Something I forgot to mention in the video. I have read reports from users on various forums who said they were getting all kinds of instability issues with their 13th and 14th gen Intel CPUs. However once they had loosened up their cooler mount a bit or switched to a thermalright contact frame their problems went away. When I was tinkering around with DDR5, this was something I also read from other users that too much MOUNTING PRESSURE was the cause of instability and they would get memtest errors if it was too tight. Haven't tested it myself for obvious reasons, but its an interesting theory.... In any case Intel denying RMAs to users is some load of crap!
It might just be related and not the cause. Reason is if this issue is pads ripping off due to thermal stress, or substrate failure, pressure changes will just reset that pad and prolong the issue to occur later. From what I can see it is substrate failure e.g. the CPU itself degrading, due to it being voltage bumps to regain stability. If a pad/mounting pressure is causing that, it's not going to degrade so linearly. Even someone commenting who knows an intel engineer said they said the hardware itself was 'effed'.
i had a theory that the instability issues may be a mix of temperature, voltage and mounting preassure. just think about it, we know for a fact that the mounting bracket of the Mobo may deform the cpu, Gamer nexus made a video about it and proved that this caused issues with temperatures, well , what i was thinking is that over time the high temperatures the voltages and the preasure would end up bending/deforming the silicon of the cpu causing physical damages to the cpu die which would be the cause of the instability or even kill the whole thing/ i could (and most likely) be wrong but until Intel or someone else find the root cause of the problem anything could be use to explain the problem, i think it would be good to know if people who are using contact frames from the moment the build the pc are also having the same instability issues.
Looks like Intel flew too close to the sun trying to beat AMD in the benchmarks. Customer loyalty and reputation be damned. All about those shareholders.
It’s time to reevaluate and think critically as us consumers. Being the “fastest on the market” doesn’t really matter anymore. Stability matters. Price to performance matters. Efficiency matters. Speed should come naturally, not through brute-force.
They didn't have to force it anyways, my i7-12700K at stock already beats in 1% lows the Ryzen 7 7700X in Hogsmead area for Hogwarts Legacy, only a 7800X3D can beat my chip. I'm cooling my 12700K with an overkill Deepcool Assassin 3, power consumption while gaming is between 30 and 70 watts, just like a 7800X3D.
Intel basically did this- > Hastily came up with a CPU architecture in less than an year to beat Zen 4 > Used 50-100% more power to beat X3D chips > Found out that their chips are crashing > Not sure what caused them, but shifted the blame onto motherboard vendors anyway > Found out its a fundamental architectural flaw > Trying to ignore it and handling communications poorly Im curious to see what they will do next lmao
But The IPC on 12th-14th gen is higher than Zen 4. Definetly it's an optimization issue and quality control... AMD said that 95 degrees on Cinebech for the Ryzen 7 7700X is fine, meanwhile my 12700K with 4 extra cores doesn't go past 76 degrees and clocks 800MHz lower. My 12700K is faster in Hogwarts Legacy and Dragon's Dogma. "beat X3D chips" It's only the 7800X3D. The 7900X3D and 7950X3D sucked (scheduling a way bigger mess than an i9), no one talks about the latter two. Hardware Unboxed tried the APO feature for the 14900K, power consumption was reduced by 50% and pulled ahead of the 7800X3D by 30% in some games. The perfomance per watt of the 14900K is actually the best if the optimization is there. Flawed architecture my @ss. What is flawed are Intel's executives.
We all know what causes those crashes, it's untested clock/voltage curves for 5.8GHz clockspeeds and degradation. This is why i refuse to buy any CPU that clocks above 5GHz. Including the 7700X nonsense that AMD launched, not just the terrible core i9s.
@@Andy_S79Then they pull a Coffee Lake Refresh moment when the i7 (Ultra 7) tier because of the removal of Hyperthreading loses against a Ryzen 7 in gaming. The i7-9700K sucked so much, then Intel made the i9-11900K...
@@saricubra2867 An architecture can be well performant and flawed at the same time. I wouldn't rule out faulty design in the SOC, but it also could be because they pushed the chip too far (system agent or ringbus or cache clockspeed/voltages). We don't know the root cause yet.
Not to mention the previous generation spawn was like what, 6 years? Enough time to iron out kinks, and build up a consumer base instead of continuously fracturing it by changing the socket every 2 years.
Do a recall and you piss off the shareholders. Don’t do it and you piss off your customers and then your shareholders when your sales start dropping. No need to do anything, Intel. Problems only get better by ignoring them. 😂
The biggest issue I've experienced with my 13900K is the USB ports randomly disconnecting and reconnecting. Nobody seems to be talking about that. And, that's on top of the latency spikes, desktop lag, BSOD's using XMP RAM settings, game crashes and the unreliable 2.5Gb onboard network. My 9900K was flawless and my second PC, a 12700K, is flawless. My 13900K, which is my SECOND one (replaced under warranty by Intel in April), has been nothing but a nightmare. This weekend I'm driving to Micro Center and picking up an AMD 7950X3d and X670E AM5 motherboard. I feel like snapping this 13900K in half. Pure garbage.
My 2.5Gb onboard network port is unreliable too, but I have an AMD board. Seems to disconnect often when connected to the 10Gbe port on my modem (so it has a link speed of 2.5Gbps) but it works fine when on the 1Gbps. It's a Realtek. So maybe that's not Intel's fault (unless yours is an Intel network adapter)
@@jeffb.6642 On my 12700K system, I installed a Realtek based 2.5Gbe PCIE NIC and disabled the onboard NIC. On my 13900K system, I installed an Intel 82599 based 10Gbe m.2 NIC, which has been 100% reliable. I've also had good luck using Marvel based 10Gbe NIC's.
The 7950X3D is a terrible CPU, only one CCD has the extra cache. Better buy a regular one, not a scheduling nightmare. People compared the 7950X3D vs fully normal and stable 13900K, the latter feels faster because of Thread Director, AMD has nothing like it.
Losing trust is the last thing you lose as a business before going up in flames. Intel could afford losing to AMD in performance reviews, it can't afford this.
You understand you guys are the cult now right? A little gamer cult. Many Intel people are more mature, we arent fanboys, or brand loyal. We just have money, and dont just game with our computers. We don't take part in your exaggerated nonsense, and to you that makes us fanboys, I guess. Really, it's just that we have experience, and you don't. We know how this will play out. Intel always makes good. We have good reason to stand by Intel while the issue is investigated. Intel has taken care of many of us for decades. My daily rig is a 3930k (12 year old CPU). I've saved so much money with Intel. What number Ryzen cpu are you on? 😂 AMD just recently started being good after decades of sucking. Excuse us if we dont immediately jump ship like rats.
@@DingleBerryschnapps I am disappointed with my i7 6700k. You can just overclock the i7 2700k and get same performance. I sold it and bought Apple M1 Macbook. Thats what decades of monopoly does to a company. Plus Intel wanted to go diversity with the workforce and that cause 3 top engineers to leave the company and they couldn't get pass 10 nm. The company is a huge mess from inside.
Man, this whole situation is getting really bad for Intel and I feel for them… I have a 13900k system which I have been running for 13 months now and thankfully have not experienced any instability or crashing issues. (EDIT) nah I actually don’t have any sympathy for Intel, but I do feel for their customers experiencing such an issue. Maybe it will be fixed. If not then maybe switching to AMD in the future might be the right decision.
Don't feel bad for a corporation. Ever. They knew the problem existed during testing, and decided to push this crap out anyway. Lots of people are losing money over this fiasco, both businesses, data centers, and enthusiasts alike.
@@punishthemeatpocket No s###!!! Intel has a long history of shady business practices that have screwed consumers for decades now. I'm actually glad this happened. AMD is currently in a great spot to finally knock them off their high horse. I feel for the people that have wasted their time and money dealing with this. But I couldn't give a rats ### about Intel here. I hope this hits their wallet real hard. They need to be reminded that these calculated risks don't always work out.
@@meeponinthbit3466lmao someone doesn't remember AMD FX 💀💀 You guys are delusional if you don't think many corps released unfinished/ broken products. It is what keep the world turning. If everything was perfect, noone would make anything better
@@ThisIsExile91 you don't remember the REAL AMD FX as in FX 55 and FX 57. Intel paid most retailers(like best buy) and all OEMS(Dell, HP) to NOT use AMD CPU's for over ten years. They did this in USA, Asia and Europe. Completely bankrupted AMD and AMD could not quickly move on from the "bad" FX series that you children know today. You probably were a literal infant when intel started doing this.
I had an X570 MSI motherboard all of a sudden have USB overvoltage issues after having run just fine for 2 years. Switched to an ASUS TUF MB and good so far with FAR more usb C ports than I ran on the MSI. All the same peripherals.
@@sparkeyjames Sounds like the USB controller on the board, or perhaps a 5v regulator on the board. A solid 80% of the time I see weird failures, it's MSI cutting corners with low quality components.
I remember Wendel said something to Steve from GN about one of the failures that was happening on a test that ran purely on the cpu cache. The test was so small that it didn't have to use system memory to run. So in that light, the IO issue is sticking out as a likely culprit.
Thing is that it doesn't seem to be just power and temperatures at fault. There seems to be a fundamental design flaw in the cpus that over time gets worse due to the heat and power draw.
@@oxaile4021 from what I heard, heat has little effect on it. For some reason the CPU slowly fails from high voltage. Even if it has really good cooling. It fails faster if it is very hot, but even water cooled CPUs fail eventually. The only thing which seems to prevent them from failing is to consistently keeping the voltage very low, which would imply much less heat and low clock cycles too for stability. This may be a design error in the CPU ring bus, or maybe even a silicon technology fault making the CPU less tolerant against high voltages. It is important to keep in mind that those CPUs use nominally 1.0V-0.9V transistors in the cores and ring bus, but some can boost to 1.5V-1.3V automatically.
I purchased a i9-13900k for a Asus Strix Z790 ddr5 6400 ram for my gaming pc when they first became available and I paid around $600 for the specified performance. When the cpu was new it was blazing fast, but I rarely even played my pc and over time it has gotten weaker and if I use those Intel recommendations that cpu doesn't perform as good as my i7-11700k did in the pc I had before. Everyone should get the $600 performance they paid for or get their money refunded because mine doesn't even come close to running 5.8 using intels crappie weak recommendations, more like 4.7 Cinebench and that's the p cores.😢
My 13900K seemed to slow down over time, it was strange. It went from booting up right away, to getting strange shutdown errors, slightly slower (but still quick) bootups, and some strange lag watching RU-vid clips and multitasking. Then one day, I shut it down, went outside for a couple of hours, and it never worked again with a CPU light on the motherboard. Bought a Z790 motherboard, thinking the Z690 went bad, same thing. New 14700K worked on both motherboards. I just did the RMA today.
Intel needs to be careful with their server market once you piss them off it will be a struggle to get them back, I am surprised they hoped onto the AMD bandwagon really as they were pretty pissed about the Phenom II TLB bug issue from a decade and a half ago. That's how long they hold grudges for... but when you are refusing RMA/ add additional costs to RMA about a year has passed and still not known exactly what's going on that's a pretty big issue for them since time is money and the more down time they have they will be very, very upset to put it nicely.
If all 13th and 14th gen i9 processors will die with time Intel should upgrade us to the 15th gen chip and refund our motherboard purchase. I am never buying Intel again unless they make me whole.
Well, enjoy your new Amd pc because i doubt intel would RMA a old cpu with a new generation, also they would refuse refunding a mobo because that's hardware from a different company that doesnt have anything to do with them.
@@XFXGX but the motherboard isn't faulty, it's the CPU. And if you've been RMA'd with a 15th gen CPU, you cannot reuse your current motherboard. Conventionally nobody will replace your non-faulty board for you, you have to do it yourself. Intel's short mobo chipset support is now going to bite them in the arse.
What makes you think they won't repeat the same problem on 15th gen? This is happening because they couldn't compete with AMD and "had" to overclock the CPUs by default, from what I'm seeing with their hyper-threading dumping and no clear announcement to reply to AMDs 9xxx announcement, I doubt they have some secret super-performant CPU tech on their labs.
Some ways to fix some issues with 13th and 14th gen (13900k/14900k) 1. Limit watt usage to 262 watts in BIOS, gives some wiggle room over 253 watts from clocking down 2. Pin P-Cores to 5.4 or 5.5 ghz (whatever you are comfortable with), in my case I do 5.5 and 4.3 on E-Cores. 3. Turn off Multicore Enhancement in BIOS (otherwise CPU will go over 300 watts at stock) 4. Turn off Intel Speed Shift / Speed Step, both can cause stutters in games, but it will also run your cpu at higher idle wattage of about 50-60 watts 5. Make sure that 2 cores that auto boost to 5.8 go only to 5.4 or 5.5, or this will also cause a spike in voltage and stutter in games 6. Undervolt the CPU. I ran mine at 1.28v V-Core and Load Line Level 6 on Asus motherboard 7. Test stability in AIDA64, and then test stability of ram in TM5 with Extreme777 profile. IF everything is passing in stress tests, run games and watch for any instability, if your games crash, slightly increase your voltage to 1.285 or 1.29 volts. This will take time, but it will stabilize your system.
The whole issue can be traced to the memory controller overheating. The common thread in all the failures is the memory controller and its associated components. What the problem is unknown. But all the problems suggest that the problem is the controller.
People haven’t been paying attention for the last couple of years of how crappy the quality of everything has gotten. Motherboards, laptops, ssds, gpus, I have noticed so many things that fail far too easily and companies have been watering down customer service, just look at Asus. Asus used to have the best repair service (well in my neck of the woods) and you see how that’s going. I shouldn’t have to send a motherboard back multiple times to get a functional one let alone one that doesn’t look hella used. Intel probably knew before they even launched these chips there was going to be issues with them. They were just hoping to run out the clock on warranty. Intel will stay quiet, do the bare minimum and speed up the launch of 15th gen and start screaming “look at the new shiny”. Sadly Intel hardcores will totally buy into it like they always do. The fact amd has been kicking intels ass in performance and efficiency for years now and hasn’t completely overtaken Intel in market share just goes to show that the system works.
True the Intel and Nvidia greed again, very overpriced Hardware and have all Kind of Problems the RTX 3090 that dies or the RTX 3060/4060 Problem or the RTX 4080/4090 melting connectors 😂 and breaking pcb's ... Yes greed IS huge in Intel and Nvidia, Just dont buy ther products until they learn how to drop prices but AS WE See in Intel they dont learn at all and Just Put ther overpriced crap Out to everyone
On the contrary, I think that the quality of motherboards, GPUs and CPUs are great right now. But with everything "auto overclocked"; GPUs running over 300W and CPU over 200W, then yes, something has to break.
I had to disable enhanced multicore perf on my 13900k to stop the constant locking up - it's now less frequent. I don't even play games and it's never been overclocked, it's just being used for 3D work. Should have just stuck with the 10900k.
A 7950x absolutely dumpsters a 13900k for blender and other 3d work. The E cores fall flat when they need to work together with the main cores, the ring bus that makes the intel cpus fast for gaming really gimps any cores past 8.
I still have my 10900k it ran great for years never crashed.. I upgraded to 14th Gen and took weeks of tinkering to get it stable .. lol .. And it locks up still from time to time… Had to loosen a bunch of settings to get it to run right and I don’t even game on it😂
Lower noise? Have you ever been to a datacenter? It sounds like a jet is starting next to you... Noise is pretty much the only metric your absolutely not concerned with. ~~ /edit: But still, thumbsup. Good video.
Those server computer fans really sound like jet engine its scary 😅 I was helping my friend installing new server pc on his office and when he start to turn it on it immediately making this high pitch fan sound "dont worry its normal" he said 😂 boy it was scary i thought the cpu gonna explode
@@Napoleonic_S Seems you've never seen a server. I have 3 enterprise machines in my homelab and the ones with 40mm industrial fans go to ~21k RPM and the ones with 80mm fans go to ~16k RPM. Servers are very loud, individually or in a group doesn't matter.
@@Napoleonic_S Noise is absolutely a non issue in a server environment. As long as basic hearing protection will protect a technician from hearing loss, anything goes.
AMD stock went up, and they didn't even have to do ntn for real, but I feel bad for the ppl that was affected tho and those cpu's are mad expensive I can smell a lawsuit.
@@ofon2000 Yeah, in one video he was slating AMD for not innovating, saying they should remove 3d cache for a HBM. He doesn't even know how much slower HBM is compared to L3 cache.
This its like going back to Zen1 era first gen ryzen, the amount of "tweak", workaround and mental gymnastic to make it run and stable was annoying, i remember how Intel fanboys was blasting claiming how stable, how easy and the out of box experience of intel was, now things turn around i guess.
But how long will it be stable @ 140W if he is running a notorious CPU on the edge of stability The silicon might still be degrading @ an unacceptable rate. He should probably go down to 125w to give the CPU some breathing room. If he does not want to have to go down to 110w in a years time.
@@wessonliam7423 A properly designed chip should last at least 20 to 30 years even. If your chip does not reach even one year of use then it's simply bad quality. Intel should be sued for that.
Great video! I have fully switched back to the 7800X3D and the AM5 platform. I am happy to report that all is well. I wish I had this level of stability with AM5 a year ago. I also wish I had this level of stability with my Intel system. I have learned a lot and spent a ton of money this past year lol. But in short, I have a new AM5 video coming soon and I am giving the one year update on it. In short, everything is so much better than where it was a year ago. I am happy with my AMD system and it is very reliable so far.
So glad it worked out for you. I'll be looking forward to the update video. Like you I'm sure there will be many who will be prompted to make the switch for the peace of mind. For Intel's sake hopefully they figure something out soon, they're essentially just running a free marketing campaign for AMD 😂
Intel Owners 13th/14th gen DO THIS right away as your CPU's WILL degrade: I don't know who will see this - please thumbs up - on the new Intel - I've built may enthusiast systems and had the same issue with several as a lot of people has had. There is a way to make it stable (should work for most of you). Do step 1 - you should be FINE. If not go to step 2. Full explanation below but here's the fix for normies: 1. Go into Intel XTU - lock all of your P-Cores to 5.3 GHZ or 53 multiplier. Apply/Close. DONE 2. If your CPU is still not stable at 53, you need to add a positive voltage offset to make it stable. .05v or 50mv should be enough. That should make you stable. DO THIS ASAP: ALL of your 13900k CPU's and 14900k CPU's will degrade. Explanation as to what I found below. EXPLANATION: Why capping your multiplier makes you stable By capping your CPU at a 53 multiplier your CPU will not do a super high single core boost. The higher a core boosts the MORE voltage it uses. Intel has a boosting table and it shoves very high voltages into those cores going to 58 or 60. Sometimes EXTREMELY high. If only 1 PCORE is degraded - your computer will throw errors. Capping your WATTAGE will NOT WORK because a single core can STILL boost WAY too high as it only needs about 55 watts or so to go all the way to 6ghz. This is why the SERVER motherboards still degrade quickly - they are power limited but will still boost too high. For step 2 above - some CPU's are SERIOUSLY degraded - why adding an offset works Ever wonder why your computer only throws errors when it's under a workload of some type? That's because your degraded core (might be only one) is SO degraded that you need to add that 50MV positive offset to your CPU to make it stable. IF this still doesn't work for you - you need to lower your multiplier further (keep your voltage offset) maybe try 52. I've never had a CPU that degraded though. Questions? I'm here to help and will respond to this comment. If you need voltages I can do that for you. CT
Anybody with a brain would know that ramping up 4096W or a max current limit of 512A (which is technically more then enough to start welding at that point) is not good long term. I'm a all AMD user since the 486, OC'ed so many chips and one thing chips don't like is high currents, overvoltage (shootouts) and core's running into their maximum of what they capable of in regards of temps. These CPU's are not the only one lacking - there where NUC's, NIC's of intel camp that simply degraded over time. I had a AMD board with a Intel NIC that started to give erratic behaviour, one day just disappears out of the bios, followed with taking the whole motherboard with it in weeks. I think someone at Intel figured that nobody is going to use chips for a good 10+ years and knowing the avg lifespan of those chips they could cut it down by increasing clocks and thus shortening it's lifespan.
Sure and this IS why Intel keep doing IT because otherwise they Had No chances against ryzen 7000 😂 and Intel pricing was sabotatet by someone or how Intel scam ther customers with 4c8t for years and every CPU Gen need new mobo that was overpriced AS well 😂 you really think someone sabotatet Intel after ther snake oil Video leak ? Intel IS done a Long Time ago and all they do IS scaming and sabotating instead of doing good.
How much you want to bet that when the Dragons Dogma 2 fiasco was going on Charlie’s 13900ks was unstable when he tried playing it? Dude doesn’t even know, he’s rich anyway 🤣
What you said about DDR5 is 100% true. The auto voltages are often dangerous and unstable. Regarding raptor lake instability, it can be just one core being bad, you can try ratio-limiting individual core(s) to see if that fixes the stability problems.
There's 125W or more going through a tiny processor which was never designed to allow for so many cores / layers. If you inspect a broken CPU under the microscope I bet you will find that the material that is supposed to separate the lanes has weakened / melted, resulting in a short circuit when pressure is applied. This would also explain why it only happens after a year or so. Bending happens gradually, but of course it happens faster when the material is weakened and under too much pressure (from the infamous CPU latch or overtightened cooler).
Higher boost clocks, and moar cores moar bigger better multi-threading. Jokes aside, I've seen people buy into i9 CPUs and only play games while their CPU sucks back hundreds of watts, when a 7800X3D would've been faster and consume an order of magnitude less power. Some people are allergic to AMD for some reason, which I don't get - how are you allergic to a company? AMD-Intel-Nvidia they're all the same, you should only buy what is good value. I hope all these people are suffering from PC issues right now.
@@zdspider6778it was confirmed to be mostly Asus's mobo issue and there was 1 gigabyte board on a gamers Nexus video. 7800x3d works perfectly fine on Msi, Asrock and most of gigabyte boards. Heck after lot of bios updates Asus is stable too.
If you get it, lock the pcores to 5.5GHz and limit the power draw to 288w. You won’t have any degradation issues like these folks are experiencing. Set a negative voltage offset immediately
Then they are degrading. Simple as that. I recall Intel NIC's, NUC Cpu's and such having the same stuff. Its just a design flaw - intel damn well knows it but won't come out with a admission that would lead to a mass lawsuit.
Ive had an i5-11400 to replace the i5-10400 that had already burned out, it died on me within 12 hours of mounting it into my machine, this has been an issue for a fat second ngl, intel chips just BURN OUT FAST. Ive only had a single AMD motherboard fail so far, it's memory channels died of heat stress from being used for over a decade
I suspect Intel knows exactly what cause(s) is(are) and it's inherent in the design and they're hoping to get arrow lake out fast enough to burry the story and cover their tracks. Now complex electrical devices can be hard to DX, especially ones that also have to deal with QM effects, and if this was merely a few percentage points higher than normally expected, or even double, I could see it being still uncertain the exact cause, but with it being most 13/14900 k(ks/kf,etc) skews and many 13/14700k series parts I find it hard to believe with that many returned and examined they're still completely in the dark on this. They can set up hundreds of test systems and try thousands of variations until they have a repeatable set of failure cases. At the very least they should know what specific part is failing. Unless they've literally run into some new physics at the quantum level and are about re-write our understanding of QM in some way.
If I understand correctly it seems it’s mainly affecting the 13900k and 14900k but it’s starting to sound like a hardware issue. If that’s the case it’s possible that all 13/14 gen could be affected and weaker chips may just take longer to see the effects of the issues. Right now it seems only the 900k models are affected.
I read a report from a user with a Core i5 13600K who also experienced crashes, so I would not rule out that all 13/14 generation Intel CPUs are affected.
I have an Intel 13th gen processor and haven’t experienced any issues on my laptop? do I need to worry about my 13th gen processor? or what maybe I didn’t overclock my Intel cpu because I am still new to pc and have lots of learning to do, I do hope Intel has a fix for this because if we only have AMD then there is no competition and AMD tried with Nvidia but they couldn’t keep up with Nvidia so hopefully Intel doesn’t give up and keeps trying since they are really Important just like AMD is important.
Laptops are mostly fine. It's really only the 13th/14th I7 and I9 k chips. It's because they're being pushed so extremely hard to get a minor edge over the competition. Laptop chips underclock instead and all of the i3s and i5s are perfectly fine (other issues of course in buying them lol).
@@cuteAvancer The recent video from Gamers Nexus on this issue points to a design flaw, not just a core frequency that is too high, because even when the CPUs were clocked down, they still crashed.
Is Intel going to stay silent? I think so. They will replace some of the chips, they will silently make some deal with their big customers and hope the problem goes away with the next gen. At some point, they may push some micro-code update to downclock some part of the CPU or lower thermal protection max values...
Glad we went with a 12th generation processor when we finally upgraded last year. Considered AMD, but they didn't offer anything with equivalent performance for a similar price.
Intel needs to pull every trick to be on par with the far superior AMD parts. Now this backfires and the CPUs are dying because of reckless overclocking by factory.
i have a intel ultra core 9 laptop. everything about it from a subjective lvl is instable. if the NPU does something it causes "jitter" "delay" "lag" on the whole pc. some windows apps (for me smartphone link) causes periodically INSAINE frame drops, 100fps periodically 4 fps back to 100. vsync is straigh up broken and causes the weirdest copy frame buffer bug ever, where i see a .5 sec old frame pop up. even with vsync of i get that copy fram buffer bug sometimes with video overlays. and much more minor things. win 11 and intel core ultra chips are such dog poop its insaine.
Apart from the thing with the 4096w 500A that i had to lower to PL1=PL2=385w 420A because Games would crash on Shader compilation, I luckily didn't have issues with my 14900K. I don't use an contact frame. Cooler is an 420mm AIO. I now hope that it stays that way and that it isn't a defect that can suddenly appear later.
Nobody is talking about it, but it looks like these CPUs will have a horrible resell value. Intel customers wanting to "upgrade" will either be forced to skip 13/14gen and buy next generation Arrow Lake CPUs from intel (so that's good for future sales) or what's worse for intel, they will switch to AMD 🙃
Maybe they've reached the limits of their existing monolithic architecture and are now trying to maintain their dominant position over AMD by deliberately releasing processors way over real spec, resulting in processors that live fast and die young from overheating? Intel's history of sticking with an architecture regardless of how bad it's getting goes back to at least the Pentium 4. Until then, Intel 13th+ Gen owners really needs to underclock and under-volt their systems to a performance level comparable to older processors that are still happy to run just fine on air to alleviate the issue.
Thank God that I went for AMD during my last laptop purchase (although nVidia card still uses Intel I think). I chose AMD because I wanted my laptop to overheat less when gaming. I think overheating is what killed the battery on my previous laptop.
Intel should require lowering the power limits and should be the default profile. It is just a matter of time that more would fail due to chip degradation..
As a long-term Intel customer, I have tried AMD but it's just not doing the job for what I play game wise. My intel I9-14900K has zero problems but I don't overclock anything and don't XMP my DDR5 its 7800 running at 7200 its just fine at this with my Intel chip. no stutters, no BSOD no lock ups during gaming in apex/fortnite etc running a 4090 GPU. I am running a 420 AIO though with a high-pressure fan setup to get around the hi temp nonsense hi peak for a moment only hits 82C during gaming the 2 other 360 AIO's i used prior would get into the 90C range pretty annoying.
I bought the 14900k with the msi z790 edge and have gotten the “out of video memory” error quite a few times. This was my first all out build and it’s pretty disappointing knowing I have a ticking time bomb that costs $600 waiting to fail. Shoulda went with the 7800x3d like my gut said but iNtEl CpU gO BuRrR
I have a 12700H i7 on my laptop it is a good processor but I recognize an issue you showed at the beginning, the guy that was doing video editing on Adobe Premiere, I had issues like that, sometimes I listen to a video, unzip a 300MB file on my exter^nal SSD, and it lags exactly like that, the explorer to restart itself to fix it, I think it not just Intel, it's their drivers, and Microsoft too, all of those have a crisis issue they're in dire need to make an impactful and successful switch to ARM otherwise they will all die, we could totally imagine a world were most people than can put 2000 dollars or more on a laptop buy exclusively worldwide Macs with an Nvidia GPU, the people with lower budget resolving to a somewhat functional ARM Google laptop or Microsoft ARM with limited support.
I'm glad I did my RMA the day before Wendell dropped his video if intel is denying them all the sudden I contacted intel about my 13900 k and they took it back no problem I'm waiting for the new one to come right now, with that being said if I have any issues with my new 13900k the 7950x got my name all over it
@punishthemeatpocket noice! I have a 13600kf in my system till my i9 comes back and I've been super impressed, it is in the low 80s on cinebench clocked to 5.3 all core at 1.27v but I have also been debating when Intel sends me my new CPU instead of opening it selling it and going am5 I really like to Tinker so I am also looking at the 7950x3d because I don't mind having to use lasso, and that is if they haven't fixed the scheduling yet, they probably have.
@@danielabbott9312 I'd check out the non-x3D variants. Easier to cool from what I understand because of the v-cache forming a barrier. They are stupid fast, and you won't to worry about the scheduling stuff. Best of luck either way!
BTW if you live in a hot place VCCIO & SA can be unstable on 1.35v but thats perfectly fine voltage for these chips. You can set it to around 1.47 max if you have the cooling & you can max out your RAM latency +Core & Cache frequency (provided you have the cooling). E-cores should be disabled (or set properly which is not as easy), No AVX. 12,13 & 14Gen have dual IMCs & very often one is worse than the other. Thats where the problem is IMO.
I built a new rig 3 months ago. 14900k on apex encore. I work so much that I’ve barely had time to play anything on it but here in the last two weeks I tried to play both Pacific Drive and Once Human and both crashed multiple times on me. Only thing I’ve changed as of this moment is took the ASUS AI optimization off. I’m at work now and trying to decide what I should do. Get off work and try applying these bios changes and hope for the best or RMA and switch to AMD… and hope for the best and the RMAs go through. 🙄
Built my new PC a bit over a month ago and I'm using a i5 13600k, really hope I don't end up with problems like this. Didn't do enough research on CPUs unfortunately, probably would have gone with the 7600X if I had.
13600K OC 5.7 P-cores 4.5 E-cores 5.0 Cache. 1.37v AVG LLC4. 200W 90c. Not affected unless I want it to be affected. Pretty easy to recreate those errors, crashes and fix them back but for a person who has no clue what to do it can be impossible to solve. The CPU's are fine, it's all in the settings.
12:46 "...they are going to be complacent to gamers." Don't you mean down right condescending? A bit like Apple admiting hardware issues but only a year or two later and then setting up a *compensation fund* (doing the right thing). Journalists should absolutely pester Intel for a real longterm RMA solution for the affected 50% or 100% (c.f. oxidation). This will cost Intel millions (13 and 14 gen) and will tank their share price. Obviously Intel will *initially* avoid the subject or skirt around...
Some of the instabilities can be blamed on RAM (XMP on DDR5 can straight up not work at all, but baking XMP timings in manually might work for some reason - also since DDR5 now has its own voltage controllers, those things don´t have to work properly, since nobody is pushing RAM manufacturers to use the same level of parts as motherboard vendors use - i can definitely see DDR5 failing over time and much sooner, than DDR4 or older kits. I´ve already encountered RAM stick, that was reporting voltage fluctuations - it would drop to 1,08V running 1,1V JEDEC spec and 1,32V running XMP 1,35V - although it might be just faulty measurement, replacing the sticks with another set greatly stabilized whole computer. PSUs can also cause instability, especially if it cannot handle intel power saving modes (easily fixable for as long, as your motherboard has "dummy load" setting). Then, there are the BIOS settings. Before intel base profile was enforced, motherboard vendors used some kind of "optimization" profiles, which allowed the CPUs to boost freely, while running quite low Vcore and multiple other settings modified. It wouldn´t be an issue, if there wasn´t silicon lottery involved - my 13600K was unstable with "optimization" 1,2V, but with intel profile and Vcore 1,28V works like a charm. On the other hand, there was an issue with early 14th gen BIOSes on certain motherboards, when it would push 1,6V into the CPU. While i didn´t encounter that, i had an express failure on 14500, which started failing within days of use.
I never inserted my i9 13900K in original ILM on mobo,i use Thermalright BCF.If the CPU dies,i don't have warranty to it because of that. That would be problem if i want RMA with the seller.
Factory overlocks = High Heat plus voltage spikes and high clocks this accelerates current drift migration crosstalk through insulation barriers cause internal short circuits burning through insulation Layers … Moores law is Dead .. Need better design and new tech… It will take another 10 years to fix😂… New releases are just bandages to patch up previous designs.. lol… We also need better more efficient programming and code .. !!
The problem you described are several different problems not one. The lag for ex. is just E-Cores low speed & bad scheduling. Crashing in games is also incompatibility with E-cores.
Meh, they will all fail...... eventually, Intel moved the I/O block which is visible on the left handside if you take a look at the Die shots online. On the 12900K and or the 13700 and or 14700k this hasn't been moved, but to eek out more performance they moved it on the 13900 and 14900 series. This is going to 100% kamikazi all cpu's of these series eventually. Its just a matter of time. It will start with simple I/O errors and it will end up with you having to use those "performance" defaults which will stretch that warranty period just enough for Intel to go like "warranty expired Go f*ck yourself fanboy"
I tried to build a 13900k system, wouldn't even boot, so I sent it back for a 14900k the day it came out, it was so unstable I couldn't even get through a cinebench test without crashing. I have a top of the line system with a quality water cooler. I had to bring my computer out into the 40 degree garage just to barely pass it. I swapped out a AMD AM5 MB and got a 7950X3d and it runs perfectly. Get your act together Intel. And no it wasn't the OC on the MB I even set the CPU to enforce all intels limits which slowed it way way down and still it was crashing on any serious load.
Almost reminds me about my i9-9900k case. memkit rated speed was overclocking cpu memory controller. 6 months ok then failed to boot to windows -> dropped to jedec and worked for 3 months then had to RMA. When I got new cpu then verified that I run memory with controller specs. Now that setup is used by my niece and it's working ok and I'm using i9-12900k. It seems I will postpone my system update.
I don't understand what people mean by 100% failure rate? When it reaches that point does that mean it no longer works? That when you turn it on it immediately fails and cannot be used? Or does it mean that if you use it for more than an hour it will give yo a BSOD? Can you reboot it and continue to use it?
By 100% failure rate he means that the games and applications crashes everytime, and the processors performance will degrade so much over time. Just like someone's 13900k chip performance degraded so much that even a 11700k can keep up or even beat it
I had a 14700K, all stock settings, and every since the very first boot it gave me even more issues than the system it was replacing. Freezing, lag, stutters, random reboots, BSODs, you name it & I promise it was happening, from the fresh boot all the way to the second I tore the system down and returned it (Only to never be refunded, thanks Amazon, much appreciated, I'm sure Newegg will thank ya for the new customer as I will NEVER be buying PC parts from Amazon again.)
15:29 why couldnt Microcenter just hand him a new cpu from their store ? Bit ridiculous of them to have their customer deal with Intel, that should have been their job.
Oh and finally Wendels data is from Servers. From one hand thats good because they work 100% 24/7 & can be seen as longevity test but on the other the server load is totally different than a nomal PC user. Also the data is not large enough sample for final answer. Many of these chips crashed only one time. Just get a better cooler (or set manual pump & fan speeds) & dont push the RAM speeds & you should be good. Changing the LLC is also preferable.
Wow... I must of got very lucky on the CPU and mobo lottery's because I haven't had a single issue and I game with very high levels of CPU utilization with flight Sims and other intensive task very often also, never had any of these issues report and even under stress test running it at 300+ watt it's never bsod or froze or crashed.. it's without question one of my most stable systems I've ever owned..
I am now curious about 65W limit rather than 100W limit. latency vs throughput (cache vs CISC, single threaded vs multithreaded). We need tokens to manage latency via logic and memory. Cache hit rate ends up driving power loss which makes dynamic switching less problematic. Programmers need to use a high hit design. (This creates problems though.) It used to be we could experience a gain in latency through power limit, but I am not sure that is true anymore. I guess we can try cache performance and capacity. Windows can rebuke us. This is just sad. Now we will have real power users. Good luck to anyone in GUI. These people are considering a run on FPGA, get low. Talk about little Caesars pizza, but we ain't talking crazy bread. We talking spices.
what Bryan was talking about was not the same thing, he ws getting lag in his software stack when moving the mouse and he said first that it had to to witht eh io hub moving to the "mobo chipset", then that the cpu was a frankenstein of a cpu and not a monolithic cpu, which it is.
Two huge pointd you hit on ive exoerience was running something like a 14900KS stock and voltage being outrageously high and different XMP kits failing Karhu within a minute or two with nothing touched