, but coperate users? Game developers? they need to be extra frustrated about the issue to come out and speak, risking their own discount/offers from the largest CPU maker in the world.

Wasn't the game developer making a dinosaur simulator? Also there was a lot of frustration with Zen 2 with many believing that 1.35v was unsafe on them. It just wasn't such a tech media storm because the tech media was in love with Zen at the time.

But stop distracting focus from the source of the problem: too high of volts from motherboard manufacturer settings at light load causing degradation and often at the same time too low of volts under heavy load causing instability.

Did you know Fram Chasers got his hands on a RPL degraded from a motherboard auto oc and got it running by giving it +50mv. That sounds like garden variety degradation to me.

 

[thestryker]

A lot of RPL users are in socket upgrade using the Z690 boards since ADL, which is supposedly unaffected under the LLC and vdroop "problem", and intel is surely aware of everyone using "dangerous LLC" before they even release the RPL SKUs, it would be easy to enforce all Z790 boards to use what they deemed to be safe, and issue advisory on the "potential dangerous LLC", either not recommend/ said that old boards with a list being incompatible, or required to have specific bios settings to rectify it before it would even be moved to a compatible list, much like how the QVL list done be board partners. They didn't do so.

Instead, they do release the CPU, saying it's compatible with all existing LGA 1700 Z6xx series boards, with mandatory bios update (which still defaults to the high vdroop LLC, no Vcore max cap, and unsafe microcode and TVB behaviour to get those extreme voltages out). Hack AMD can ask the board partners to use the default AMD settings on their boards, don't tell me intel can't.

The thing you, and several others, seem to be overlooking is the actual scale of the problem. This is what was interesting about the information that had been released by Puget. Given that Intel doesn't put many hard restrictions on motherboard manufacturers and the overall failure rate was relatively low the issues likely didn't raise a red flag early on. As I mentioned earlier chances are none of this triggered anything outside of their normal RMA process so it's unlikely anything in depth was done to examine initial failed CPUs.

Even after it was clear that something was specifically happening it wasn't universal and has taken Intel a long time to narrow down as much as they have so far.

Intel has confirmed that thanks to their new architecture, the company's next-generation Core Ultra 200 Lunar Lake and Arrow Lake CPUs aren't affected by the Vmin Shift Instability issue.

So, It basically means they admit that it IS architecture issue that causes the degradation🙄

No they're simply trying to make a clear, broad, line in the sand so people understand without nuance. If it was architectural the problem would be far more widespread and affect all RPL CPUs regardless of power profiles.

 

[mac_angel]

Two things.
One, does the BIOS update flash the microcode on the CPU? or is it purely a BIOS setting?

Two. I was skeptical about the new BIOS update because I spent a LOT of time tweaking my BIOS to be able to run at stock speeds, and stay there under load. Doing that I was able to score a lot higher than many well known overclockers saying they got their computer to run at 6.2GHz all core. While they might have been able to type that in for the BIOS, they were never able to get it to actually run at that speed without thermal throttling. I'd rather have my CPU running at stock and not be thermal throttled than brag about getting a 6.2GHz all core that never actually does that and ends up down clocking below stock from thermal throttling.
Anyway, I did end up trying the new BIOS in hopes that maybe it might help with spikes and maybe I might get a chance to tweak a bit more out of it. I spent all day yesterday trying to just get it to do the same as the older BIOS I had, but never really came close. I'm lucky and glad that I was able to reflash the old BIOS back and get my performance back.
I have my CPU voltage on override and set to 1.35V. Nothing added for System Agent or anything else other than the DDR.

 

[rluker5]

Two things.
One, does the BIOS update flash the microcode on the CPU? or is it purely a BIOS setting?

Two. I was skeptical about the new BIOS update because I spent a LOT of time tweaking my BIOS to be able to run at stock speeds, and stay there under load. Doing that I was able to score a lot higher than many well known overclockers saying they got their computer to run at 6.2GHz all core. While they might have been able to type that in for the BIOS, they were never able to get it to actually run at that speed without thermal throttling. I'd rather have my CPU running at stock and not be thermal throttled than brag about getting a 6.2GHz all core that never actually does that and ends up down clocking below stock from thermal throttling.
Anyway, I did end up trying the new BIOS in hopes that maybe it might help with spikes and maybe I might get a chance to tweak a bit more out of it. I spent all day yesterday trying to just get it to do the same as the older BIOS I had, but never really came close. I'm lucky and glad that I was able to reflash the old BIOS back and get my performance back.
I have my CPU voltage on override and set to 1.35V. Nothing added for System Agent or anything else other than the DDR.

The microcode update is a bios thing. So if you were to take your CPU to a different motherboard, whatever microcode is in the different motherboard's bios is what would be running on your CPU. Your CPU microcode can even be changed at Windows startup by Windows. They used to do this with older chips with the Spectre mitigations, but I don't think they have any bios overrides for newer CPUs. Yet. That means your microcode is probably not 129 unless the older bioses were updated with the newer microcode. HWinfo will tell you which one you are running on the summary page just below your TDP.

And if you have your system already running well at 1.35v you really don't need a cap placed at 1.55v.

I was a bit eager to try out the new microcode for the same reasons as you, but after I saw FrameChasers get not a lot of good and no big improvements on an Actually Hardcore Overclocking oscilloscope check I decided it wasn't worth it to redo my bios and maybe lose the Intel APO functionality on my 13900kf.

I've also found that the last 100mhz I can get out of my CPU before it crashes generally isn't worth it in performance improvements. Maybe I'll get a lucky bench, but nothing consistent to warrant the extra voltage.

Nice to hear somebody getting good results taking a different approach than I would. I haven't set my CPU volts to override since I did Bclk overclocking. I might have to see what I could get out of that. It would definitely simplify things.

 

[YSCCC]

The thing you, and several others, seem to be overlooking is the actual scale of the problem. This is what was interesting about the information that had been released by Puget. Given that Intel doesn't put many hard restrictions on motherboard manufacturers and the overall failure rate was relatively low the issues likely didn't raise a red flag early on. As I mentioned earlier chances are none of this triggered anything outside of their normal RMA process so it's unlikely anything in depth was done to examine initial failed CPUs.

Even after it was clear that something was specifically happening it wasn't universal and has taken Intel a long time to narrow down as much as they have so far.

No they're simply trying to make a clear, broad, line in the sand so people understand without nuance. If it was architectural the problem would be far more widespread and affect all RPL CPUs regardless of power profiles.

We don’t know what contractually Intel could ask board partners to do, but surely make sure them enable Intel defaults by default and let the user opt for the “optimisations” is one of them, as what they’ve been asking and quickly done since they announce investigating in early July/late June.

Puget is, nerfing the CPU by not listening to intel’s recommendation on 13900k and 14900k as they’ve said themselves, they use the older definition of pl1 and pl2 and set them as They think is safe, that makes perfect sense they will have a lower rate of failure.

They had the 0x125 fixing something wrong in eTVB algorithm and then later 0x129 which stops the chip from asking for 1.55v. These shouldn’t have passed the pre release QC as they should have those stress test with even higher voltages and measurements to know around what voltages it will start to have problem and put a software cap on those before release.

As I said before if this is all normal bad luck and board issues, and that Intel can’t tell the board partners what to do by default, and with Intel not disclosing something like “ it’s definitely just the voltages, we have limited it from board partners by default now, so next gen is safe”.

How ppl would just trust them they can make sure the top SKUs won’t repeat the disaster?

 

[thestryker]

We don’t know what contractually Intel could ask board partners to do, but surely make sure them enable Intel defaults by default and let the user opt for the “optimisations” is one of them, as what they’ve been asking and quickly done since they announce investigating in early July/late June.

Not talking hypothetically, but rather stating what they currently do which shaped the investigation and I'm not sure why you keep repeating this. Intel absolutely could mandate whatever they want to there's no question about that as at the end of the day they control the platform and can withhold chipsets etc.

Puget is, nerfing the CPU by not listening to intel’s recommendation on 13900k and 14900k as they’ve said themselves, they use the older definition of pl1 and pl2 and set them as They think is safe, that makes perfect sense they will have a lower rate of failure.

PL1/PL2 settings by themselves don't have anything to do with the failures which has been stated repeatedly. Unless Puget was running the CPUs way below their stock settings and running them like the T series it would have zero impact. It's voltage which is why the CPUs with higher failure rates are ones which were maximum boosting and high power consumption isn't required.

These shouldn’t have passed the pre release QC as they should have those stress test with even higher voltages and measurements to know around what voltages it will start to have problem and put a software cap on those before release.

This is absolutely nonsense given that not every CPU is failing despite them all having the same issue. It's entirely possible none of the tested CPUs failed in this fashion so they had no reason to suspect the specific problem. I also go back to pointing out again if this issue was so simple to figure out why did it take Intel months to narrow it down?

As I said before if this is all normal bad luck and board issues, and that Intel can’t tell the board partners what to do by default, and with Intel not disclosing something like “ it’s definitely just the voltages, we have limited it from board partners by default now, so next gen is safe”.

Intel can make board partners do whatever they want to as they have all the power in that relationship.

You're making the assumption for some reason that every CPU allows high voltage by default when they absolutely don't. Intel expanded the possible VID tables with RPL which is likely related to maintaining high boost clocks. There was clearly a lot that went wrong allowing the voltages to go as high as they can, but none of this is architectural as it has nothing to do with the hardware, but rather how it's configured.

How ppl would just trust them they can make sure the top SKUs won’t repeat the disaster?

Why would anyone who actually understands the problem assume that Intel would configure the next generation to run in exactly the same way despite being a different architecture with different manufacturing node?

I certainly understand being wary of what Intel is putting out, but voltage can be tested for and any in depth review should be at the very least monitoring it with software.

 

[YSCCC]

You're making the assumption for some reason that every CPU allows high voltage by default when they absolutely don't. Intel expanded the possible VID tables with RPL which is likely related to maintaining high boost clocks. There was clearly a lot that went wrong allowing the voltages to go as high as they can, but none of this is architectural as it has nothing to do with the hardware, but rather how it's configured.

Why would anyone who actually understands the problem assume that Intel would configure the next generation to run in exactly the same way despite being a different architecture with different manufacturing node?

I certainly understand being wary of what Intel is putting out, but voltage can be tested for and and in depth review should be at the very least monitoring it with software.

The VID table expansion is the problem and they’re not like having only a week to test and decide the frequencies they can run on. As a few post before I’ve said that it’s the lack of safety margin to danger zone and no hard cap implemented by software microcode or hardware being the issue.
It’s not like Chips can’t do accelerated stress test as in stuffs like engines or speaker drivers, there is expected behaviour curve of the chip subjected to say, 1.8v degrades in 3 hours, 1.7v in 5 days or so to project the safe limit to run 24/7 for the warranted 3 years duration. In house QC to test for potential issues pre release always do such tests and not run it at default and see if it got any error in a week.
Combined with the calculated values and what a few biggest board partners have in their board hardware/bios config you’ll know roughly what vid and frequency you should release the chip on.

Now it took them months to realise the issue is because that the symptoms of the degraded cpus are so similar to things like ram and gpu issues, so that early reports on 13900k don’t even think of it’s a degraded CPU! It runs web browsing fine, other games fine, just in heavy shader loading games like UE5 engine does it started to show the degradation/instability. And that pop up as out of VRAM! 99% of users will just think it’s the new game’s coding issue or NVIDIA driver issues so they don’t even rma the cpu, they RMA everything else until half a year later it’s found to be CPU.

Ppl including myself was wary of the zen 4 x3d literally burning issue as well as all those tech tuber, but those are really at launch and quickly fixed via bios limitation, and ppl did test that and it was fixed, so it kept on selling for its product lifetime, Intel this time around first blame the boards for unlimited pl1/2 wattage (which could, be fair, a factor other than the voltage), and then get the baseline profile the board partners need to follow unclear and such a mess and eventually voltage limit.

They are having the new architecture and new node, this time by TSMC which should be fine, but then the PL2 power isn’t far from 14900k, it’ll be a hard time to convince ppl that Intel won’t have similar behaviour.

Hack they don’t even admit the 2023 oxidation via batches nor announce the batch no. Of those RPL series until it was leaked by the tech tubers. Even then, it’s just a statement of “its past, new batches don’t have the issue” and never disclose the serial no. Of those chips and issue a recall for those affected parts! Those are outright defective or handicapped out of the factory

 

[thestryker]

They are having the new architecture and new node, this time by TSMC which should be fine, but then the PL2 power isn’t far from 14900k, it’ll be a hard time to convince ppl that Intel won’t have similar behaviour.

What is your fixation on PL1/2? These are wattage limits which by themselves have zero impact on voltage limits. The 12700K and 12900K/KS have higher PL2 than the 13600K/14600K and are fine because they don't have the same sort of voltage request configuration. It's the same situation for the mobile and low power SKUs that do use RPL B0 (mobile also have their own RC based J0/Q0) die.

Now it took them months to realise the issue is because that the symptoms of the degraded cpus are so similar to things like ram and gpu issues, so that early reports on 13900k don’t even think of it’s a degraded CPU! It runs web browsing fine, other games fine, just in heavy shader loading games like UE5 engine does it started to show the degradation/instability. And that pop up as out of VRAM! 99% of users will just think it’s the new game’s coding issue or NVIDIA driver issues so they don’t even rma the cpu, they RMA everything else until half a year later it’s found to be CPU.

No I'm talking about from when they acknowledged there was a problem to when the first solutions started rolling out. It was months because the problem isn't something easy or simple to figure out.

Ppl including myself was wary of the zen 4 x3d literally burning issue as well as all those tech tuber, but those are really at launch and quickly fixed via bios limitation, and ppl did test that and it was fixed, so it kept on selling for its product lifetime, Intel this time around first blame the boards for unlimited pl1/2 wattage (which could, be fair, a factor other than the voltage), and then get the baseline profile the board partners need to follow unclear and such a mess and eventually voltage limit.

AMD's problem isn't really comparable aside from the dead CPUs aspect because their problem was due to improper guidance. They knew up front there needed to be an upper bounds voltage limit for SoC voltage and that didn't get coded into the CPUs or part of the BIOS specifications for motherboard manufacturers.

 

[YSCCC]

What is your fixation on PL1/2? These are wattage limits which by themselves have zero impact on voltage limits. The 12700K and 12900K/KS have higher PL2 than the 13600K/14600K and are fine because they don't have the same sort of voltage request configuration. It's the same situation for the mobile and low power SKUs that do use RPL B0 (mobile also have their own RC based J0/Q0) die.

They didn't rule out excessive PL1/2 will degrade the chip, that's why beside the 1.55V they still mandates or advise to keep the PL1/2 to extreme profile, as per common knowledge (which could be wrong), electron migration is accelerated with both voltage and power.

The thing is with similar TDP/PL spec, it suggested that they arn't much more conservative on the power usage this gen also, and in a new process which would logically have different tolerance to max voltage. With the track record of unable to know that the VID table in RPL is dangerous without setting a cap, and that the oxidation via batch isn't recalled don't give faith on high TDP parts from intel.

Note that for those published affected CPUs are all 125W TDP parts for RPL.

No I'm talking about from when they acknowledged there was a problem to when the first solutions started rolling out. It was months because the problem isn't something easy or simple to figure out.

I understand that, but that the issue of overvolting SHOULD be catched during accelerated wear test months before the release of the CPUs, acknowledging the problem months after launch and took another months to identify a problem is a failure in the competency on the design process. Again, why one should trust them knowing the limit of the CPU this time around and not pushing too far? Arrow lake and Lunar lake development and target spec was already in progress for some time when the RPL issues were acknowledged. When they weren't able to test and validate the safe design limit of RPL, why should one trust ARL would be just fine?

And even when the oxidation issue was internally known for months, they did nothing and just wait for RMA, and if the chip luckily survived the warranty or the user don't bother RMA, they just ignore they've shipped a defective product out. Problem is at the coporate culture level and not a single product failure which ppl could just trust the next gen will be fine.

AMD's problem isn't really comparable aside from the dead CPUs aspect because their problem was due to improper guidance. They knew up front there needed to be an upper bounds voltage limit for SoC voltage and that didn't get coded into the CPUs or part of the BIOS specifications for motherboard manufacturers.

Completely agree on this, but still, it's a miss on AMD side of not warning the partners that the X3D stack could burn out exceeding the upper bounds, what I respect was that they at least know the issue and upper bound of their chips and act quickly, that's what matters to consumers IMO

 

[TheHerald]

Again, why one should trust them knowing the limit of the CPU this time around and not pushing too far?

For the same reason you trust any other company They've been making CPUs with no issues for 45 years.

To reverse uno card your argument, since 12th gen is flawless, how can I not trust that arrow lake will be also flawless?

 

[YSCCC]

For the same reason you trust any other company They've been making CPUs with no issues for 45 years.

To reverse uno card your argument, since 12th gen is flawless, how can I not trust that arrow lake will be also flawless?

Since 12th gen is flawless, so that’s why ppl trusted 13th and 14th will be also, which turns out to be not the case, now you have to show that you have taken care of the issue. Reverse fail chief. Try tell your boss it’s fine coz you have been without big issues but slipped some massive mistake and found out only 2 years later.

You have to know that both camp beside this repl gen have done enough accelerated wear test and other tests to ensure they know the limit, and design accordingly, now you have 2 full gen where this slipped passed the radar and just admitted before 15th gen is releasing, so why one would trust them blindly is beyond reasoning.

Plus you have to know that one can still not go AMD but stay on current system until necessary to go for the next gen, much like ppl buying 11th gen systems when 12th is released, and still buying 12th gen at 14th gen release.

Edit:

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

I just feel the same as buildzoid at around 9:00 mark, what is the real issue the backlash is going this far is that they just missed the whole situation during the testing, even at 14th gen release. Either they just have really crazy crappy validation, or they have crazy crappy design team, and yet they didn't admit and rectify in mid 13th gen life cycle.

 

[TheHerald]

Since 12th gen is flawless, so that’s why ppl trusted 13th and 14th will be also, which turns out to be not the case, now you have to show that you have taken care of the issue. Reverse fail chief. Try tell your boss it’s fine coz you have been without big issues but slipped some massive mistake and found out only 2 years later.

If I have a 45 year track record my boss will be fine actually.

"Big issues" is an overstatement. How big? Nobody has specific numbers (besides puget I guess).

The problem is you keep moving the goalposts. Is the real issue (as you just claimed) that they missed the whole situtation during their testing? Okay then, didn't the same happen to amd with x3d? Why isn't there an issue? Now you are going to move the goalposts and say "because they fixed it" (even thought they have not).

 

[YSCCC]

If I have a 45 year track record my boss will be fine actually.

"Big issues" is an overstatement. How big? Nobody has specific numbers (besides puget I guess).

The problem is you keep moving the goalposts. Is the real issue (as you just claimed) that they missed the whole situtation during their testing? Okay then, didn't the same happen to amd with x3d? Why isn't there an issue? Now you are going to move the goalposts and say "because they fixed it" (even thought they have not).

You do see high rank officials step down with decades of perfect track records due to one big mishap like Boeing no?

No, I didn't move the goal post.

It is a big issue, as big as they need to get the fix out and put out these public statement, if it wasn't big, why didn't they say that the failure rate is just in line with the past gen and you'll be fine just sticking to the default settings? it is dying in accelerated rate and UE5 engine expose that much quicker than it used to be.

The problem I keep saying is HOW they handled it at the EOL of the gen and say next gen is fine, it passed the QC, didn't recall the oxidation affected parts as a gesture of QC, which they did identified to be rectified by 2023, don't you think those chips out there, with oxidation issues, even not degraded to be broken should be recalled and replaced??? ok just to wait for RMA?

X3D as thestryker have said, isn't comparable as they actually know that there should be a hard cap, which have been missed out and quickly put back into place. have them denying the issue and release the same fix until now, where 9000 serise Zen 5 released, I will have the SAME reaction as to Intel, and I say it the tenth time to you, cheif, when the X3D burning out issue surfaced, I bought the 14900k, and recommended friends to stay clear from X3D and even recommended the more power hungry RPL until it cleared out, beacuse precisely "it seems safer for the LGA1700 chips from ADL track record", which sadly, it turned out the X3D burning issue (similarly condamned by the clickbait GN or Jayz), were rectified really quickly as they have tested that during development, just missed to enforce that to board partners. Both Asus and AMD did replace the parts promptly and not saying it's user OC them to death.

I tell you once more, it's they've missed the design and claim it's difficult to isolate the issue for 2 freaking years, AKA the whole product cycle of 2 generations to admit and fix the issue, this is tech industry, it is fine to have teething issues in one or two new architecture for initial (software) stability which need bios tuning, or have big one to be fixed in weeks at most, not years.

 

[YSCCC]

If I have a 45 year track record my boss will be fine actually.

Btw that track record is meaningless to certain degree, how many generations of the leadership and management and engineer have changed????

Boeing have a great track record since 1916! that's 118 years, for the first 110 years they have the best track record in the world, but see how it goes after the 737 Max incident? it's how the company's current leadership uphold the value of the company, not how their grandfather do.

 

[TheHerald]

Both Asus and AMD did replace the parts promptly and not saying it's user OC them to death.

See this is the problem right here. You are literally making stuff up for no reason. Did Intel deny RMA and responded with "you oced them to death"? Just stop it dude. You are arguing entirely on bad faith here.

Btw that track record is meaningless to certain degree, how many generations of the leadership and management and engineer have changed????

Boeing have a great track record since 1916! that's 118 years, for the first 110 years they have the best track record in the world, but see how it goes after the 737 Max incident? it's how the company's current leadership uphold the value of the company, not how their grandfather do.

It's an entirely different situation. You need to research the boeing situation to find out what really happened cause it's too complicated to expand here, but basically boeing knew the issue from the start, they just tried to make their airplanes more appealing (lower the costs) by not making pilot retraining necessary which led to the incidents. It was a cost saving measure for boeings clients in order to make their airplanes more appealing.

Intel didn't cut any corners or any cost saving measures. In fact quite the contrary, they are making huge CPUs huge dies at very low costs for the end consumer.

 

[thestryker]

You do see high rank officials step down with decades of perfect track records due to one big mishap like Boeing no?

Boeing isn't exactly a good example as their merger with McDonnell Douglas and subsequent front office takeover by those from McDonnell Douglas led them to 20+ years of failing miserably which was simply capped off by catastrophic incidents. Intel sure started down the same path by pushing financials over all, but at this point only time will tell if they can get off that path or not.

 

[YSCCC]

Boeing isn't exactly a good example as their merger with McDonnell Douglas and subsequent front office takeover by those from McDonnell Douglas led them to 20+ years of failing miserably which was simply capped off by catastrophic incidents. Intel sure started down the same path by pushing financials over all, but at this point only time will tell if they can get off that path or not.

I do know it is not directly comparable, that's why just use as an extreme example of some big F up could do to kill off the reputation, and the new MD management is as much as any director board change in big multi billion firm, once you moved pass the quality first, performance or market share second morale, things go bad pretty quickly as in Murphy's law

See this is the problem right here. You are literally making stuff up for no reason. Did Intel deny RMA and responded with "you oced them to death"? Just stop it dude. You are arguing entirely on bad faith here.

Man you are twisting the point again, did I say intel said "you OCed them to death"? I am saying, they identified that it is a fatal mistake that they didn't remind the board partners the voltage cap must be enforced due to the X3D heat transfer, which is known in their internal test, fixed the issue very quickly AND not denying RMA, any of the above steps gone wrong I would be against them as much as I am against Intel now!!!

It is inexcusable for AMD missed the issue that it is problematic to me in the first 3 months of X3D release to not recommending it to anyone as is intel now, but they proved they fixed it bloody early in the life cycle, not hiding 2 years, you are still trying to muddy the water to say those are worse or equivalent.

I say my point again in point form for you to read:

1) they missed those huge VID requests in their testing, which from your very early post, "every tech savy should know that 1.4v+ isn't safe", so they are below "average tech savy" level for chip safe limit(which they now still cap at 1.55v, and I believe this time round that is the safe cap as they should've do some real testing by now).

2) First 13900k instability discussion was around half year since it's release, and it took another almost 1.5 years since they found out, admit and give some mitigation microcode, which is at the end of productcycle, you can't compare someone fix their fatal issue within weeks vs something lingering for 2 years.

It's either one of 3 things for intel as the buildzoid video above:

1) They missed the very obvious of what the most common board brands can have in default for the VID requests, and for 2 years, not 2 weeks.

2) They know it but in order to keep showing performance gain, they chose to not speak out or require board partners to implement the safeguards

3) They know that it could have as high as 1.6x volt request and supplied voltage, but somehow think it is safe for their product in the reasonable lifetime, which turned out to be not true

None of the above seems acceptable

 

[TheHerald]

I do know it is not directly comparable, that's why just use as an extreme example of some big F up could do to kill off the reputation, and the new MD management is as much as any director board change in big multi billion firm, once you moved pass the quality first, performance or market share second morale, things go bad pretty quickly as in Murphy's law


Man you are twisting the point again, did I say intel said "you OCed them to death"? I am saying, they identified that it is a fatal mistake that they didn't remind the board partners the voltage cap must be enforced due to the X3D heat transfer, which is known in their internal test, fixed the issue very quickly AND not denying RMA, any of the above steps gone wrong I would be against them as much as I am against Intel now!!!

It is inexcusable for AMD missed the issue that it is problematic to me in the first 3 months of X3D release to not recommending it to anyone as is intel now, but they proved they fixed it bloody early in the life cycle, not hiding 2 years, you are still trying to muddy the water to say those are worse or equivalent.

I say my point again in point form for you to read:

1) they missed those huge VID requests in their testing, which from your very early post, "every tech savy should know that 1.4v+ isn't safe", so they are below "average tech savy" level for chip safe limit(which they now still cap at 1.55v, and I believe this time round that is the safe cap as they should've do some real testing by now).

2) First 13900k instability discussion was around half year since it's release, and it took another almost 1.5 years since they found out, admit and give some mitigation microcode, which is at the end of productcycle, you can't compare someone fix their fatal issue within weeks vs something lingering for 2 years.

It's either one of 3 things for intel as the buildzoid video above:

1) They missed the very obvious of what the most common board brands can have in default for the VID requests, and for 2 years, not 2 weeks.

2) They know it but in order to keep showing performance gain, they chose to not speak out or require board partners to implement the safeguards

3) They know that it could have as high as 1.6x volt request and supplied voltage, but somehow think it is safe for their product in the reasonable lifetime, which turned out to be not true

None of the above seems acceptable

Do you realize that every CPU (both from intel and amd) boosts to unacceptable (for me) voltages? Especially on ST boost clocks. It's just so happened to not pay off for intel this time around, exactly like it did with zen 2 and zen 3 having very very similar issues, especially on the high end 3950x and 5950x models. There was a huge issue back in 2020 with zen 3 using as much as 1.5 volt (!!) to hit those single core clocks, with a lot of CPUs actually never even reaching the advertised speeds.

So yes, Intel did ship them with very high - unsafe voltages - but so does AMD, it's a really a moot point. As far as I know only one of them increased warranty by 2 years though. Imagine having a half roasted x3d that is on it's last legs due to high vsoc (before the fix) and it dies just outside warranty, right? That's why I feel safer with intel, i got those 5 years - and the track record that they are willing to extend the warranty - to back me up.

 

[YSCCC]

Do you realize that every CPU (both from intel and amd) boosts to unacceptable (for me) voltages? Especially on ST boost clocks. It's just so happened to not pay off for intel this time around, exactly like it did with zen 2 and zen 3 having very very similar issues, especially on the high end 3950x and 5950x models. There was a huge issue back in 2020 with zen 3 using as much as 1.5 volt (!!) to hit those single core clocks, with a lot of CPUs actually never even reaching the advertised speeds.

So yes, Intel did ship them with very high - unsafe voltages - but so does AMD, it's a really a moot point. As far as I know only one of them increased warranty by 2 years though. Imagine having a half roasted x3d that is on it's last legs due to high vsoc (before the fix) and it dies just outside warranty, right? That's why I feel safer with intel, i got those 5 years - and the track record that they are willing to extend the warranty - to back me up.

That is unacceptable FOR YOU, not even for the CPU, they did internal testing andvalidation to determine what voltages are safe for the lifetime, which should be done given they have months and dozens if not hundreds of engineering samples to do all those validation, not you or me. and Intel is not even 1.5, they post to 1.65, and even after the fix, it still have a fair share of CPUs reported to constantly hit 1.5 sustained in some bios, and 1.55v capping isn't uncommon.

And you just count how many reports on reddit given the X3D and 14900k are on the market since the issue and still surfacing or not? they put the 2 year extended warranty is only because they know they F up big, and to salvage their reputation at the 2 year mark they need to get it to 3 years from now on for the earliest part, that is the 13900k released 2 years ago.

And still, the already publicly announced oxidation issue, that is DEFECTIVE out of the box and did they issue the batch no. affected and recalled??? No, that is the track record of the current management of Intel, just saying "it is history". You do aware that the 5950x have some backlash from users also no? but even debauer's test showed at constant, 24/7 1.45v only 1 out of 3 of his samples showed any degradation, that alone is much lower than say wendell or the dino game developer have in their 14900ks keep degrading in similar time frame no?

Btw, melting is almost instant process and half melt chips still fails, degradation is electron migration, much more subtle and will really shorten the life. So are you suggesting, given Intel have similar or even slight worse MT performance to the AMD Zen 5 X3D you will be still buying Intel? since they gave you more confidence?

 

[NinoPino]

If I have a 45 year track record my boss will be fine actually.

Not 45 but about 25 because they did a similar mistake with P III 1.13.

 

[TheHerald]

That is unacceptable FOR YOU, not even for the CPU, they did internal testing andvalidation to determine what voltages are safe for the lifetime, which should be done given they have months and dozens if not hundreds of engineering samples to do all those validation, not you or me. and Intel is not even 1.5, they post to 1.65, and even after the fix, it still have a fair share of CPUs reported to constantly hit 1.5 sustained in some bios, and 1.55v capping isn't uncommon.

And you just count how many reports on reddit given the X3D and 14900k are on the market since the issue and still surfacing or not? they put the 2 year extended warranty is only because they know they F up big, and to salvage their reputation at the 2 year mark they need to get it to 3 years from now on for the earliest part, that is the 13900k released 2 years ago.

And still, the already publicly announced oxidation issue, that is DEFECTIVE out of the box and did they issue the batch no. affected and recalled??? No, that is the track record of the current management of Intel, just saying "it is history". You do aware that the 5950x have some backlash from users also no? but even debauer's test showed at constant, 24/7 1.45v only 1 out of 3 of his samples showed any degradation, that alone is much lower than say wendell or the dino game developer have in their 14900ks keep degrading in similar time frame no?

Btw, melting is almost instant process and half melt chips still fails, degradation is electron migration, much more subtle and will really shorten the life. So are you suggesting, given Intel have similar or even slight worse MT performance to the AMD Zen 5 X3D you will be still buying Intel? since they gave you more confidence?

1 out of 3 means 33%. That seems much higher than any legit numbers I've seen for Intel.

No melting isn't an instant process. The safety guards on the x3d are gone first, and that leads to the cpu melting eventually. If it was an instant process then everyone that used over 1.25 vsoc wouldnt even boot past the bios, which isn't the case. The chip was slowly cooking over a period of days. Have fun with half cooked cpus.

 

[Amdlova]

Intel Notebook cpu on desktop lol its crashing free with a 45w before that melting the sun!

 

[YSCCC]

1 out of 3 means 33%. That seems much higher than any legit numbers I've seen for Intel.

No melting isn't an instant process. The safety guards on the x3d are gone first, and that leads to the cpu melting eventually. If it was an instant process then everyone that used over 1.25 vsoc wouldnt even boot past the bios, which isn't the case. The chip was slowly cooking over a period of days. Have fun with half cooked cpus.

1 out of 3 having a mild degradation under 1.45v non stop 24/7 for 6 months, that means no throttling at 1.45v, 8 hours a day for 1.5 years, which is unrealisticly low, on intel, you could try similar. nobody have done such test on intel yet. and as much as you wanna twist the plot to this gen AMD being worse, sales will tell. Remember AMD is evil, don't go for X3D even if it performs

and by the way, for the X3D melting, melting is a process you reaching a temperature threshold where the material itself turns into liquid and deforms, hence shorting, you apply the voltage, either it is hot enough to melt the silicon and deform, and smokes out, or it isn't, so it is instant melting, but not instant death from voltage, it isn't electromigration, so it isn't accumulative from time, it is though, each time individual cooked to below the temp and the silicon is still fine, or it reaches the melting point and it melts, whenever the melting happens, it will not boot, or it is just hot enough and not melt or flow. And that is precisely why after the bios fix AMD have done the X3D chips don't have reported deaths appearing here and there on reddit, but 13900k users with those unsafe bios keep degrading and slowing dies at approximately same period