CrackongIn Intel spec document it is described that MB vendor should measure and set their own values. (Since no such value is provided by Intel) [...]

AnonymousGuy767My 14900K wasn't stable at 6Ghz turbo boost. Didn't touch any frequencies or voltages manually, and it would crash at idle when it happened to clock up to 6Ghz because of some background process.

That was a fun one trying to troubleshoot. Eventually I locked it to 5.7 or 5.5 max regardless of core active count and called it a day, and it didn't crash after that. And then it degraded and had to get RMA'd anyways.

So I basically said "f this noise" and put a non-K in. No more TVB to go to unrealistic clocks or weird all core frequency stuff that is beyond specs. And seems to idle at a sane voltage so probably won't degrade. I have the replacement 14900K here so I'm tempted to unseal it and try the new Asus bios to see if they fixed that, but I've done so many CPU remounts and swaps and stuff fixing problems that I'm kinda over it.

Solid State BrainAgain, motherboard manufacturers seemingly tend to use use more-or-less random (or at best, one-fits-all) values for this, and end-users pay the price for it.

AnonymousGuy767My 14900K wasn't stable at 6Ghz turbo boost. Didn't touch any frequencies or voltages manually, and it would crash at idle when it happened to clock up to 6Ghz because of some background process.

That was a fun one trying to troubleshoot. Eventually I locked it to 5.7 or 5.5 max regardless of core active count and called it a day, and it didn't crash after that. And then it degraded and had to get RMA'd anyways.

So I basically said "f this noise" and put a non-K in. No more TVB to go to unrealistic clocks or weird all core frequency stuff that is beyond specs. And seems to idle at a sane voltage so probably won't degrade. I have the replacement 14900K here so I'm tempted to unseal it and try the new Asus bios to see if they fixed that, but I've done so many CPU remounts and swaps and stuff fixing problems that I'm kinda over it.

Dr. DroInteresting, you must have gotten one heck of a dud sample. I've always expected a few 14900K samples to be bad, considered it's really just slapping 13900KS clocks onto a 13900K chip and calling it a day. It relies too much on its binning grade, and we all know not every processor that rolls off the line will meet the same standard of quality. I think installing this new 14900K would work, but perhaps not worth your time if you got a non-K 14900 installed already. Kind of why I didn't bother buying a 14900KS either, I don't think it's worth the effort to seek out and spend the resources to purchase one.

Thankfully I didn't share in your experience, things have been great with my 13900KS.

Solaris17This is wild. They literally post these specs on the intel ark pages. lol.

Crackong



Since there is no reference value provided by Intel.
MB vendors had to cook up their own values, by testing the ES CPUs provided by Intel.
IDK how many they've tested.
But juding from the reality, it doesn't cover the whole silicon lottery spectrum.

dgianstefaniWould not be surprised to see them coming out swinging this year.

ZubasaYou know, maybe it is because Intel pushed 14th gen which are the exact same silicon even further than 13th gen, thus all the problems.

:D:DDoesn't Intel provide POR?

:D:DES CPU's? The VRTT is a tool that sits in the socket instead of a CPU. Presumably it tests and analyses the mainboard voltage regulator, board traces / power planes and components to get specific values that are dependent on the manufacturers board, so no, Intel cannot provide those details as it's pertinent to the mainboard being used. From these values presumably custom values could be used by the mainboard manufacturer via BIOS to improve performance.

:D:DWould Intel locking down TDP stop people running beyond those limits? No, on it's own there are ways to circumvent.

If Intel allow changes that cause problems then perhaps it's partly their problem as well as manufacturers and even users who decide to take components and build their own system rather than buying a prebuilt PC with warranty. JM2c

Just for the record, W1zzard deliberately disables the thermal and current protection for these tests

Carlyle2020hsI don´t like TVB since it always tries to shoot fot the moon.

Not that i trust the numbers in this pic, but the burst captured does produce a heat spike, so something is overdriven:

That´s how TVB looks like for me with a 14600k

Carlyle2020hsYeah!

Got me pretty high scores though, that i can´t take real credit for:

www.3dmark.com/cpu/1700909

I played around with setting tvb targets (+2 +3 +4), like 4 months ago and stopped since i tuned a silent system and those burst are annoying due to the fans ramping up.

So sorry, no real conclusion here except parts of tvb work even with a 14600k.
Setting 5.7Ghz manually gets me close but without those heat spikes at the beginning of the tests.

dgianstefaniAmazing, thanks :laugh:


What?

chrcolukASRock on a Z790 board set TJMAX to 115C as default (this also happened on my board).

Vya DomusIt's a meme to you maybe, I haven't had any significant issues and I jumped on 7000 series pretty early on.


AMD also has the benefit of typically not lying and that helps a lot, so even if something goes wrong they don't get as much flack for it. Maybe Intel shouldn't have that said in the past that it's actually totally within spec to have a gazillion watt power limit on their CPU, they should have kept their mouths shut and now the narrative that it was the motherboard maker's fault would have been more believable.

Unfortunately they didn't kept their mouths shut and now if you have a brain it's hard to believe they weren't at fault for this.

chrcolukSee link I posted in my previous reply, these claims of them saying its in spec are largely from media. Their tech support isnt telling customers it in spec.

Media reps I wouldnt consider as gospel either, their job is to sell stuff, what support staff and documents says has more meaning. The documents posted in this thread dont state unlimited power as in spec.

Intel are guilty of not policing things properly, but its all gone wild when people are claiming what the board vendors have done is considered spec by Intel.

:D:DWell really I would put that one on Intel for allowing it. Started doing that on some chips from 4th/5th gen IIRC

ZubasaTBH customer / tech support can be hit and miss depending who you get as well.
GN did a video a couple years back about Intel CS regarding things like XMP. One CS says its OC and thus not covered by warranty they other didn't even bother asking about XMP being enabled or not.
Intel and AMD has many EULA / Disclamier clause that are usually not enforced but are there so they can weasel out of legal trouble.

chrcolukThis just isnt rationale thinking. Board vendor writes the BIOS and provides the board, they even told the customer they did it for performance.

BoggledBeagle80% of 14900K unstable with auto profile???

wccftech.com/only-5-out-of-10-core-i9-13900k-2-out-of-10-core-i9-14900k-cpus-stable-in-auto-profile-intel-board-partners-stability-issues/

I often buy processors because I have a big studio with dozens of machines, and I occasionally install them and sell them. So starting from the 13th generation, 13900K started to be used, and 14900K was released to replace 14900K in batches. So far, at least more than 100 14900K have been passed through. Step by step from 13900K to 14900K, it’s hard to describe in one sentence. In one sentence, it can be summed up as follows: It is already enough and I need to pull more. In addition, I have bought a lot of both boxed and loose pieces. By the way, I can tell you my feelings about the boxed and loose pieces. Mainly the main motherboards are Z690 B660 and Z790 B760, used from 6 series to 7 series. Not to mention the BIOS. I have to go to the ASUS official website once a week to see if the BIOS update can solve the problem. Before I begin to summarize, let me first talk about my standards. [Fully defaulted and safe to use] The productivity software runs continuously for a week without freezing, crashing, restarting, or reporting errors. (Productivity software refers to various data calculation categories, including but not limited to matlab, R, finite element, VC, VS, large-scale office batch processing, etc.) Yes, it is this productivity stability requirement that PASS has lost many Us. First, set the Z board to full AUTO and default to no overclocking. The power consumption is 253W. If it fails, go to the B board. If it still fails, it will be determined to be unstable. Let’s summarize: 1. The probability that 13900K can be fully defaulted and used safely is about 40% to 50% (at most 4 to 5 out of 10 can be fully defaulted and used safely). Improving the anti-drop pressure can increase the probability by about 10% to 20%. Setting the voltage to worst on the Z board or going to the B board can increase the probability by 20%. The probability that 13900KS can be used safely by default is between 13900K and 14900K. 2. The probability that 14900K can be fully defaulted and used safely is about 20% (at most 2 out of 10 can be fully defaulted and used safely). Improving the anti-drop pressure can increase the probability by less than 10%. One special thing about 14900K is that it is basically unsafe to use on the Z board, even if the failsafe voltage mode is set. So I mainly use B boards. 3. The power consumption and temperature of loose chips under the same load are generally about 10~20% higher than that of boxed chips, so it is easier to trigger the power wall or temperature wall. 4. The stability of loose chips is slightly better than that of boxed ones (this conclusion is only responsible for brand new loose chips. Second-hand loose chips are most likely to have been picked up even if there are no indentations). The slightly higher voltage of loose chips may be a life-saving straw. But what's the use of easily decelerating at 100 degrees (the most outrageous 13900K chip I've ever seen has a 180W power consumption that reaches 100 degrees of thermal deceleration). 13900K loose chips are easier to stabilize, while 14900K loose chips are useless and completely slag. 5. Intel recently added baseline settings. I tested it immediately, and the result was: completely useless. It is just equivalent to helping you limit the failsafe voltage (Z board), default CEP, current wall, and power consumption wall. Doesn't help improve stability. After Z board adopts this setting, it is equivalent to becoming B board. 6. Before the baseline option was available, the stability of board B was better than that of board Z. However, the stability of Z board after adopting the baseline option is no different from that of B board. 7. I have only bought 3 pieces of 14900KS so far. It is too IQ taxing, so I will not comment on 14900KS for the time being. I will make a separate post later. 8. In particular, once I accidentally set the power consumption wall to 33 watts and found that the energy consumption ratio was satisfactory. It is strongly recommended that you simulate the low power consumption situation, there will be surprises. 9. Regarding the voltage, I have tried Intel failsafe on the Z board before. If you choose this voltage mode, the motherboard will indirectly become the B board. 10. Supplement: I have bought dozens of 13700K before, and all of them are as stable as old dogs, and there is no problem with any of them. Therefore, I personally think that if you want to close your eyes and use it with peace of mind, the 13700K is the most stable. 14700K is not as cost-effective as 13900K, so I have never been exposed to 14700K and cannot comment on 14700K. Conclusion analysis: The above is just my personal experience as a productivity user, which is equivalent to the conclusion that I spent money to try for everyone. Pressurizing is definitely the right answer, but how to increase it to a stable point, especially when long-term testing is required for productivity, is too energy-consuming. Sometimes the test may be fine for 2 days, but if there is a problem on the third day, it is very time-consuming. . So every time I encounter something that fails to test, I resell it directly. If you are just an average user, there is no need to care about my extreme productivity conclusion. It can be relaxed a lot. After all, ordinary users can still tolerate unstable situations that are not easy to relapse. For example, my own personal computer, the 14900K+Z version, by default only encounters a crash and restart [doge] once a month at most.

BoggledBeaglekmdkai on the chinese forum wrote, that the unstable chip mail fail after a few days of his testing and that his findings are not very relevant for a "casual user".