Thread 106003612

FUCK this fucking piece of FUCKING shit. Had to RMA the first one which took forever and now the replacement chip is failing exactly the same way the first one did. FUCK intel. Never going with their shitty CPU's ever again

"10% failure rate per year keeps the goyim in fear"
- Intelaviv

>>106003612 (OP)
I switched to AMD on the second gen's of Ryzen. Never looked back. Intel can suck it.

>>106003612 (OP)
>>106003961
How do their true failure rates compare?

>>106003612 (OP)
your motherboard is doing something stupid

It's actually amazing how bad intel's reputation is now. People that actually do buy them are getting burned hard, this is how you lose customers forever

the one thing they always had was stability. I've still got 4th and 10th gens in use

I heard 12th gen was good, but shit, at this rate AMD won't have competition again for a very long time

>>106003992
yeah it's trying to run an intel CPU
OP bro what you need to do is gimp your CPU so it runs like a 12100F then it won't cook itself oxidise itself and shit itself to death just use that cpu like it's 4c4t

>>106004001
I think Intel's latest laptop chips are actually quite good

>>106003612 (OP)
Did you update bios to the newest microcode?

antisemitism is not welcome here

my i7-10700k is still going strong

>>106004245
same. just make sure you are running the latest bios for the microcode fixes.
intel did fuck up, but if you have a recent one from after the microcode fixes, and run them on the microcode fixes, you quite literately won't have any issues.

>>106003612 (OP)
I had to RMA my 13700f. its been 3 months and still no replacement yet. so for now I've picked up a 12400f and it seems to not have any issues. This is the last time I buy intel and thats coming from someone who has used intel for the last 25 years. The way they've handled this is a joke.

>>106004341
this isn't true at all. I've had the microcode update from the day it came out and my cpu still died. I'm convinced the microcode update just prolongs the CPU from failing not actually fixing it.

>>106004341
yep
might still die earlier than a 10700K/10900K would but that applies to all modern CPUs, those transistors are fragile and tiny
>>106004404
your BIOS is misconfigured and you lack the agency to fix it yourself
i would recommend a macbook

>>106004433
no its not and I was using a f series CPU so you can't even overclock it. your just shilling because you also got tricked into buying intel like I did

>>106004445
>no its not
yes it is
>you can't even overclock it
yes you can: power limits, AC/DC LL, voltage offsets, etc
>your just shilling because you also got tricked into buying intel like I did
i have a 5800X

Works on my machine :)

>>106004433
10th gen is bullet proof after years of refinement on 14nm

>>106004460
How come my motherboard works fine with the 12400f?. Are 13th and 14th intel chips such dogshit you have to have to tinker tranny with setting just to get it to rune at stock settings?

>>106004475
indeed
i ran a 1700X at 1.4V all core on an air cooler for years and it was fine lol
you can't apply the same mentality to 10nm and below
>>106004480
i would stick to apple products from now on

>>106004464
No it doesn't, you butt hurt silly boy! :3

>>106003612 (OP)
I bought and used i9 13900k and I'm using it as my main PC, zero issues. I would suggest limiting the voltage request of your CPU to 1.4V. The issue is that high, transient spikes kill those CPUs and limit the voltage prevents any degradation.
I did that since I got mine and degradation is nearly impossible that way.
You might lose 100mhz but loosing that is nothing compared to killing your CPU

>>106004480
the problem are the 13th/14th gen I9s and the I7s K variants. Any other chip has a tiny change of degrading. Bullzoid did some testing and it seems like the issue is that Intel tries to predict workload spikes and ask more voltage than needed using a mechanism called "load lines". The higher power chips fail more as the single clock speed is higher, hence they request more voltage by default and a lot of mispredictions make the CPU request 1.6V hence frying them.
That's why lower power chips so not fail.
The video where Bullzoid explains Load Lines on detail, why they exist and why Intel is not using them on Core Ultra:
https://www.youtube.com/watch?v=9slwXKUwmnE&t=5657s

>>106004602
Ya fuck intel never again. I'm not rich so dropping $500 AUD for a CPU just to have it die and need to spend another $160 for a weaker replacement while I wait months and months for a replacement is not acceptable at all.

>>106004627
I agree with you, Intel really messed up this launch. Although Amd has also messed up too with Ryzen 5000 series as those are already degrading, but no big youtuber mentions it, for some reason.
The problem is that no one is testing for safety, these companies care about raw benchmark numbers and looking great on launch, then fix any mistake on the go with patches.

>>106004341
It could be that you are running a fake OS.

>>106004538
Can you walk me through that process? I'm tech illiterate.

>>106004001
2th gen was the last one good Intel CPU
t. went from i7 2600k to 7950x3D

>>106003612 (OP)
>mfw golden sample 14900k that never goes above 1.425v

>>106004445
there is still stuff you have to do. load line, current limits, dual tau boosting, etc.
i have pic related, i still had to go into my bios and manually set things. like limiting current to "intel's default" over "let the motherboard decide based on vrm quality." asus also wanted to do a fake overclock by enabling dual tau boosting which i had to disable. motherboards will still try to auto overclock your processor, regardless if its "f" series or not.

>>106005009
https://www.lmgt.org/?q=i9+13900k+voltage+limit+hardware+site%3Areddit.com

>>106004994
well if its because i am using a "fake operating system" (linux) then it goes to show, linux won't kill your processor like windows will.

>>106004341
>>106005150
>>106005156
I've also had this 14900K since September of 2024. After Intel released that big microcode update. Its only ran on the "fixed" microcode. Its been an absolute champ. I've abused the shit out of it. From 14+ hour long gaming sessions, to actually compiling software, video encoding, you name it.

Processors being used before the microcode fixes, yeah, are probably fucked. I am sorry. But for those who got them after the microcode, they should be fine.

>>106005138
what's your sp value? mine is 96.

>>106004707
>AMD has also messed up [...] with the Ryzen 5000 series as those are already degrading
Do you recall where you learned that?

>>106004181
> laptop, intel, good?
No. Equivalent ARM compute uses 25% the power of intel’s best possible achievements.

>>106005101
>2th gen

>>106006068
TechYesCity did a video touching this topic:
https://m.youtube.com/watch?v=KW1OZWH1tPg
He found a Ryzen 7 5800x for dirt cheap, it boots fine but the moment you apply any load it crashes. The fix? Reduce max turbo by 200mhz and it works now.
But he is not the first one with this issues. Just search for "Ryzen 7 5800x unstable" and you can search for a lot of cases of people complaining that their system crashes. The first link I found, from 2022:
https://www.overclock.net/threads/another-unstable-r7-5800x-even-at-stock-settings.1799521/
Again: no idea why Kiketubers quickly picked Intel degradation but stayed very quiet with this one.

>>106006161
>Ryzen 7 5800X
>no big YouTuber mentions it
>stayed very quiet with this one
If I recall right, I saw a post that said the 5800X and 5800X3D were quietly replaced with the 5700X and 5700X3D due to degradation.

My 12400F works fine bruv

>>106003992
>CPU requests fatal voltage
>motherboard complies as per spec
>i-its the motherboard's fault!
faggot

>>106005222
IDK of SP value specifically because I think ASRock uses a different scoring
It's up there tho

>>106006793
The CPU requests voltages based on a few assumptions that have to come from the motherboard and they have to match the actual electrical characteristics of the board like AC/DC load line
Board makers have been playing fast and loose for years on this like reporting the mobo has a much more lossy vcore plane than it really does.
CPU requests voltages expecting a certain vdroop out the VRM into the CPU, gets higher voltages than expected because there actually isn't much droop across the vcore plane.

Hasn't been a problem because CPUs could take some additional voltage but 13900ks and 14900ks cannot.
It's not strictly the mobo makers fault but years of bad practices has exacerbated the issue.

>>106006246
5800X is still available and 5800X3D was discontinued because there weren't enough good CCD bins left
>>106008469
this is /g/, people don't understand how computers work here
you plug it in and it does things on its own beyond your control

>>106008469
>Board makers have been playing fast and loose for years on this like reporting the mobo has a much more lossy Vcore plane than it really does
Do you know how long board makers have been doing that? What does lossy mean?

>>106008469
>Board makers have been playing fast and loose for years on this
So why doesn't it affect AMD cpus?

>>106003612 (OP)

My 14900K failed last year and it took months for them to refund me. The replacement has been good so far though and I do have the microcode updates installed. My next PC will probably have an Nvidia CPU in it.

>mfw still running ancient microcode 0x11D that gives you +100Mhz throughout on my 13900K since launch
lmaoing at you negro faggots
I just fixed my AC/DC LL and have been running this thing at unthrottled and uncapped power and current limits since I got it in early 2023. Ran everything under the sun known to cause degradation and killing but its been running super stable.

>>106010894
>I just fixed my AC/DC LL
How?

>>106008469
I will cite Bullzoid here: the problem is not the motherboard makers, is Intel. The CPU request a voltage to the motherboard, it gets that voltage. Motherboards can only influence through Load Lines how much voltage to apply as compensation which adds/removes a small delta on the CPU request. Heck, motherboard makes actually saved Intel by using a smaller Load Line resistance by default (for I9s Intel default is 1.1mohns, they were using 0.4-0.6mohns which reduce the Vcore spikes on single core)
>>106010798
You are kinda correct, although AMD also messes up they chips. Let's not forget that the 9800x3ds are cooking themselves. Although more frequent in ASRock motherboards, AMD issued a fix on AGESA 1.2.0.3e for all vendors to appply. Here is a link for MSI mb in Leddit.
https://www.reddit.com/r/MSI_Gaming/comments/1lprkhz/agesa_1203e_the_most_important_bios_update_that/
Ask yourself this question: if ASRock was the sole problem (and they have some blame here) why did AMD had to issue a CPU patch? It was a CPU issue! As, again: they have no incentive to overvolt the CPUs. If the CPU request X voltage, it gets X voltage.

>>106010748
>Do you know how long board makers have been doing that?
Not the original poster but from my own experimentation, they've been doing it since at least the launch of 9900K.

>What does lossy mean?
You have some voltage drop across motherboard VRMs (basically Ohm's law and the set VRM switching efficiency). So for example, if CPU requested 1.2v and there is a drop of 0.05v across VRMs, the CPU will actually get 1.15v delivered which could make it crash. To counter this, motherboard sets some values in the CPU called AC/DC LL. Motherboard tells the CPU how much of a voltage drop is expected on the VRM plane. The CPU then adjusts the requested volatge accordingly to get its desired voltage. So in the previous example, CPU will now actually request 1.25v to get the requested 1.2v. If you incorrectly set the AC/DC LL values, you can either undervolt or overvolt the CPU.

Motherboard vendors have been undervolting the Raptor Lake CPUs since launch as it lowers the temps and causes them to boost for longer, making their motherboard look good in reviews. After some CPUs started crashing because not all of them can maintain undervolt, motherboard vendors released some BIOS, due to panic and incompetency, that OVERVOLTED the CPUs causing them to degrade and kill them.

>>106010748
>Do you know how long board makers have been doing that?
Not sure on specifically they started deviating from guidelines, it just hasn't been an issue until now.
The VID system has been in place for well over a decade. Wouldn't worry about anything pre-10nm

>>106010798
The fast and loose attitude is how you got 7800X3Ds melting or the current crop of 9800X3Ds degrading.
The problems are unrelated it seems to AMD's SVI system and more board vendors just allowing the CPU to request extremely high voltages when it comes to 9800X3Ds, Its assumed these voltages are technically within AMD guidelines but right up against what's permissable and some CPUs seem to can't handle it. Asrock in particular had a higher limit than others.
On the case of 7800X3Ds it was board vendors purposely setting high VSOC voltages in an attempt to increase memory compatibility. Asus in particular set basically no limits for voltage or current and fed CPUs until they literally melted.

>>106003612 (OP)
>he boughtered Intelaviv

>>106011017
It depends on the motherboard. Search the web for your specific board, all data is available. If you have Asus board. then I can help you. Or, just install the latest microcode fixes.

>>106003612 (OP)
bluesisters... not like this

>>106011055
>I will cite Bullzoid here: the problem is not the motherboard makers, is Intel. The
The only mistake Intel made was not clipping the VID requests which they fixed in the newer microcode. 95% of the blame is on the motherboard vendors.

>>106011119
>I just fixed my AC/DC LL and have been running this thing at unthrottled and uncapped power and current limits since I got it in early 2023. Ran everything under the sun known to cause degradation and killing, but it's been running super stable
To get what you have, someone would have to get a 13900K (or any other CPU with the same issue), not apply the microcode update (or revert to an earlier BIOS that doesn't have it), and manually fix the AC/DC LL? Would that work if someone were to buy one of the newest 13900Ks (or others) around as far as what was placed on a brick and mortar store's shelf or stored in an e-tailer's warehouse?

>>106003612 (OP)
>buy from one of the two known problem gens
>complain
You have only yourself to blame. If you didn't win the silicon lottery and get a good 13th gen the first time like I did, you should have just gone for 12th gen (same mobo support) or their post-rebrand gens.

>>106011094
>9800X3Ds degrading
It's asrock mbs killing them.

t. my asrock MB killed my 9800x3d

>>106006083
>meme ARMshit

>>106003612 (OP)
works on my machine, great cpu

>>106004404
The microcode only stops accelerated degradation, it doesnt undo any existing damage,

those chips were never going to last long even without the bugs, because they are pushed to the edge out of the box , 10 year lifespan im guessing. so if you have a partially damaged degraded chip, even with the microcode fixes , it probably has a dramatically shortened lifespan.

>>106011378
Yes. It would work on any 13900K/14900K if you have a board with an older microcode/BIOS. You get +100Mhz extra (5.5Ghz all core instead of 5.4Ghz in case of 13900K) and it can hold two-core max boosts (5.8GHz for 13900K) for days. You just have to manually fix some AC/DC loadline settings. After the microcode fixes, it got severely circumcised.

>>106011094
its also the microcode, its bugged , early microcode allows the highest boost state at 90c instead of the defined 70c, and it also for whatever reason causes huge vid spikes at the beginning and end of a workload, like 1.65-1.72v , this alone is likely the main culprit, since alot of boards default to infinity amps allowed, that means regular spikes of >1.65v at potentially very high current at 90c, which will 100% damage the metal oxide layers of the cpu, even if its only for a brief periods

honestly if the temp limited werent screwed up, Its very possible it would have take another few years for this issue to show up, since temp accelerates degradation by roughly 2x for every 10 c increase.

>>106011731
>vid spikes at the beginning and end of a workload, like 1.65-1.72v
That only happens on Gigabyte boards AFAIK because they set AC/DC LL even beyond Intel specified range during the fix. Even then, some boards, like Asus, do have the option to limit the voltage output manually. Just limit it to 1.45-1.55, or whatever you're comfortable with, there.

>>106011793
happens on basically all motherboards, but you need an oscilloscope to see it, it was just particularly bad on gigabyte so you can actually see it happening with software monitoring.

>>106011793
I have a gigabyte laptop and it freezes completely on random. Forcing me to hard restart it. Could it be due to this?
Also is it possible at all to check your cpu’s health?

>>106012015
you can try running a light load that triggers the 1t max boost state frequently , the core cycler script might be useful for this. https://github.com/sp00n/CoreCycler

>>106011724
>13900K
Do you recall why you chose that one?

>>106012516
It was the best available at that time (March 2023). I needed the best single core speed for a specific emulator and best multicore performance for large software compiles.

>>106012887
>a specific emulator
For the PlayStation? If so, which generation?

>>106003612 (OP)
works on my puta

>>106003612 (OP)
LMAO, gen13 intelaviv goyslop strikes again.

>>106013060
N64 with software rendering

>>106003612 (OP)
No such cases with the glorious i5-2500K.
This is what you deserve for thinking you need more.

>>106013766
>N64
Were you trying to get a particular game to run optimally? Or were you going for better performance across the board?
>software rendering
Why?

>>106006085
1th gen, 2th gen, 3th gen, 4rd, what's it matter?
Anything with more than 4 cores BAD.

Imagine accepting a task manager that looks like this
>>106005138
LMAO
They're not even real cores. Only eight are real. You're getting 8 fake cores (threads) and another 16 fake estrogen cores. That's 4x the CPU charts that should be displayed.
E-cores should be grouped up and hyperthreading is a meme that should be turned off anyway.

>release 8 core CPU
>spams 32 CPU graphs in task manager
They go to great lengths to mask CPU inadequacy as this translates to a maxed out CPU struggling on a single-core workload only registering as ~3% CPU in benchmarking tools even though it's running at its hardest.

S C A M

>>106013967
I was making an experimental N64 emulator with software renderer.

>>106014148
>an experimental N64 emulator with software renderer
CPU emulation only? No GPU emulation? The GPU is used only for outputting the game? Do you have a GitHub (or similar) of the emulator?

>>106014201
Both CPU and GPU (RDP). Yes, it did boot and play some games.

>Do you have a GitHub (or similar) of the emulator?
No. It was a personal fun project.

>>106013986
>muh e cores
Post system

My 22nm, 662mm^2 18-core E5-2696v3 will be crunching for millennia compared to these. What peasant uses such tiny transistors? Got a dual-socket Broadwell EP rig with dual E5-2673v4 with 20 cores each in the works, with 128GB ram in octa-channel. X99 will run my personal AI for the next 10 years lol. Chinese are genius for making new board to run the old server stuff.

I bought a i7-13700k for my sister's PC because I have it and it's working fine for 2 years
And I bought it from a private seller
But it has been failing since the beginning, I upgraded the Intel drivers which fixed it for a short while but it's started failing again this month
and in the EU there's this law that I can return the product within 180 days even to a private seller for a refund
and it's been 178 days when she told me it's failing again so I filed the thing and now trying to get the cash back, the guy who sold it to me will be able to RMA it if he wants but as a consumer I was supposed to get a working chip from him
kinda feel bad for the guy but it's in my favor so I will take it if I can, and ultimately it's Intel's fault

but anyway when I get a refund I will buy i9-12900k for my sister instead. it's a little worse which she won't notice, but stable

>>106014672
>E5-2696v3
How's the power usage?
>a dual-socket Broadwell EP rig with dual E5-2673v4 [...] in the works
Is it up and running? If so, same as above.

>>106014914
No big deal at idle. 145W TDP but it runs very cool for what it is, and single core turbo is 3.8GHz. Tried turbo unlock bios but the Huananzhi QD4 couldn’t handle 3.2GHz on all 18 cores. The big draw is the R9 295X2 with the 500W TDP. Had to 3D printer a 5.25” bay to 120mm fan intake duct with a Corsair ML120 Pro that spins up to 2400RPM to keep it cool. Just posted all my rigs in the battlestations thread. Dual socket rig isn’t done yet, been too busy but it’s on an EATX Huananzhi F8D-Plus with a matching orange and black Gigabyte G1 Gaming GTX 1080.

>>106015495
>>106014914
>>106014672
Pic related.

>>106003612 (OP)
>buy a used 13700k
>immediately undervolt it by 100mV
>perfectly stable with no issues for >2 years now
>runs much cooler, which also allows it to clock higher
The stock voltage curves manufacturers are trying to push so they can offload shit silicon is insane, I also undervolted my 3080 by like 150mV and saw massive improvements in cooling and clockrates.

>>106015511
Does vertical do much for thermals? My 5070ti hit 52c gpu 56c junction temp yesterday, I was shitting bricks

>>106015678
You should have bought palit gamerock 3080. Best coolers on the market

>>106008469
>Board makers have been playing fast and loose for years on this like reporting the mobo has a much more lossy vcore plane than it really does.
That results in an undervolt though. It would be stable just like all the older i9 9900ks to 12900ks if the 13900k/14900k didnt have degradation problems.

>>106015852
All rtx3080/90 air coolers save for the 3090ti are underbuilt

>>106010894
I ran my 13900KS undervolted for the year or so I had the CPU and by the time I sold it, it was starting to show signs of instability from crashing in remnant 2 which is a UE5 game

>>106015838
No idea, it’s my first build trying it out, always loved inverted ATX because it showcases the interesting side of the GPU, same idea with vertical, just wanted to try it for fun and the PCI-E 3.0 riders are cheap. First modern case I’ve had that supports it, it’s a Sama 3509 that is the smallest case I could find for EATX and it was on sale for $75 CAD delivered from newegg. Not much risk with a 1080 that uses less than 200W, but newer stuff could be hard just because of the width and 3.5-4 slot thickness of newer stuff.

>>106015953
It results in an overshoot of voltage actually
If a board reports to the CPU that the power plane resistance is something like 1.1mohm but the ACTUAL plane is 0.540ohm, you get an overshoot because the CPU is requesting a voltage from the VRM expecting a loss from the actual electrical resistance of the vcore plane that actually isn't there.

If you did the opposite like set a 0.540mohm LLC on a crappy board that has a 1.1mohm vcore plane then you would get a lower voltage to the CPU then what it expects.

On my Asrock taichi board specifically, they had two separately released options going into the 13th, 14th gen fixes.
"Baseline" set LLC to 1.1mohm, this resulted in 1.45v VIDs with something like 1.42v on the external Vcore sensor.
"BIOS default" set 0.540mohm, this resulted in ~1.425v VIDs and the core sensor being ~1.4v which is exactly the voltage the CPU wants at 6ghz

>>106003612 (OP)
i7-13850HX Chad inferencing on Qubes OS here, just werks for me... must be a skill issue on your end, /b/rotato chip!

>>106004001
after they stagnated CPU progress for an entire decade they don't have any goodwill to burn away with the Alzheimer Lake fiasco.

>>106015511
these chink boards are a godsend, fuck ewaste and fuck planned obsolescence jews. AM4 tray CPUs are another fire option.

>>106016159
I'm talking about the initial behaviour of motherboards prior to intel acknowledging any problems and issuing any fixes, so pre-2024 lga1700 behaviour.
All the Asus and MSI boards I've used always underreport AC loadline at stock (~0.80mOhm) while running vrm loadline at intel spec 1.10mOhm so they're undervolting at stock
Same for the older 9900k (1.2mOhm/1.60mOhm) and 10900k (0.80mOhm/1.10mOhm) too and they've all been stable without problems

>>106014672
Yeah 14nm is the final process node with reliable transistors. Anything smaller is pretty questionable

>>106015997
>>106010894
it wasn't just high voltage and it wasn't just buggy power management sending the wrong voltage.... but voltage shift issues as well so that ended up frying the clock tree as well even if the cpu sat at idle drawing 700mv of power. i won't go into heavy detail about it but it was just all retarded shit. intel really fucked up so much basic shit. the entire voltage handling, its entire power management was all buggy and wrong. they literately did no testing before release.

the clock tree issue, would have never been an issue, if intel actually tested their all clock trees are sensitive to heat but the issue intel had, was they actually, admitted, they didn't bother doing any quality assurance / testing to figure out its limitations and set everything up appropriately.

>>106016526
> if intel actually tested their all clock trees are sensitive to heat
if intel actually tested their shit* all clock trees are sensitive to heat, but the issue intel had, was they actually, admitted, they didn't bother doing any quality assurance / testing to figure out its limitations and set everything up appropriately.*

so you had an issue of:
>transient spikes going higher than they thought
>motherboard vendors auto overclocking something that was already pushed to its limit, beyond that limit
>power management that was sending the wrong voltages, and was overshooting because of it
>voltage shift issues that was slowly causing electromigration issues, causing the cpu to need MORE voltage, even at idle, to stay stable.

>>106016559
now all of those issues are fixed (if you trust intels word) but all of that shit should have been caught during testing.... the fact they admitted they didn't do any long term reliability testing nor any testing to see how the clock tree behaved, even though they did a complete redesign of it with raptor lake over alder lake, is just embarrassing and criminal. i don't know why no class action lawsuit has been fired over their admitted incompetence.

>>106016559
>voltage shift issues that was slowly causing electromigration issues, causing the cpu to need MORE voltage, even at idle, to stay stable.
and this one should get more attention than anything else, because this issue resulted in, cpu's frying themselves, at idle, doing absolutely nothing because the default voltage to remain stable at idle, kept increasing higher as time went on due to electromigration issues developing around THE FUCKING CLOCK TREE.

idle voltage going from ~600-700mv at idle to 1-1.1 volts. now, those voltage might sound low and safe, but, for idle, lower frequency, that was TO HIGH. causing to much heat building up around the clock tree, and FRYING THE CLOCK TREE. since the transistors / registers were not opening and closing fast, helping to dissipate the heat, the heat would build up as a hotspot.

>>106016601
its just so fucking retarded how intel didn't catch any of this shit. all basic shit. no excuse on their behalf. they honestly should give refunds back to every single purchaser of a 13 / 14th gen processor. if not please someone file a class action lawsuit.

>>106003612 (OP)
>one day later op is still a faggot

>>106016526
Yeah the voltage regulation stuff they blamed motherboard makers for has behaved that way since 10th gen but was never an issue since those CPUs were properly QC'd and didn't degrade from 1.55-1.70v requests

>>106016951
>isac the mentally ill turbo autist calling someone else a faggot
wew

>>106017444
The max VID for Intel has always been 1.52V (since 14nm). You can look at the datasheets for sixth gen onwards for that. On 9th gen, they introduced 200mV "offset" for some reason but noted that it needed some special support. This is where 1.72V comes from (1.52V + 200mV). I don't understand why they did that. The VID would always be <= 1.52V though (this was bugged).
Anyways, the reason those CPUs didn't degrade was simply because they couldn't clock that high. Therefor, the VF curve on them is mostly < 1.4V usually. Even 12th gen, which is exactly the same CPU is safe simply because it can't clock that high. Raptor Lake on the other hand pushes 5.8Ghz-6.2Ghz at stock. This required higher VIDs (most sample have VF curve at 1.55v) hence the issue showed up for them.
If Skylake 14nm(+++) series could be clocked to 6Ghz (at >=1.5v), the issue would show up for them as well.

>>106017756
>xbox autist jargon
Go back to playing runescape
>>106016559
On MSI Z790 motherboard it was actually downclocking the 100mhz clock, after degradation the fucker couldn't boot cinebench if the downclock was disabled.

>>106017952
1.52v-1.72v vids worked perfectly on the couple of Asus boards I owned.
And you underestimate how desperate intel was. Starting with 10th gen, shit binned i9-10900k's had vids as high as 1.558v for 5.3ghz while later batches of 11900k's went up to as high as 1.65v for the 5.3ghz 1T boost. With the 12900k, it was relaxed slightly down to 1.4v ish and later went up to 1.5v again with the 12900KS

>>106018061
>picrel
ignore the 5700mhz on the 11900k since the owner was OCing. 1.653v is the stock voltage of the 5300mhz point of his cpu

>>106018167
wtf lol that is wild but that is worst case scenario and I doubt it would've lasted long if ran under 1t load at TVB temps
Anyways, I never tested with Comet Lake so can't comment there but 11th gen had highest failure rates as per Puget Systems.
I also have 11900KF on my secondary system and its 5.3Ghz frequency is 1.35V at stock but I've undervolted it by 25mV and its working perfectly stable since over a year now.

>>106018550
I'm surprised 10th gen had that low of a failure rate since they are notorious for internal parity WHEAs in minecraft java, AAA slop with RT/PT enabled and presumably any workload that has a lot of memory accesses

>>106018585
random intermittent crashes are always blamed on the developer

>>106013986
>Anything with more than 4 cores BAD
Why? You can go up to 10 cores without destabilization.

>>106018929
Because he's retarded and has no clue what he's talking about.

>>106015838
It makes it shit. Most extender cables are only pcie gen 3, when everything is made for gen 4 or 5 anymore, and the entire point of using an extender to vert-mount a gpu is pure aestheticslop.

>>106015678
>immediately undervolt it by 100mV
>>106015997
>I ran my 13900KS undervolted
>>106018550
>I've undervolted it by 25mV
Have you noticed issues while undervolting a CPU? 2 posts here said CPU undervolting should be avoided. And if one wants to lower power usage, they should use C-states instead. Something along the lines of CPU undervolting can cause errors. If I recall right, one example being copying files (data corruption). And it could take a long time (perhaps years) before errors become obvious. Even if one were to run stability tests, and, in their mind, verify that the undervolt is stable. If I'm phrasing it right, stability doesn't necessarily mean integrity.

>>106021336
Tard

My 13900k has been running 24/7 for 2 years with the latest bios patches.
Intel gave an extended warranty, but no one should trust these chips for anything. They are flawed by design and you get the added "opps sorry we shipped rusting batches to save money and didn't think anyone would find out!"
I hope Intel dies a well-deserved death.

>>106021336
>Have you noticed issues while undervolting a CPU?
Never. I am running all my CPUs (and GPUs too) undervolted.

>And if one wants to lower power usage, they should use C-states instead.
C-States are something else entirely and should be used together with undervolting.

>Something along the lines of CPU undervolting can cause errors.
Impossible, if you've tested for stability. Its more possible that your RAM XMP/OC might be the culprit as it take way more longer and is much more involved to check it for stability.

>>106018585
i never had issues with my 10850k. no whea errors, nothing. it ran like a champ.
>a lot of memory accesses
that sounds more like unstable ram than actually being a cpu issue.

>>106013986
>2th gen
>3th gen
>4rd gen
My brother in Christ... it's 2nd, 3rd and 4th. Go back to posting on Facebook Pajeet

>>106013986
>1th gen
>2th gen
>3th gen
>4rd gen
Pojeet my brother... it's 1st, 2nd, 3rd and 4th. Go back to posting on Facebook

>>106022037
My 10700K had WHEA errors on minecraft
I also ran 3200Mhz CL22 (JEDEC spec) DIMMs

It's in my NAS now, and as stable as anything. Hard to say if it's really an issue

>>106003612 (OP)
>Never going with their shitty CPU's ever again

For me it's total opposite. After I boughted AMD Ryzen 2600 and had to deal with this street shitted CPU, literally wasted time of my life on the CPU, literally tried to make it work, I will never EVER EVER buy AMD ever again. My i5 worked so long until I got shilled into buying AMD. Fuck my old i5 would still work and perform better than the piece of shit Ryzen trash.