It is the 1.7v stuff Intel does that is the reason for such bad degradation.
2 Likes
yeah new guidance is abs max 1.55 but reality is more like 1.4 is safe? I donāt really know, and I donāt think Intel does either
1.7v has probably degraded your CPU already
itās not just the voltage but also load lines and what voltage your CPU requests and what it actually gets aew two entirely different things
5 Likes
My chip topped out at requesting ~1.732 in R23, when OC is applied. Guess Iām building a new PC soonā¦
I will reboot and lock lower values. The voltage requested from my CPU has been relatively high from the beginning. Mostly I did nothing extreme, except for a few benchmarks.
Dunno what documentation you were reading, because 1.6v was already degrading chips. Normal voltages range <1.53v, with 1.4v being more normal. 1.7v is enough to degrade and eventually kill any CPU made in the last 20 years.
It was on Intelās homepage:
https://www.intel.com/content/www/us/en/content-details/743844/13th-generation-intel-core-and-intel-core-14th-generation-processors-datasheet-volume-1-of-2.html
Datasheet one should contain that information. However, to me itās currently unavailable.
The average voltage under partial load is 1.5V - 1.55, sometimes ~1.56.
Iāve configured 1,7 as the point, where the system does not increase the voltage anymore, no matter what the CPU may request. That is a feature on my ASUS TUF mainboard.
Note that the system wonāt post, when the latter value is set too low.
EDIT:
It looks like, when I drop the mutli 50x, that the demand in voltage sinks by ~200mV. (E-Cores deactivated)
A āmax of 1.72vā on Intelās spec is ludicrous, thatās high enough to potentially just kill chips outright or in short order. As Wendell said itās likely already damaged your processor. Continued use at that voltage is going to kill it. There is probably a temperature maximum involved, for example subzero cooling would make 1.72v safe for use. But it certainly is not safe at room temps.
Intel provides the normal operating voltage range in the same PDF spec you reference, but apparently Intel has removed it from their website. The VID range for your chip is around 1.4-1.53. See this table of base VIDS from ~150 14700K parts via Igorās Lab https://www.igorslab.de/wp-content/uploads/2023/11/refreshbin_i7_vid-1536x995.png
Just because the instrument cluster in a low HP car goes up to 180mph or higher doesnāt mean the vehicle can safely achieve it without significant wear or outright damage, often itās just the same speedometer cluster thatās also used in vehicles with much more powerful engines in them. Most motherboards allow you to set voltages that will instantly kill almost any of the hardware plugged into them too, as one example because there are subzero overclockers that want the functionality.
Intelās support suggested to change a few settings, which actually helped quite a bit.
I have not run my CPU at those voltages, for more than a short test. Maybe it was also my mainboard, that liked āoptimizeā the voltages and currents.
Running the chip at maximum voltage, that is. as you wrote, just a bad idea.
Some funny anectada on user RMA experience with intel support. When I said in my last post that intel probably intends to defeat customers in detail, I was just half serious.
I already feel the āuse anything to deny RMAā instructions in the air.
https://www.reddit.com/r/hardware/comments/1ei1zvm/intel_has_denied_two_of_my_14900k_rmas/
Any idea if this is isolated case or pattern of behavior? I wonder if The steve or better @wendell noticed that already and making contact, if true it might nice interview material.
EDIT: Tech jesus was already informed via email, intel support is in for a walloping then 
if this pans out.
I love the if you send it in for RMA, we will confiscate it with no reimbursement, so don even try buddy.
OP is lucky that the original resellers are not actively trying to fuck him over and will accept rma.
Question is, will intel try to fuck them over just hard as they did threaten OP?
1 Like
TomsHardware did an article yesterday. I am trying to find another source though.
I saw that thread. The issue is nobody can trust anything from Amazon, an item can have a dozen sellers but all get sourced out of the same bin, the same bin returns often are tossed into after a faked inspection by unhappy Amazon employees. So it could very well be true the guy bought a non-legit chip.
If the reddit OP was telling the truth then I donāt see Microcenter selling a non-legitimate processor though. So Intel seems in trouble from that angle.
1 Like
3 Likes
Not related to to what I posted, but still nice summation of intel responses and behavior regarding this saga to date:
@kougar
Yes, there is always possibility of co-mingled stock with outright lemons or fakes, but reported agent behavior, items description, and its generat nature conflict with those original claims.
Even if item fake, which is really hard to pull off without user noticing, agent threats of confiscation without reimbursement are massive red flag.
Unless laws differ massively, only thing intel legally can do is refuse service, and charge you for inspection and transportation cost when returning the item afterwards.
Anything else is between intel and retailers, unless they can somehow prove that customer bought fake knowingly and is trying to pass of as real to rma for personal gain, ie clean case of fraud.
I am reading that this either case of:
- incompetent agent
- agent executing checklist from higher up how to deny RMA coverage (and being very transparent due to mistakes) and finally using hail mary option when customer is not letting it go easily
Note - denying warranty with item is fake, without providing proof and also denying any insight how decision was made is ancient and reliable way to fuck customer.
Seeing it applied to something like modern CPU, where having fake is extremely hard and rare to pull off without being obvious (i.e you can do it realistically only by shipping entirely different item or do mislabeling of real intel but inferior cpu ).
If there were fake cpus circulating that were practically undetectable by end customers, that would be both major news and major intel fuckup.
I.e what I am hinting at - fake 14 gen intel cpu? Very idea is novel and laugable. You can fake some FTDI microcontroller, but not massive piece of silicon running on standardized plaftorm with every piece of firmware and microcode signed and encrypted.
But if it was case, then we have juicy news story developing.
3 Likes
As for Intel destroying fake chips, itās understandable why a company would do it. The optics just look extremely terrible when the same company is claiming legitimate products are fake, which is separate problem entirely. Intel was upfront warning him before he could send in the chips.
The issue of Intel claiming chips are fake when not, thatās a genuine problem. More and more threads are already popping up from claimed Intel customers reporting this. By all accounts it appears that Intel is doing exactly what ASUS did with its warranty service, as Steve calls out directly in his latest video. That intel would claim these are non-legitimate CPUs without physically having them in custody does seem pretty absurd.
Steve is all over the place in the video, but heās pretty quick to catch the changed statements and documentation. That Intel refused to even give its large customers the date range and batch numbers is pretty damning. I remembered the 21% pay cut Gelsinger took in 2021, but it really does just make the 43% pay raise he got for 2023 even worse. Nice to know Gelsinger was all talk and no substance.
As for Intel destroying fake chips, itās understandable why a company would do it.
Sure, but there is correct way to do this, and this isnāt it. Everything about described interaction is off. If there was actually fake chip in play, then there are two victims - the customer who was defrauded and intel, whose product and image integrity was damaged. And finally that fake chip is customers property, intel cannot just confiscate it as it wontās.
Look , I am commenting from european legal perspective, where withholding someones else property like this is actually very illegal move. Maybe its different in US.
From the communication errors, to overall approach and and finally the threatening statements. All they have in common, intel as direct beneficiary in the end.
It almost step by step what not do when interacting with customer, unless you absolutely intend to fuck them over from get go and are also absolutely certain nothing will blow back on you.
If thats so, something also tell me those support agents are outsourced to third party contractor and if this blows up, intelās hand will end up clean in the end anyway.
Intel was not withholding any property. The customer never sent in either physical CPU to Intel. He specifically declined to do so even though Intel finally gave him permission to RMA the second CPU.
As I said above the rest of the issue is a separate discussion and we donāt really disagree, Intel used the same underhanded tactics ASUS was doing with its own motherboard RMAs. Thereās no excuse for it. And it sounds like the usual minimum wage outsourced contract employee, even if the physical chips are mailed to a US address.
It seems they fucked up badly and it started way before Pat with those delays in the 14nm forever story arc. I hope intel doesnāt follow the light to meet with Sun, SGI, 3DFX, and friends.
1 Like
Non native speaker - I am not claiming the are withholding anything, they have threatened to withhold it if fake, and stated that they believe its fake without any explanation and justification.
Clear enough? Everything else stands with that in mind.
Anyone that comes to mind who can actually do confiscation like that is police or customs, and those processes give you proof you need to get your money back in full, if indirectly.
What asus does is also iffy, but over here you have leg to stand on if you pass your complaint to civil commerce authority.
It might take a few weeks, but if you documented what you sent to rma and communication between you and service center, the center gets nasty fine for trying to extort you.
I have thankfully never needed to go to such extreme measures, just threat of passing complaint to that authority was enough to unclog the process.
But I have never had to deal with pricks like asus.
Has anyone got their CPU replaced? I lodged a Intel support ticket, after a 10 days of back and forth questions they approved my RMA, but then havenāt responded in 2 weeks despite requests for updates for details on how the RMA works.
If anyone is still interested , intel finally published more comprehensive root cause analysis report.
Cause of instability itself was degradation of clock tree, and degradation was caused by multiple factors.
So it was both intel and oems fault :).
3 Likes