Thank you being the only Tech tuber TMK to actually put an oscilloscope on a board and checking the best you can. everyone else just miss the spikes from what I have seen
Oh I can help you with that! Sorry to hear about your issue. Rest assured, Intel values all our customers. Just keep trying to use our RMA service so we can ... jerk you along for 3 more months with repeated denials and delays and pray you give up.
@@The_Noticer. big corpos, including AMD, Nvidia, Apple, Google, ... (the list goes on and on) lie all the time and only backtrack, when they get caught.
I feel like Intel is just trying it's hardest to not give up on the advertised boost clock. They are trying to frame the situation as if the voltage is accidentally too high, when in reality too many samples needed the voltage for those frequencies. The degradation is probably still happening, they are just slowing the bleeding.
They just have to slow it down enough for the chips to get past the extended warranty, and for Intel to have new chips on the market to replace them. I asked at the local computer chain if people had stopped buying Intel CPUs given the reliability issues, but they said it hadn't had a major impact on sales.
@@strangestecho5088 Well most consumers are ill informed in general, but AMD has been outselling Intel in the DIY space for quite some time anyway. So it hasn't effected the people dead set on giving Intel their money.
@@Denbot.Gaming amd is the same these days. 9000 is near-0 progress cpus made only to make 7000 with discounts to look better then they are. because CPU sales are down and nobody needs new cpus without progress with affordable GPUs. and progress with GPUs was killed to sell more consoles chips amd makes for sony/ms.
Many countries with stronger consumer protections would be extremely upset if the boost clocks dropped, this is Intel playing CYA to avoid having to refund customers there.
So this microcode is basically "if intel defaults selected, then turn on voltage limiter, else let the cpu request anything it wants".... Got it... Someone got paid 100k+ to write this microcode 🤣🤣
I am running 13900k at 5.2ghz with AC/DC loadline 0.60 both on 0x125 microcode, Ring ratio 4.5, E-core 4ghz my CPU is sitting at 1.268-1.3v. I don't need 5.8ghz or 6ghz for game performances as most single players run on GPU and multiplayer uses CPU cause of players nearby and my core does fine at 5.2ghz and i didn't have to undervolt. P1/p2 253W and 307A ICCmax.
i would expect further microcode updates in the near future lol it seems like theres plenty of ... room for refinement .... in the current fix tho perhaps understandable that they would rush this out given the situation
The new microcode settings don’t change much because if they send out a proper fix that limits the 6.0/6.2 boost on single core to a fixed all core boost of 5.7/5.8 which fixes all the high voltage spikes and degradation issues they are then set up for a legal challenge in court for false advertisement which leads to more issues than an extended warranty on degradation failure…
they can just argue core clocks are like ISP speeds, never hitting the "up to" number, this already has precedent with laptops. and magnuson-moss warranty claims are out of the window with extended warranties if they actually respect claims made by regular users.
@@KasasagiWad3 True, that's why the new microcode is still trash it still spikes voltage to 1.5v but only now it spikes for a lot less time so they can argue it hits advertised speeds still on 1 core and they reduced degradation hopefully enough where it doesn't completely require a recall while they screw people over denying claims for replacements due to degradation with other loopholes. If people would just manually change the stupid 1-2 core turbo overclock in their bio and cap the voltage to 1.4x max it would fix all of the voltage spikes and degradation on their cpu's and work a lot better than these failed new bios microcode intel fixes.
Hi Buildzoid, I know you said you don't find this topic all that interesting but could you make a short video showing a few boiler plate explainations of how to properly set up "safe" settings that don't completely gimp the performance for few different Motherboard BIOS like, ASUS, MSI, etc because the BIOS names and location seem be be pretty diffent across manufacturers. This kind of info would be super helpful & appreciated by the communiuty.
I second this, I really could use a straight forward, informative guide on how to get my 14700K properly safe and setup on my msi motherboard. This is obviously a channel I can trust with my pc's important bios settings, I'm reluctant to try other channels because they do not know what they are doing half as well as Buildzoid does! And I hate just copying someone else's settings directly and randomly implementing something as important as a bios, without actually understanding what these changes are and why we are doing them, copy pasta'ing a BIOS just isn't a good idea lol. Buildzoid would actually explain what the changes are, and which would be ideal and why. Here's to hoping maybe he will take pity on us poor simple peasants, and help us out with a guide to setting up a safe intel cpu, without compromising performance as much as possible
Same. I copied his settings from his last video (high/55/55/-0.130v offset) in my Gigabyte and everything seems fine, temps are much lower, voltage is fine... but he didn't mention it to do this video if you didn't trust Intel and their stock high voltages. So yeah, I'm a bit more confused now.
@@Alonne1 It's a fucking disgrace, I'll just say it man. This should have been implemented on such a hardcore low level that whatever any user does in BIOS, it should never override this, a true fix. I wonder if they have actually done this on purpose, to just push away anyone trying to RMA later down the line but who wasn't using their intel default profile. You can't help but wonder. It's just so lazy. After all this time...
@@club4ghzCooling has nothing to do with this. Apparently the problem is oxidation of connections inside the chip, which is greatly accelerated by the voltage that Intel chose to grill them with. Thus eventually the CPU stops being stable at factory speeds.
It seems that 0x129 is just another delay tactics by Intel, unless they show technical analysis on their end that it solves the issue and show that it is the root cause of the issue
They made a faulty product and now they do damage control with a code to prevent their costs, not your costs where you payed for an inferior product. It’s like you bought a car and the engine gets to much fuel and that damaged the engine and the fix the manufacturer do is lowering the fuel consumption to prevent further damage resulting in that the engine isn’t anymore as they advertised it to be with the horsepowers and performance you paid for!. Shiiit the whole tech market seems to be massively infected with greedy a-holes that skip corners for an easy buck and you get low grade products!.
@@tmsphereMinimal impact of new code on performance with the caveat that you’ve already applied the other incremental fixes that themselves lowered performance relative to launch day / advertised. So, not any additional performance losses, but still a net performance loss to have a chip that will degrade slower.
Incidentally raptor lake = dinosaurs. Meteor lake came when raptor lake started dying = comet exterminated the dinosaurs. Arrow lake = old arrowhead caverns post dinosaur age and the rise of manmals lol
I’m really glad I downclocked my 14900kf to 5.5 ghz all cores and locked my voltage to 1.30v right off the bat. I wanted it to run cooler and the voltages didn’t seem right, despite Intel saying they were perfectly fine. Excess voltage for an extra 200mhz on all cores didn’t seem worth it. On a side note, buildzoid reminded me of my brother when he wondered how many people just deal with it. I built a pc for my brother with a 13700k and he’s perfectly okay with dealing with random blue screen on rare occasions. He’s so stubborn -_- lol. I got him a 12900KS to replace it and he’s still like nah, it’s okay 😂.
there is no need to downclock it, set sync all cores to 57 so you disable this trash 2 core 6ghz boost and your voltage worries will be gone. If you want you can set IA VR to like 1450 to make sure the cpu doesnt get more than 1.45v ever. Then the only thing left to do is undervolt for more performance (since youre gonna be power limited if you go by intel limits as seen by buildzoids previous video / 253w = 40.4k~ cbr23)
Bro at 1.3v you should be stable at 5.7 all core... Currently my 14900k is running 5.8 all p-cores, 4.4 e-cores at an adaptive 1.29v max voltage 100% stable but you need to run manual settings with offsets and correct ac/dc load line settings...
@@joeysuzyjozy yeah, I’m only using manual voltage and set a specific llc. I used to run it at 1.260v but, I recently started playing the first descendant and the cpu kept crashing. It could pass stress tests of all sorts at 1.26v but that game won’t run stable. I don’t play a lot of games and this game is an Unreal 5 game, an engine 13th/14th gen seems to struggle with according to reports I’ve seen. I had to bump it up to 1.30v to keep it stable. Hasn’t crashed since I did. Good to hear you have a good chip.
My friend's 14900kf crashes for no reason during gaming. He has since purchased a 360mm aio for it, RMA'ed his 4070 Super and got a replacement and crashes still happens. I told him about Intel cpus burning itself to death, he did not believe me. He said all the techtubers videos say Intel cpu having issue are all paid by AMD. He's a proud Intel owner for 15 years and never have issues. He was going to RMA his mobo, I told him to humor me and let me flash his mobo first with 0x129, ofc him as a true Intel fanboi said flashing bios will break mobo... I told him you are going to RMA it anyways right so why do you care? so he let me do it. After that he's been gaming no issue for about 16 hours straight. And today he told me... see I told you all the techtubers are just paid by AMD, simple BIOS flash fixed all his issues, so Intel is still the king of gaming cpus... Why is he the way he is.
I have tested the microcode on a Z790 AE AX -W and the Intel fix actually decreases performance by around 2000 points in R23 compared to the previous BIOS with 253W/280W/350A/1.5V limits. For some reason, the Intel Extreme setting limits the power to 253W/253W. Even when increasing the limits to those specified before, the score doesn't increase by much, plus the voltages are clearly higher than in the previous BIOS (although under 1.55V). At this point I am satisfied with turning off the Intel Extreme setting and manually setting the limits to the 253W/280W/350A/1.5V (although the Gigabyte Optimized setting will clearly undervolt the CPU).
11:50 I'm a maintainer of some popular Binding of Isaac mods on steam, and I've had at least one person contact me because they bought a new PC and their game was crashing on random mods including mine, and it ended up being this exact issue. There's definitely a lot of people who don't realize that this is happening, and I don't like that I now have to include **out-of-the-box hardware issues** as part of my troubleshooting steps
Go no further, I have a friend that works in a retailer and he had no fucking clue this was a thing, and went as far as tell me no one is complaining so it's not a problem. Edit. I had to educate him on the spot.
@@Joao_Dario The thing is... in raw numbers it is a lot of people experiencing this issue, as many CPUs are being sold. When you would look at percentages though, the number is relatively low. There is no exact number known but based on reported return rates, one can estimate we are talking about 5-6%. So 94% of people buying these CPUs will in fact not run into this issue or are unaware. So yes, from a retailer point of view this might not even hit their radar. That being said, 5% is an unacceptable high rate of course, plus it is only an estimation and most likely a low estimate as 14th gen is relative new (so limited amount of time to get affected) AND some affected people might not even be aware.
You described me perfectly. Early days into the 14900k issues, I replaced basically every piece of hardware in my PC thinking it was to blame for game crashes. I bet there are a TON of people out there like me.
Makes me glad I only went with a 13600k this time around. Max voltages I see at stock settings are 1.27 @ 5.1 all core, easily handles 5.5 @ 1.31v. Not touching these bios updates with a 10 foot pole. They say some of the i5s could be affected, but I just run at stock speeds most of the time with a -75mv undervolt and have been since late 2022. If they don't get this sorted out by 15th or 16th gen, I may hop to team red for my next upgrade cycle by then.
I built a 14600k for a client, I typically run a week long stress test which initially passed with flying colors, XMP enabled and everything, even manually configured intel default power limits vs the terrible ASUS defaults. Fast forward only a few weeks later, the computer wasn't even being used but just powered on, and started getting random crashes. Now the system is unstable with any XMP profile, had to reduce ram to 4400, which is "within intel specs" so it won't be RMA'd even though it's significantly degraded already. It may be the initial week long stress test responsible for increased rate of degradation, but I've been doing this for 20 years and have always tested systems extensively without any issues. For my sanity I'm just going to swap to a 12700k and eat the cost on the clients behalf. I'll now be exclusively using Ryzen or EPYC 4004 for high perf client systems. These modern Intel systems are just a headache. When defaults won't even work, it's a failed product.
Good idea. I have a 13600k with no issues but I just updated the bios and then my VID went up to like 1.42v and my temps 95c. I turned on intel defaults and same deal. I ended up having to use CPU Lite load to bring everything back down. I should have stayed away.
@@rover1374 depending on your motherboard manufacturer, you may be able to downgrade the bios to the previously installed version. With my MSI board I was able to re-flash the bios I had previously (version "v1F" from november 2023).
11:48 At that point you started describing a phenomenon I noticed a long time ago in GPU land: When the game crashes Radeon users think it's the Radeon drivers. When Windows BSODs or spontaneously resets, likewise. Guess what Nvidia users think no matter what the Windows Event Log says? So now I wonder: Will an "Intel + Nvidia diehard" rather blame Intel or Nvidia in this case? :P
Reminds me of some forum post i saw of people complaining how their game runs way worse once they ve upgraded from NV to AMD GPU, and naturally they blame the radeon drivers. It turns out that Nvidias compiled shaders (who can even survive DDU) caused issues and game reinstall is required to get rid of them
It's still insane to me people consider occasional crashes normal. If my PC crashed I'd investigate the cause immediately, luckily it hasn't happened on my Linux daily driver in years now. The only crashes I got were when I started games with anticheat on my Windows dual boot.
@@PlayJasch I'm a software developer myself. Devs aren't stupid. Management is stupid. They think code output can be measured the same way you can measure how many milk cartons a machine can fill per hour. Fortunately in my line of work the wrong people die if the software isn't up to scratch, and the budgets are there to make it happen. Doesn't mean we don't have to give management push-back every so often, though. But it's not like with consumer crap where every cost saving measure must be used to make the share holders happy.
Funny thing: ASUS beta BIOSes "forgot" the Intel profiles performance/extreme are both set to ICC Unlimited = ON / ICC = 501A / P1/P2 = 4049W, also some board types from ASUS had AC at 1.1 again. At this point, you can basically trust no one anymore to handle your hardware right, you have to check every setting.
Great video as always. You forgot to mention how Intel’s ILM design has been bending CPUs over time ever since gen 12, that we’re forced to buy Contact Frames to protect them from bending.
As soon as I select the Intel Baseline Defaults on my asrock z790 steel legend wifi, my i9-13900k requests 1,55v VID. With the same settings, selecting ASRocks 'performance' baseline, it requests no more than 1,36v VID. This behavior changed from 0x125 to 0x129 as previously it didnt change much between intel baseline and asrocks performance baseline before (slight VID up with intel baseline on 0x125 to 1,40-1,45v VID).
@PrismVuxOFL "128GB DDR5 Installed 4000Mhz (For some reason?)" if all four ddr slots taken, XMP can be activated/or clocks raised, if you raise VDD and VDDQ voltages to DRAM mnf specs. Also unsync all banks and set voltages from mnf there as well.
i told intel if they don't refund me tomorrow, I'm opening an arbitration case with them with the AAA. I've done it with Dell before. Dell gave me my money back immediately after starting the case, and Intel will too, because it costs them more than $2000 just to open the case with AAA, it just makes more sense to give me my $552 back.
I locked my vcore down after a day of noticing ny 13700k running hot and well over 1.4v. Looks like this microcode is a nothingburger for those in the know
Have a water cooled system. i7 13700k CPU, a contact frame with an Asrock Z790 Steel Legend Wifi board. Before the intel bios update, it would instantly thermal throttle at 100°C whether Intel's limit was activated or not. After the bios with microcode update, it preformed EXACTLY the same when using Cinebench R23 200. I was ready to give up until visited Intel's site to see their power default and manually set them. 125°C base and 253°C Turbo. Lord knows what Asrock sets it to despite what settings you pick. It for the first time actually doesn't thermal throttle. The Asrock Microcode update was a complete failure and waste of time from my experience unless it does something else that I am unaware of.
most people still dont know that, 13th/14th gen i7/i9 run super hot... so hot that it was practically impossible to cool them with ANY kind of cooler available in the market...(no matter you do the microcode update or not) untill you actually put a power limit or undervolt and bring down the clocks. I realised this before getting my 14700k and got an air cooler. after initial tests with 310watts, i knew it would throttle, now i put limit to 235watts, and temps remain stable at 90deg C.
Glad there's a finally a working fix for people's whose cpu are still in relatively good shape. I wonder if they've discretely rolled this out to laptops with 13980HXs, which even though power limited on wattage, shouldn't be voltage limited any different than a desktop 13900K.
They should in fact aim for a completely different working point on the power-performance curve for the sake of efficiency. Power consumption scales linear with frequency and square with voltage, so any little bit they can get on driving the voltage lower helps their typically horrendous thermal situation and marketable battery run-time. I can't say whether they do or do not but it would make sense that they would, and that they would get binned correspondingly for better phase margin at low voltage, and laptop chips sell at a premium. It would also make sense then if true that intel wouldn't be concerned about early laptop failures.
14900KS here. Update from maybe a week ago. Went from A-OK to things crashing often (mostly certain games; essentially unplayable now) in the span of, well, that week or so. Unfortunate. I guess some of the more spread out game and app crashes (once a week, once every other week) were early warning signs after all.
@@soulofjacobeh oh, i understood wrongly you bought it past week. Sorry dude. I'm in the same situation with a 14900K. But i could stabilize it by lowering down voltages and freqs. Problem for me (blame myself for it) is i bought it 2nd hand, no invoice, even the previous owner. So i'm truly fucked if it keeps degrading.
@@Pacho18 I've had some RMAs that didn't require proof of purchase. Intel might be lenient, so it's worth a try. And yeah in my original comment I was (vaguely, sorry) referring to a comment I'd made on a prior video about a week ago in which I'd thought I'd avoided the degradation issues somehow. But the last 100hrs or so of runtime have rapidly deteriorated in application stability. I now have to use Intel Failsafe SVID
I am forever glad that when I decided to upgrade my 5950X to a newer, gaming-focused setup, that I went with the 7800X3D instead of going with the more expensive Intel route, cuz holy crap would I have been pissed with all this going on if I'd paid ~$600+ for that CPU.
Yep, lots of Intel shills are pissed there $600+ CPU kills itself on arrival basically. Even coping by saying "just rma it haha" until Intel rejects your RMA lmaooo
How ironic. I'm currently building an "open box" intel system with a 13700k, and to get anything that would even come close to competing I would be spending double for AMD. I don't have any brand loyalty and I was really trying to put an AMD system together this time. Not even with returned/open box products could I come close to my 13700k ($235) and Asus strix z790i ($213). Open box but unused. I really tried though. Edit: Even if I had purchased that CPU before, I don't think I would be pissed. I've been using Intel products for two decades and they've never let me down. I knew they would have fixed the situation and make good with their customers. Which they have, with fixes (still in beta) and a five year warranty. 👍
@@DingleBerryschnapps Nobody is talking about "open box" CPUs here. At one point I used Intel CPUs too, hell most of us did...because they were better however things change and AMD makes the best gaming CPU on the market right now.
@@DingleBerryschnappswhat are they going to do with all the small custom PC businesses that are going to have handle hundreds/thousands of CPUs that now need to be replaced or at least have the board BIOS's updated then tested? Is Intel going to pay for the shipping of entire PCs, the extra time and labor? I flip PCs but thankfully never sold any with 13th or 14th gen Intel units (AM4 and 11-12th gen, mostly). You bet I'd be freaking out if I had sold ANY of those cursed chips to a customer.
@@smugmode Have you seen PugetSystems article with 11th gen failure rate? It's kinda not that big of a big deal, but at the same time, the 13th/14th gen customers are prompted to seek instability because of the articles, so there definitely are more false-positives, or otherwise undetected instability issues. I think it means for the future, you should be sticking to CPU's manufacturer spec closely and disable MCE etc.
"everything is unstable" Might be a particular extra layer of fun if you are using one of those new cheaper DRAMless SSDs with Host Buffer Memory and that has to use your CPU as its memory controller.
I updated my BIOS on my Aorus z790 pro x to the 0x129 and my 14700k thermal throtled instantly on the XTU stress test. I am not good with BIOS settings unless I have a damn good guide (meaning I will find where to change something by following a guide but to find the setting on my own is something Im not good at and frankly a bit scared to mess around with) So before 0x129 I used XTU and set my core V to 1.260 and my TB's to 280W (custom loop). I have max 1.37v spikes with about a 1.27v average running stress tests like XTU and CB R23. All cores at 5.5/4.3 and max temp on CB R23 is 93C. I would love to set this in BIOS but I cant figure out how because I get to many volts and 0 stability (due to my lack of BIOS knowlage) and also XTU likes to reset to default sometimes hence why I would like to have this BIOS side. Also is it just my MB or leaving the Intel Preformance setting on wont let you undervolt even if you disable it BIOS side?
I'm just confused with all this situation. So if Intel's defaults are off, cpu doesn't get enough voltage on load and so some workloads will crash, but if you turn them on with the new microcode it just limits max vcore because before and without it was too much. I'm just lost.
Intel is trying to burn new code into cpus so motherboards no longer feed the cpu - voltage above 1.45v or so. When the board is at a certain voltage and Intel's boost requests the next high-level voltage, the cpu may get over voltage'd by the mainboard when it's is so close. Its a delicate balance that is required to be adjusted in two different ways because that is how the components are designed, there is just a lot of interplay going on between them. This is not an easy problem to fix with the devices in peoples hands. Cya!
Also add the ILM bending to the list of LGA1700 issues, and according to Noctua the bending gets worse over time And they removed AVX512 from 13/14 gen, what a shitshow
Still fine for 99% of the people out there... Not every kid wants to play RPCS3... Stop the drama. If you are not happy, switch to AMD, problem solved and no more whining.
@@FEGTTTSDH They can't, its fused off in all but the early batches of Alder Lake. Also since the e-cores don't have AVX-512, they must be turned off to use AVX-512 on the P-cores.
I got back my RMA'd 13th gen (13700k) a few days ago and its been night and day. I can't believe I waited this long for an RMA - I purchased the original CPU Jan 2023 and had it undervolted the entire time (CPU Lite Load 1, MSI motherboard) because I heard that it runs hot and I prioritize heat/noise. Instability occurred over time but I thought it was a problem with my wifi, a problem with my games, windows, etc. Windows wasn't crashing every day, but something would freeze/crash at least once a day. My computer effectively runs much faster and smoother now. It's crazy how much tolerance retail has - maybe because this is my only desktop. If I was working between two desktops with different CPUs, I probably would have complained earlier.
I power limit my 13700K to 95W, and it hits 1.1V under load. Still perfect for my Civilization II games 🤣 It's actually quite efficient when you don't drive it hard. (Seriously, it does suck for people who bought it and need the full power of these chips)
@@peterwstacey I used to undervolt, until Asrock completely broke the Microcode update, even updating to a more recent BIOS and selecting the bios that would let you undervolt... nope... Its funny how when you set 1.3 as max in Bios, you boot into Windows, and see 1.57 LMAO Intel is fucked. No more Intel for me after this. 620 euro CPU thats fucked...
@@Yes-sw8gh Totally understand the frustration. I have an insanely long platform refresh cycle (P4 -> Q6600 -> 4770K -> 13700K) so I'll see how Intel looks in 15 years. If I need to get a new platform now, I'd definitely go AMD 7800 non-x. Hope you get yours sorted.
Your modern high performance CPU can run a game from 28 years ago while undervolted, not sure what point you're trying to prove here lol@@peterwstacey
I have a 13600KF on a "low-rent" Gigabyte B760 DSH3 DDR5 board with 6400Mhz RAM. I have never had any issues (apart from high VRM temps which were ameliorated by a BIOS update back in the mists of time last year). I did however, for the sake of peace of mind, update my mobo to the latest 0x129 bios for my board. I have not experienced any slowdowns, my CBr23 scores remain at around 23K. What I *HAVE* noticed though is that total package power now never exceeds 155W, whereas before it would routinely hit the 181W limit. Maximum VIDs have reduced by around 30-40mV (averaging just over 1.33V on the P-cores now) as far as I can tell. VRM temps have dropped even further to max out at around 70'C. I generally leave everything on AUTO in the bios except for enabling XMP and rebar and what have you - I'm not interested in overclocking it. My bump of trouble tells me that most people running B-series chipsets won't experience issues even without an 0x129 update (even with 14900Ks) but have no data to back this suspicion up. B-Series is just pure bondage and discipline for K-series SKUs, although it seems it _had_ been supplying more power than my CPU actually needed at full bore.
I was under the impression that the microcode update applied to non-K processors. I mean when I'm reading the BIOS updates from MSI or ASUS, it seems to be non-K processors. I haven't seen it confirmed for both k and on k
14900k sufferer here. One question that I have is if manually setting that voltage limit to 1.55 makes it safe to turn off the intel defaults? I'm a big tinkerer, and I'd love to be able to overclock and tinker with my chip. I want to stray away from the Intel defaults in some ways (e.g., I am considering removing wattage limits). Is setting that voltage limit manually going to work equivalent to the Intel defaults (except for where I diverge manually) in terms of chip safety?
Thanks for the video! But why didn't I try tuned IA AC/DC Loadline (for example 60/110) on Gigabyte default profile? Also, the "CPU Internal AC/DC load line" enabled on the "Perfomance" on the Gigabyte default profile can help with instability.
It's strange how some corporations are able to hypnotize customers to the degree, that they buy a product that is advertised for its high performance 6GHz and its reliability and them downclocking it to even reach that reliability and being just fine with lower performance despite them having paid that premium price in the first place. No wonder corporations keep doing what they do.
I have a question for the AC/DC load line video: Is AC load line measured on a specified frequency or manufacturers tune it for their VRM controller? Simply put PCB traces have different caracteristics on different frequencies, so for example a motherboard have 0,400mR at 100kHz and 0,442mR az 400kHz. Some cheap motherboards have a fixed 200kHz freq, other more expensive ones have a range 200-600kHz. What is the minimum specified VRM response time for AMD and intel? I mean the time between when the CPU request a Voltage and when it gets the requested Voltage.
I personally turned on intel default settings with the microcode update plus turned on XMP 6000 on the RAM. I have high bandwidth/low latency disabled as well. Before all this information came out I was just running whatever profile was on it, think it was the gigabyte optimization setting then turned off multi core performance. This PC was built only like 9 months ago, i have a 14700k
I can say that my old Z690 from Asrock was giving me very reasonable values. It died and I had to switch to a ROG B760 and the values were completely bonkers. I tried the X129 and the problem I have is that my 13600k still hits 100C° and keeps thermal throttling. I have a quite decent Cooler master twin Fan waterblock (getting hot, so it works) and I already set the Long duration boost to 125W from the 181W that the Bios sets as standard. (181W for the short duration too.) It is quite hot these days but a CPU should not hit 100C° under normal gaming load. The ASUS advanced Bios allows to set a 90C° as a threshold , and still it never hits above 80-85C° under ASUS optimization and I often get better benchmark results, despite manually setting the CPU values to Intel recommended. The Voltage spyke often hits 1.46v through, which is quite high for a 13600K but that happens under Intel default as well. I will try to set a limit to 1.4V and see if it's stable but I cannot have my CPU thermal throttling constantly just because the Intel default settings on x129 lacks the tools to tame the CPU at reasonable temp values. The CPU also Spyke's anyway to 1.465v for no reason while switching applications on desktop environment: I'm not impressed to say the least! Even opening HW64 is enough to see spykes. You are right: the platform is a mess, but it wasn't in 2022. Intel and the MB manufacturers made it like that to be competitive with The Ryzen 7xx0 platform: Intel wants it's offer to look fast and is burning our CPUs to do that.
n1 videos even if sometimes a bit longer. did u test disabled tvb/singlecoreboost to get rid of the high vcore spikes or is this a urban myth in youtube? thx have a n1 one
y-cruncher has a 2nd mode that is mostly CPU intensive, not sure how to use it via bench mate. If you start y-cruncher directly, select option 1 and then 3 for a quick test or higher numbers for longer tests.
3:40 I have a z790 Aorus Pro X and never touched my AC LL because I never had the experience to change that nor understand it. I knew I was right that the curve my motherboard was doing was not supplying enough voltage on high loads most especially multicore / more than 2 cores active loads. I never figured out how I could set the AVX offset to off instead of the default Auto that keeps changing settings depending on what my motherboard feels like. On default, it kept playing at -5 or -7 and I hated that the negative transients was caushing blue screens on my PC while I was undervolting it. I might as well as copy majority of the settings that you have on your 14900k into my 14700k and then play around with it to get stable. This band-aid fix is REALLY disappointing.
What do you do when you have an HP Workstation with a limited user controled BIOS? One thing I'm going to do is swap in a better screw down thermal clamp on the CPU instead of the default lever clamp. It may void the warranty, but at least the cooling efficiency will be a little better on the chip.
Great video. Question is im on z790 aorus prox and i got the intel defult settings after i update my bios what if i go to my old update F3 ( the bios that My pc came with ) which is the previous update from F4 that shows intel default settings is it safe to do? Plus after updating the gigabyte control center tells me that under vault protection is enabled in cpu tap in the app I didn’t hat this notification before when i
Funny enough my CPU was spared because when i bought it 2 years ago, top youtubers called to disable multithreading enhancement which is the future that pushes CPU even further. Well, thank you guys for saving my CPUs. I did also applied a small downvolt for temps just enough not to lose performance. Checked my VID now. is below 1.270 using adaptive voltage offset of -0.100 My bios is from january and my Mobo is a Gigabyte z690 UDD4 with f28 bios, I will not update to microcode
@@Summanis the only setting i made was to limit power to 235watts after microcode update... i could not find IA VR or VCore voltage limits, which i wanted to set at 1.4 Volts, i have the same gigabyte motherboard, but slightly different interface, its all so confusing...
@@realracing3specter295 For real. I try searching up what all the settings in my MSI motherboard do and half the time nothing related comes up. I still have no idea what "memory extension mode" is. I haven't even checked to see if my motherboard has a similar voltage limit (12th gen) but I'm sure it would turn into a rabbit hole.
I set mine to 1.45 max. It will hit a 40300 on cinebench with that limit. Now that is higher wattage but voltage limited. Intel defaults are down from 38500 to 37700.
I updated to the beta bios with the new microcode on my 14700K intel performance (Asus doesnt give me an extreme option). ICCmax to 400A, PL1/2 to 320W, IA volt limit to 1.5V, AC/DC loadline to 0.5/0.5. And the VRM loadline to the recommended 4. As is, that runs CBr15. Overclocked the cpu to 59x and now looking into undervolting the VF curve... Seems that this chip runs 59x on 1.44V, but is capable of a -0.05V undervolt. Which is a worse result than before, where it could handle a -0.08V undervolt, so the new microcode is doing something. Checking the next VF point: 56x runs with a -0.20V undervolt around the 1.22V. Work in progress.... Still lots more testing needed, but the chip still seems to get way more voltage than it actually needs on the new microcode....
Enable 'Enhanced Thermal Velocity Boost' or in a more techy name 'VMaxStress': is just a voltage reliability on/off switch. Optimized AIB settings always disable it. Also for XTU users: touching anything in XTU to overclock will disable the feature too. Intel's fault(and the reason they are lax with warranty) was making a mistake in their voltage reliability feature which was allowing the CPU to run at unrealiable voltages: that's what 0x129 microcode fixes. AIB do know how to properly configure Intel CPU: my ASUS Laptop has AC_LL and DC_LL perfectly configured, as well as other things like TVB, PL1, PL2, ICCMax and even ASUS propietary 'Regulate Frequency Above Threshold'. For Desktop they just use AC_LL for whatever their reasons are: maybe it's some social engineering for home overclockers, ease of configuration,...
Thank you for the video, really helpful. Can we run the XMP mem profile with the intel default settings on ? or is it better to keep the profile disable with the intel default on. Also after applying the intel default settings the temps seem to be really high 80-100. If this keeps up would you reccomend doing the 1.4V limit? ( if it would help)
People just all core overclock your CPU -> fixed Pcore, Ecore and Vcore. No troubles then. Or turn indeed the IA VR voltage limit setting on 1.40v (value is 1400 in BIOS). Maybe KS owners do put it at 1.425v (1425), that is stil safe. Or put fixed max on your Pcores and E cores, then the VID never requests high values, so P core 5500mhz, E core 4200mhz or so. Good luck it is so easy to solve. Don't do BIOS upgrades, all not needed.
I wonder if there's a specific setting that enforces the 1.55 voltage limit, or if it's really a hidden flag. So far I was thinking that the Intel Defaults would just be a preset for the available BIOS settings, but they might actually do more, like setting such kind of hidden flags. 🤔
id really like to learn more about how vid requests are generated cause it really does seem to me like the microcode fix doesnt get to the root cause here rather it's the only feasible fix for cpus that have already been manufactured
I just got a computer with the i9 13900k in it and learned about these issues. I did my gigabyte motherboard bios update. Some people told me to do the gigabyte default clock. Do you agree to do that and will loading the defaults hurt anything or delete any data/apps?
The GOPsters are making moves to have off-grid people pay for "lost utility company income" in some districts. You'd better churn a lot of butter, Eli!
@@LilMissMurder3409/videos I'm not even American so it doesn't really affect me, but I wouldn't vote for the Democrat scum even if they paid me like they do with everyone else. Not that I think any of this is relevant to Intel being a terrible company.
Considering I spent 4 grand on my computer and another 4 grand on my sound system for my computer, it would certainly be cheaper to be Amish. No BIOS headaches.
I think you are probing the correct way, 50 ohm termination resistor and a coax cable with the difference in shield and signal wires cut as short as possible and ground point soldered as close to the signal wire as possible to decrease the inductive loop. If you find a voltage probe point on the CPU substrate consider using channels 1 and 3 to prevent multiplexing of the ADC's if it has two, otherwise ignore my advise, my own scope is pedestrian at best. Also YES some explainers on the various BIOS settings and OC flowcharts are very welcome.
No way 1.55 doesn't still cause degradation. Maybe they'll also limit how often you can have your advertised boost clocks and most chips will just about about make it through the warranty?
If I leave intel default enabled my 13600k runs 1.45vcore in Cinebench r23 multicore. and instantly hits 100c+ If I change svid behaviour to typical it lowers vcore to a normal-ish 1.25ish but my r23 multicore score is 2000 less than normal for a 13600. This is on a 690-A from Asus.
People just all core overclock your CPU -> fixed Pcore, Ecore and Vcore. No troubles then. Or turn indeed the IA VR voltage limit setting on 1.40v (value is 1400 in BIOS). Maybe KS owners do put it at 1.425v (1425), that is stil safe. Or put fixed max on your Pcores and E cores, then the VID never requests high values, so P core 5500mhz, E core 4200mhz or so. Good luck it is so easy to solve. Don't do BIOS upgrades, all not needed.
Same deal with me on 13600k/msi b660 baked. 1.42v and 95+ temps. I enabled ‘normal’ mode and then chose the cpu lite mode setting 6. Temps back down to 75c and at 1.30v. Cinebench score is about 22600, so pretty close to the standard score. What a stupid fix. I wasn’t having any issues with my 13600k but decided to update the bios for the ‘fix’. The fix would have cooked my CPU.
I had to return my 14700k with the performance degrading by the day for a 14900k. (I know risky.) - But so far, before and after the microcode this thing runs 15c cooler than my 14700k at the same intel default settings that the 14700k was overheating with. With the same exact set up, the 14900k ran cool and no issues so far in comparison to the 14700k was thermal throttling and at times, unstable. Also.. I've been enjoying your videos and here is my sub!
@@nightwing8794 In short, I had bought mine at Microcenter on Black Friday it came with the MSI Z790 bundle. About less than a year, approximately around 6 months.
@@nightwing8794 To add to this comment, what's weird is no matter what, it was overheating with even undervolting / underclocking. We thought initially it might be the aio, but like funny enough none of these issues have appeared in the 14900k yet, it's been rock solid thank god.
Hey mate. Thanks for all these videos and tests. Right now, i have 14900k-msi z790 with latest 0x129 non beta update. Im using it with lite load 9 and ia cep disabled because it comes with ll mode 16 and that causing idle voltages being too high. Like 1.3-1.3 vcore and 1.48 vid. And if i dont turn off cep and lower ll mode to 11 or lower, cep kicks in and cpu performance getting capper from %20 to %50. So i wonder, if i dont change any settings from default intel profile but only change ll mode 9 and disable ia cep, would that also cancel the microcode update? I dont gent voltage spikes in hw info but there still could be transient spikes. Can you test? Thanks.
Does this mean that the VID-Limit will also be disabled when I change the AC/DC LL or set a Vcore offset when using the "Intel default settings"? That would be a bummer. Want a little UV but also the Vcore Limit.
Great video, I'm tempted to set up my Tektronix oscope then capture voltages on my MSI motherboard. I'd love to see what's happening in real time although I'm uncertain where to pick up the signal. CPU power plane is easy to find but would that be the correct voltage. Hmm...
So for my undervolteted 13700k its more save to keep my undervolt settings ? Never hit over 1.3 Volts. I did -95mV all cores expact the 2 Fast cores. They got just -65mV to get the 5.7 GHz done.
13900K/MSI Z790-A wifi bought in August 23, since the very beginning I ran adaptive undervolt of -0.035V on frequencies above 5GHz, with the MSI default profile for water cooler (i.e. no other limits). Never experienced any crashes, passed all the usual stress tests like Cinebench, prime95 etc. Just ran Cinebench multicore, 40293 and max VID during the test in HWinfo was 1.42, with averages around 1.33V. Don't think I'll be updating to 0x129.
I love actually hard core overcloking. it's a lot of fun to watch. intel, historically i've liked your cpus as well as amds. But honest guys, while i like watching this, i don't want to live it. What i'd like is to forget about voltage and use my computer care free. I'll wait for the update and just use it i guess. I mean for now i'm running at 90w, and 1.05/-.07v (adaptive offset, because i like cool and quiet anyway 52/22 multipler. I like to set these settings in the board though and don't want to deal with this problem. BTW i can't tell the difference from the voltage melting settings, and the computer run quiet with no heat. Food for thought. Maybe we've all gotten too focused on benchmarks. Amd is paying that price now. maybe we are all to blame for this. to be sure it's probably mainly intel, and some measure of board. I mean those defaults are strange. This is what happens when everyone keeps stripping away every last bit of buffer. There was a time 15 years ago where cpus had a 20% buffer. It was actually pretty nice. That was a bigger process node too which are way less susceptible to electromigration. I m not even convinced we are entirely mitigating that. What's 3nm and smaller going to be like. This can't continue forever unless we want our stuff lasting for 2 years.
Totally unrelated to the Intel case, but I have a unique recurring crash on AMD 3900X, on a Gigabyte X570 Aorus Pro rev 1.0 with a 12+2 VRM, which I believe to be voltage related. And I've been suspecting the board. But I don't really have the voltage-expertise. So I guess posing the question here can do no harm. This crash first happened when running Starfield. Consistently, every 5 to 20 minutes of running the game. A crash with a brief moment of the sound hanging, and then a perpetual black screen, with everything including the reset button unresponsive. But turning the system off by holding the power button down was still possible. Then rebooting to a black screen, which stays black for several minutes, after which suddenly the system comes back to life again. When entering the BIOS, it says that the BIOS has been restored. In other words, this crash corrupts the BIOS, and the board flashes it with a backup on reboot, to ... reincarnate. Whatever the hell kind of game crash has the ability to corrupt the BIOS? I have been messing with PC's since the eighties, I've seen a thousand crashes or more. But I had never before seen a crash that corrupted the BIOS. Somehow something triggers the clear CMOS, or something draws so much power that it might as well be a short, dropping the voltage. Ever since my dance with Sarfield, and swearing the game off completely, my PC has a random crash every few weeks, to the same black screen with the same symptoms. I have seen it happen while idle, running just a browser. I'm thinking about waiting until a few more 9000 series CPU's are launched and tested, and the upcoming chipsets, and then swapping out the whole system. I'm running it stock, only EXPO and SMART enabled. Updating the BIOS firmware, GPU drivers and everything else made no difference.
With a good CPU there’s no need to use a new microcode. Just limiting PL1/2 and ICCMAX, and also, if you want, as TVB doesn’t work to cap turbo boost for the best to cores, one option to avoid the Intel algorithm to boost best two cores is just cap this to cores to 5.7 and those cores would be cap, and the remaining will use the Intel algorithm to boost. So, for light loads I use 57x with no crash and for Ycruncher or heavy workloads the PLs cap or downclock the frequency to 5.4MHz, however, for shader compilation or light loads, i always use 5.7 and I do pass CB15 exteeme, link pack and other test. I have a good CPU 14900K SP106, Pcores 115 and Ecores 88 and VF curve at 1.425 max voltage at 6MHz.
Problem is that I'm on an Asus TUF Gaming B760M-PLUS (non-wifi ddr5), and this sh!tty motherboard doesn't have any option to limit the voltage. It even used to have an option to set a voltage offset in the microcode 0x104 but doesn't have that option in the latest "stable" 0x129 microcode, but now the offset option is gone and it only has a manual static option to set the voltage, which feels stupid to even set the voltage to 1.4V at all times as that may do more harm than good. I never thought I'd ever do overclocking or undervolting in order to squeeze more performance out of the system but Intel degradation situation has got me wanting to do this for a completely different reason that I never even thought I'd encounter in a modern CPU nowadays.
I updated my Asus motherboard to the last bios with Intel microcode 0x129. If I use Intel Default settings, I reach 100 degrees Celsius in all games. If I use asus profile, I get 80 degrees as max, but you are telling me it disables the VID/VCORE LIMIT... Idk what to do lol, any help?
My PC used to crash. However, I solved it after getting a new case that has better airflow. My problem was temps of my boot drive. I’m someone that only likes using mini-itx cases, so temps are higher than if I used a mATX or ATX case.
If using an all core, fixed voltage manual setting (5.4p/4.2e/4.8 ring 1.32 vcore set in bios) with a standard LLC (MSI level 6 in my case, 1.248 vcore under load) rather than the adaptive/default cores/voltages and advanced tweaking with ac/dc llc, is it still required to tweak these settings shown for safety purposes?
Thank you being the only Tech tuber TMK to actually put an oscilloscope on a board and checking the best you can. everyone else just miss the spikes from what I have seen
"You didn't update and use Intel defaults? No refunds" - Intel probably
Bingo! This gives Intel a second excuse to cancel your warranty!
Oh I can help you with that! Sorry to hear about your issue. Rest assured, Intel values all our customers. Just keep trying to use our RMA service so we can ... jerk you along for 3 more months with repeated denials and delays and pray you give up.
Just lie about it, and if you feel bad about that, note that they've been lying to you for more than a year.
"Prove me i didn´t use intel defaults" - me and i hope i won´t have to do it like this...
@@The_Noticer. big corpos, including AMD, Nvidia, Apple, Google, ... (the list goes on and on) lie all the time and only backtrack, when they get caught.
I feel like Intel is just trying it's hardest to not give up on the advertised boost clock. They are trying to frame the situation as if the voltage is accidentally too high, when in reality too many samples needed the voltage for those frequencies. The degradation is probably still happening, they are just slowing the bleeding.
They just have to slow it down enough for the chips to get past the extended warranty, and for Intel to have new chips on the market to replace them. I asked at the local computer chain if people had stopped buying Intel CPUs given the reliability issues, but they said it hadn't had a major impact on sales.
@@strangestecho5088 Well most consumers are ill informed in general, but AMD has been outselling Intel in the DIY space for quite some time anyway. So it hasn't effected the people dead set on giving Intel their money.
hoping all the hard core fans buy the new CPU and move away from the current CPUs... I'd buy AMD if I wasn't already on and AMD system
@@Denbot.Gaming amd is the same these days. 9000 is near-0 progress cpus made only to make 7000 with discounts to look better then they are. because CPU sales are down and nobody needs new cpus without progress with affordable GPUs. and progress with GPUs was killed to sell more consoles chips amd makes for sony/ms.
Many countries with stronger consumer protections would be extremely upset if the boost clocks dropped, this is Intel playing CYA to avoid having to refund customers there.
So this microcode is basically "if intel defaults selected, then turn on voltage limiter, else let the cpu request anything it wants".... Got it... Someone got paid 100k+ to write this microcode 🤣🤣
Any proof of this amount?
@DrNoBrazil I think he is referring to the tech salary
I am running 13900k at 5.2ghz with AC/DC loadline 0.60 both on 0x125 microcode, Ring ratio 4.5, E-core 4ghz my CPU is sitting at 1.268-1.3v. I don't need 5.8ghz or 6ghz for game performances as most single players run on GPU and multiplayer uses CPU cause of players nearby and my core does fine at 5.2ghz and i didn't have to undervolt. P1/p2 253W and 307A ICCmax.
i would expect further microcode updates in the near future lol
it seems like theres plenty of ... room for refinement .... in the current fix
tho perhaps understandable that they would rush this out given the situation
I mean, that's kinda the point of an unlocked CPU. You can choose the manufacturer's recommended specs or go wild and tune it to your liking
Imagine how many normies will never update their bios and will just slowly degrade their chips.
True
Are normies the minority or the majority with these high end CPUs?
normies will never buy enthusiast grade cpu
Most of them I think
Normies with OEM/SI systems will get the BIOS update pushed via the system bundle-ware.
The new microcode settings don’t change much because if they send out a proper fix that limits the 6.0/6.2 boost on single core to a fixed all core boost of 5.7/5.8 which fixes all the high voltage spikes and degradation issues they are then set up for a legal challenge in court for false advertisement which leads to more issues than an extended warranty on degradation failure…
they can just argue core clocks are like ISP speeds, never hitting the "up to" number, this already has precedent with laptops. and magnuson-moss warranty claims are out of the window with extended warranties if they actually respect claims made by regular users.
@@KasasagiWad3 True, that's why the new microcode is still trash it still spikes voltage to 1.5v but only now it spikes for a lot less time so they can argue it hits advertised speeds still on 1 core and they reduced degradation hopefully enough where it doesn't completely require a recall while they screw people over denying claims for replacements due to degradation with other loopholes.
If people would just manually change the stupid 1-2 core turbo overclock in their bio and cap the voltage to 1.4x max it would fix all of the voltage spikes and degradation on their cpu's and work a lot better than these failed new bios microcode intel fixes.
Hi Buildzoid, I know you said you don't find this topic all that interesting but could you make a short video showing a few boiler plate explainations of how to properly set up "safe" settings that don't completely gimp the performance for few different Motherboard BIOS like, ASUS, MSI, etc because the BIOS names and location seem be be pretty diffent across manufacturers. This kind of info would be super helpful & appreciated by the communiuty.
I second this, I really could use a straight forward, informative guide on how to get my 14700K properly safe and setup on my msi motherboard. This is obviously a channel I can trust with my pc's important bios settings, I'm reluctant to try other channels because they do not know what they are doing half as well as Buildzoid does!
And I hate just copying someone else's settings directly and randomly implementing something as important as a bios, without actually understanding what these changes are and why we are doing them, copy pasta'ing a BIOS just isn't a good idea lol. Buildzoid would actually explain what the changes are, and which would be ideal and why.
Here's to hoping maybe he will take pity on us poor simple peasants, and help us out with a guide to setting up a safe intel cpu, without compromising performance as much as possible
Im a noob aswell with a 13600k and msi z690 and hoping for some tipps
Same. Let's upvote this comment!
Up!!!
Same. I copied his settings from his last video (high/55/55/-0.130v offset) in my Gigabyte and everything seems fine, temps are much lower, voltage is fine... but he didn't mention it to do this video if you didn't trust Intel and their stock high voltages. So yeah, I'm a bit more confused now.
Holy shit this is disappointing. What a bandaid fix. We are on our own. I'm glad I undervolted from the start.
so it seems, people have been calling the "fix" exactly that, a bandaid, a fucking bandaid to have the CPU last pass the extended warranty
@@Alonne1 It's a fucking disgrace, I'll just say it man. This should have been implemented on such a hardcore low level that whatever any user does in BIOS, it should never override this, a true fix. I wonder if they have actually done this on purpose, to just push away anyone trying to RMA later down the line but who wasn't using their intel default profile. You can't help but wonder. It's just so lazy. After all this time...
Seems we wasted our time updating to this microcode then.. as predicted. You're right, fucking shambles.
Why would you buy 400 watt 1.5v CPU without radical cooling system anyway ?
@@club4ghzCooling has nothing to do with this. Apparently the problem is oxidation of connections inside the chip, which is greatly accelerated by the voltage that Intel chose to grill them with. Thus eventually the CPU stops being stable at factory speeds.
It seems that 0x129 is just another delay tactics by Intel, unless they show technical analysis on their end that it solves the issue and show that it is the root cause of the issue
its just their way to tell people to hold their horses... by calming their systems down a bit...
Thanks for all you do. Deep content! Love from Sweden.
They made a faulty product and now they do damage control with a code to prevent their costs, not your costs where you payed for an inferior product.
It’s like you bought a car and the engine gets to much fuel and that damaged the engine and the fix the manufacturer do is lowering the fuel consumption to prevent further damage resulting in that the engine isn’t anymore as they advertised it to be with the horsepowers and performance you paid for!.
Shiiit the whole tech market seems to be massively infected with greedy a-holes that skip corners for an easy buck and you get low grade products!.
Benchmark tests show minimal impact of the new microcode on performance..
@@tmsphereMinimal impact of new code on performance with the caveat that you’ve already applied the other incremental fixes that themselves lowered performance relative to launch day / advertised.
So, not any additional performance losses, but still a net performance loss to have a chip that will degrade slower.
@@tmsphere 6% in my case is not unnoticable
intel fan girls are gonna cry now and defend this POS like it's some revolutionary CPU 😂
@@awaiswasi354it is revolutionary. It revolves in and out of the RMA queue
Incidentally raptor lake = dinosaurs. Meteor lake came when raptor lake started dying = comet exterminated the dinosaurs. Arrow lake = old arrowhead caverns post dinosaur age and the rise of manmals lol
CONSPIRACY CONFIRMED!
So marsupial lake is coming?
manmals
You know they're actual lakes too, right?
Plot them on a map and what do you see?
@We_Are_I_Am ill buy intel again if they fix this problem by the time "indoor plumbing lake" is released, based on this time line 😆
I’m really glad I downclocked my 14900kf to 5.5 ghz all cores and locked my voltage to 1.30v right off the bat. I wanted it to run cooler and the voltages didn’t seem right, despite Intel saying they were perfectly fine. Excess voltage for an extra 200mhz on all cores didn’t seem worth it.
On a side note, buildzoid reminded me of my brother when he wondered how many people just deal with it. I built a pc for my brother with a 13700k and he’s perfectly okay with dealing with random blue screen on rare occasions. He’s so stubborn -_- lol. I got him a 12900KS to replace it and he’s still like nah, it’s okay 😂.
make sure his ram oc is stable, as it'll be worse if he corrupts some important files or system
there is no need to downclock it, set sync all cores to 57 so you disable this trash 2 core 6ghz boost and your voltage worries will be gone.
If you want you can set IA VR to like 1450 to make sure the cpu doesnt get more than 1.45v ever.
Then the only thing left to do is undervolt for more performance (since youre gonna be power limited if you go by intel limits as seen by buildzoids previous video / 253w = 40.4k~ cbr23)
Bro at 1.3v you should be stable at 5.7 all core... Currently my 14900k is running 5.8 all p-cores, 4.4 e-cores at an adaptive 1.29v max voltage 100% stable but you need to run manual settings with offsets and correct ac/dc load line settings...
@@joeysuzyjozy same here, certainly not the best sample but it does 58/44/50 at 1.31v while being under 70 degrees at all times
@@joeysuzyjozy yeah, I’m only using manual voltage and set a specific llc. I used to run it at 1.260v but, I recently started playing the first descendant and the cpu kept crashing. It could pass stress tests of all sorts at 1.26v but that game won’t run stable. I don’t play a lot of games and this game is an Unreal 5 game, an engine 13th/14th gen seems to struggle with according to reports I’ve seen. I had to bump it up to 1.30v to keep it stable. Hasn’t crashed since I did. Good to hear you have a good chip.
There definitely are a lot of people blaming game studios or Windows for their hardware-related instability.
My friend's 14900kf crashes for no reason during gaming. He has since purchased a 360mm aio for it, RMA'ed his 4070 Super and got a replacement and crashes still happens.
I told him about Intel cpus burning itself to death, he did not believe me. He said all the techtubers videos say Intel cpu having issue are all paid by AMD. He's a proud Intel owner for 15 years and never have issues.
He was going to RMA his mobo, I told him to humor me and let me flash his mobo first with 0x129, ofc him as a true Intel fanboi said flashing bios will break mobo... I told him you are going to RMA it anyways right so why do you care? so he let me do it. After that he's been gaming no issue for about 16 hours straight. And today he told me... see I told you all the techtubers are just paid by AMD, simple BIOS flash fixed all his issues, so Intel is still the king of gaming cpus...
Why is he the way he is.
somehow it made me laugh a lot haha
For some people, you have to let them learn their lesson the hard way.
Your friend is both and idiot and somehow a genius.
Just RMA the damn CPU
Haha wtf... Intel must read this
Another tribal consumer for the meat grinder. At that point there's no point in trying to convince him he's wrong.
I have tested the microcode on a Z790 AE AX -W and the Intel fix actually decreases performance by around 2000 points in R23 compared to the previous BIOS with 253W/280W/350A/1.5V limits. For some reason, the Intel Extreme setting limits the power to 253W/253W. Even when increasing the limits to those specified before, the score doesn't increase by much, plus the voltages are clearly higher than in the previous BIOS (although under 1.55V). At this point I am satisfied with turning off the Intel Extreme setting and manually setting the limits to the 253W/280W/350A/1.5V (although the Gigabyte Optimized setting will clearly undervolt the CPU).
I asked you yesterday for this and you made the video.
Thank you for the video bro
11:50 I'm a maintainer of some popular Binding of Isaac mods on steam, and I've had at least one person contact me because they bought a new PC and their game was crashing on random mods including mine, and it ended up being this exact issue. There's definitely a lot of people who don't realize that this is happening, and I don't like that I now have to include **out-of-the-box hardware issues** as part of my troubleshooting steps
Go no further, I have a friend that works in a retailer and he had no fucking clue this was a thing, and went as far as tell me no one is complaining so it's not a problem.
Edit. I had to educate him on the spot.
@@Joao_Dario The thing is... in raw numbers it is a lot of people experiencing this issue, as many CPUs are being sold. When you would look at percentages though, the number is relatively low. There is no exact number known but based on reported return rates, one can estimate we are talking about 5-6%. So 94% of people buying these CPUs will in fact not run into this issue or are unaware. So yes, from a retailer point of view this might not even hit their radar. That being said, 5% is an unacceptable high rate of course, plus it is only an estimation and most likely a low estimate as 14th gen is relative new (so limited amount of time to get affected) AND some affected people might not even be aware.
You described me perfectly. Early days into the 14900k issues, I replaced basically every piece of hardware in my PC thinking it was to blame for game crashes. I bet there are a TON of people out there like me.
The bios is still in beta right maybe you could submit feedback before the final release.
Makes me glad I only went with a 13600k this time around. Max voltages I see at stock settings are 1.27 @ 5.1 all core, easily handles 5.5 @ 1.31v.
Not touching these bios updates with a 10 foot pole. They say some of the i5s could be affected, but I just run at stock speeds most of the time with a -75mv undervolt and have been since late 2022. If they don't get this sorted out by 15th or 16th gen, I may hop to team red for my next upgrade cycle by then.
Same, been running 13600k for a year now with a small oc and no issues, tho I am running undervolt.
I built a 14600k for a client, I typically run a week long stress test which initially passed with flying colors, XMP enabled and everything, even manually configured intel default power limits vs the terrible ASUS defaults. Fast forward only a few weeks later, the computer wasn't even being used but just powered on, and started getting random crashes. Now the system is unstable with any XMP profile, had to reduce ram to 4400, which is "within intel specs" so it won't be RMA'd even though it's significantly degraded already.
It may be the initial week long stress test responsible for increased rate of degradation, but I've been doing this for 20 years and have always tested systems extensively without any issues. For my sanity I'm just going to swap to a 12700k and eat the cost on the clients behalf. I'll now be exclusively using Ryzen or EPYC 4004 for high perf client systems.
These modern Intel systems are just a headache.
When defaults won't even work, it's a failed product.
Good idea. I have a 13600k with no issues but I just updated the bios and then my VID went up to like 1.42v and my temps 95c. I turned on intel defaults and same deal. I ended up having to use CPU Lite load to bring everything back down. I should have stayed away.
@@rover1374 depending on your motherboard manufacturer, you may be able to downgrade the bios to the previously installed version. With my MSI board I was able to re-flash the bios I had previously (version "v1F" from november 2023).
@@terminalfx thanks, I got it going ok using lite cpu mode so I might just leave it until next time.
11:48 At that point you started describing a phenomenon I noticed a long time ago in GPU land: When the game crashes Radeon users think it's the Radeon drivers. When Windows BSODs or spontaneously resets, likewise. Guess what Nvidia users think no matter what the Windows Event Log says? So now I wonder: Will an "Intel + Nvidia diehard" rather blame Intel or Nvidia in this case? :P
Reminds me of some forum post i saw of people complaining how their game runs way worse once they ve upgraded from NV to AMD GPU, and naturally they blame the radeon drivers. It turns out that Nvidias compiled shaders (who can even survive DDU) caused issues and game reinstall is required to get rid of them
neither, it is the fault of stupid stupid devs.
It's still insane to me people consider occasional crashes normal. If my PC crashed I'd investigate the cause immediately, luckily it hasn't happened on my Linux daily driver in years now. The only crashes I got were when I started games with anticheat on my Windows dual boot.
@@PlayJasch I'm a software developer myself. Devs aren't stupid. Management is stupid. They think code output can be measured the same way you can measure how many milk cartons a machine can fill per hour. Fortunately in my line of work the wrong people die if the software isn't up to scratch, and the budgets are there to make it happen. Doesn't mean we don't have to give management push-back every so often, though. But it's not like with consumer crap where every cost saving measure must be used to make the share holders happy.
@@ThunderingRoar I found 99% of the time it's user error. The rest genuine driver problem or some software\hw incompatibility\edge case.
Funny thing: ASUS beta BIOSes "forgot" the Intel profiles performance/extreme are both set to ICC Unlimited = ON / ICC = 501A / P1/P2 = 4049W, also some board types from ASUS had AC at 1.1 again.
At this point, you can basically trust no one anymore to handle your hardware right, you have to check every setting.
Gigabyte too it was at 4049W. I set both to 253W
Great video as always. You forgot to mention how Intel’s ILM design has been bending CPUs over time ever since gen 12, that we’re forced to buy Contact Frames to protect them from bending.
As soon as I select the Intel Baseline Defaults on my asrock z790 steel legend wifi, my i9-13900k requests 1,55v VID. With the same settings, selecting ASRocks 'performance' baseline, it requests no more than 1,36v VID. This behavior changed from 0x125 to 0x129 as previously it didnt change much between intel baseline and asrocks performance baseline before (slight VID up with intel baseline on 0x125 to 1,40-1,45v VID).
@PrismVuxOFL "128GB DDR5 Installed 4000Mhz (For some reason?)" if all four ddr slots taken, XMP can be activated/or clocks raised, if you raise VDD and VDDQ voltages to DRAM mnf specs. Also unsync all banks and set voltages from mnf there as well.
Brother is 21 years in the future with that system date
i told intel if they don't refund me tomorrow, I'm opening an arbitration case with them with the AAA. I've done it with Dell before. Dell gave me my money back immediately after starting the case, and Intel will too, because it costs them more than $2000 just to open the case with AAA, it just makes more sense to give me my $552 back.
I locked my vcore down after a day of noticing ny 13700k running hot and well over 1.4v. Looks like this microcode is a nothingburger for those in the know
Have a water cooled system. i7 13700k CPU, a contact frame with an Asrock Z790 Steel Legend Wifi board. Before the intel bios update, it would instantly thermal throttle at 100°C whether Intel's limit was activated or not. After the bios with microcode update, it preformed EXACTLY the same when using Cinebench R23 200. I was ready to give up until visited Intel's site to see their power default and manually set them. 125°C base and 253°C Turbo. Lord knows what Asrock sets it to despite what settings you pick. It for the first time actually doesn't thermal throttle. The Asrock Microcode update was a complete failure and waste of time from my experience unless it does something else that I am unaware of.
most people still dont know that, 13th/14th gen i7/i9 run super hot... so hot that it was practically impossible to cool them with ANY kind of cooler available in the market...(no matter you do the microcode update or not) untill you actually put a power limit or undervolt and bring down the clocks. I realised this before getting my 14700k and got an air cooler. after initial tests with 310watts, i knew it would throttle, now i put limit to 235watts, and temps remain stable at 90deg C.
What you experienced after the BIOS update is normal. Intel actually further brings up CPU voltage in R23 benchmark
Glad there's a finally a working fix for people's whose cpu are still in relatively good shape. I wonder if they've discretely rolled this out to laptops with 13980HXs, which even though power limited on wattage, shouldn't be voltage limited any different than a desktop 13900K.
They should in fact aim for a completely different working point on the power-performance curve for the sake of efficiency. Power consumption scales linear with frequency and square with voltage, so any little bit they can get on driving the voltage lower helps their typically horrendous thermal situation and marketable battery run-time.
I can't say whether they do or do not but it would make sense that they would, and that they would get binned correspondingly for better phase margin at low voltage, and laptop chips sell at a premium. It would also make sense then if true that intel wouldn't be concerned about early laptop failures.
14900KS here. Update from maybe a week ago. Went from A-OK to things crashing often (mostly certain games; essentially unplayable now) in the span of, well, that week or so. Unfortunate. I guess some of the more spread out game and app crashes (once a week, once every other week) were early warning signs after all.
idk who'd buy intel rn knowing everything is happening with
@@Pacho18 Agreed. Unfortunately I've had this CPU since it released; it was a drop-in replacement for my 12900 (non-K).
@@soulofjacobeh oh, i understood wrongly you bought it past week. Sorry dude. I'm in the same situation with a 14900K. But i could stabilize it by lowering down voltages and freqs. Problem for me (blame myself for it) is i bought it 2nd hand, no invoice, even the previous owner. So i'm truly fucked if it keeps degrading.
@@Pacho18 I've had some RMAs that didn't require proof of purchase. Intel might be lenient, so it's worth a try.
And yeah in my original comment I was (vaguely, sorry) referring to a comment I'd made on a prior video about a week ago in which I'd thought I'd avoided the degradation issues somehow. But the last 100hrs or so of runtime have rapidly deteriorated in application stability. I now have to use Intel Failsafe SVID
Disabling Enhanced TVB also removes the 1.55v limit
thanks for the info
I am forever glad that when I decided to upgrade my 5950X to a newer, gaming-focused setup, that I went with the 7800X3D instead of going with the more expensive Intel route, cuz holy crap would I have been pissed with all this going on if I'd paid ~$600+ for that CPU.
Yep, lots of Intel shills are pissed there $600+ CPU kills itself on arrival basically. Even coping by saying "just rma it haha" until Intel rejects your RMA lmaooo
How ironic. I'm currently building an "open box" intel system with a 13700k, and to get anything that would even come close to competing I would be spending double for AMD. I don't have any brand loyalty and I was really trying to put an AMD system together this time. Not even with returned/open box products could I come close to my 13700k ($235) and Asus strix z790i ($213). Open box but unused. I really tried though.
Edit:
Even if I had purchased that CPU before, I don't think I would be pissed. I've been using Intel products for two decades and they've never let me down. I knew they would have fixed the situation and make good with their customers. Which they have, with fixes (still in beta) and a five year warranty. 👍
@@DingleBerryschnapps Nobody is talking about "open box" CPUs here. At one point I used Intel CPUs too, hell most of us did...because they were better however things change and AMD makes the best gaming CPU on the market right now.
@@DingleBerryschnappswhat are they going to do with all the small custom PC businesses that are going to have handle hundreds/thousands of CPUs that now need to be replaced or at least have the board BIOS's updated then tested? Is Intel going to pay for the shipping of entire PCs, the extra time and labor?
I flip PCs but thankfully never sold any with 13th or 14th gen Intel units (AM4 and 11-12th gen, mostly). You bet I'd be freaking out if I had sold ANY of those cursed chips to a customer.
@@smugmode
Have you seen PugetSystems article with 11th gen failure rate?
It's kinda not that big of a big deal, but at the same time, the 13th/14th gen customers are prompted to seek instability because of the articles, so there definitely are more false-positives, or otherwise undetected instability issues.
I think it means for the future, you should be sticking to CPU's manufacturer spec closely and disable MCE etc.
"everything is unstable"
Might be a particular extra layer of fun if you are using one of those new cheaper DRAMless SSDs with Host Buffer Memory and that has to use your CPU as its memory controller.
I updated my BIOS on my Aorus z790 pro x to the 0x129 and my 14700k thermal throtled instantly on the XTU stress test.
I am not good with BIOS settings unless I have a damn good guide (meaning I will find where to change something by following a guide but to find the setting on my own is something Im not good at and frankly a bit scared to mess around with)
So before 0x129 I used XTU and set my core V to 1.260 and my TB's to 280W (custom loop).
I have max 1.37v spikes with about a 1.27v average running stress tests like XTU and CB R23. All cores at 5.5/4.3 and max temp on CB R23 is 93C.
I would love to set this in BIOS but I cant figure out how because I get to many volts and 0 stability (due to my lack of BIOS knowlage) and also XTU likes to reset to default sometimes hence why I would like to have this BIOS side.
Also is it just my MB or leaving the Intel Preformance setting on wont let you undervolt even if you disable it BIOS side?
I'm just confused with all this situation. So if Intel's defaults are off, cpu doesn't get enough voltage on load and so some workloads will crash, but if you turn them on with the new microcode it just limits max vcore because before and without it was too much.
I'm just lost.
Intel is trying to burn new code into cpus so motherboards no longer feed the cpu - voltage above 1.45v or so. When the board is at a certain voltage and Intel's boost requests the next high-level voltage, the cpu may get over voltage'd by the mainboard when it's is so close. Its a delicate balance that is required to be adjusted in two different ways because that is how the components are designed, there is just a lot of interplay going on between them. This is not an easy problem to fix with the devices in peoples hands. Cya!
Also add the ILM bending to the list of LGA1700 issues, and according to Noctua the bending gets worse over time
And they removed AVX512 from 13/14 gen, what a shitshow
Still fine for 99% of the people out there... Not every kid wants to play RPCS3... Stop the drama. If you are not happy, switch to AMD, problem solved and no more whining.
@@DrNoBrazil 13th/14th gen Intel cpu even without AVX512 is still faster than other cpus with AVX 512 in RCPS3
@@DrNoBrazil congrats you got an inferior nerfed cpu with fewer instruction sets, and for no good reason other than corporate greed
@@ThunderingRoar It would be a nice gesture from Intel to re-enable those instructions. At least... But I doubt it.
@@FEGTTTSDH They can't, its fused off in all but the early batches of Alder Lake. Also since the e-cores don't have AVX-512, they must be turned off to use AVX-512 on the P-cores.
I got back my RMA'd 13th gen (13700k) a few days ago and its been night and day. I can't believe I waited this long for an RMA - I purchased the original CPU Jan 2023 and had it undervolted the entire time (CPU Lite Load 1, MSI motherboard) because I heard that it runs hot and I prioritize heat/noise. Instability occurred over time but I thought it was a problem with my wifi, a problem with my games, windows, etc. Windows wasn't crashing every day, but something would freeze/crash at least once a day. My computer effectively runs much faster and smoother now. It's crazy how much tolerance retail has - maybe because this is my only desktop. If I was working between two desktops with different CPUs, I probably would have complained earlier.
Imagine studios buying at scale
1.55v is low enough to have the CPUs last ~25 months.
I power limit my 13700K to 95W, and it hits 1.1V under load. Still perfect for my Civilization II games 🤣 It's actually quite efficient when you don't drive it hard. (Seriously, it does suck for people who bought it and need the full power of these chips)
@@peterwstacey I used to undervolt, until Asrock completely broke the Microcode update, even updating to a more recent BIOS and selecting the bios that would let you undervolt... nope... Its funny how when you set 1.3 as max in Bios, you boot into Windows, and see 1.57 LMAO
Intel is fucked. No more Intel for me after this. 620 euro CPU thats fucked...
You probably need to disable CEP.
@@Yes-sw8gh Totally understand the frustration. I have an insanely long platform refresh cycle (P4 -> Q6600 -> 4770K -> 13700K) so I'll see how Intel looks in 15 years. If I need to get a new platform now, I'd definitely go AMD 7800 non-x. Hope you get yours sorted.
Your modern high performance CPU can run a game from 28 years ago while undervolted, not sure what point you're trying to prove here lol@@peterwstacey
Yeah thanks for the bios update, 25k in cinebench r23 and temps under 90 degrees, got everything back to normal myself. 14900K
Same
My is 15k r23 on i9 13900k with (max 1.6v)
i9-13900K should be around 38k to 41k in r23.
@@vbozenko i know
@@vbozenko it was 39k before the BIOS update
I have a 13600KF on a "low-rent" Gigabyte B760 DSH3 DDR5 board with 6400Mhz RAM. I have never had any issues (apart from high VRM temps which were ameliorated by a BIOS update back in the mists of time last year). I did however, for the sake of peace of mind, update my mobo to the latest 0x129 bios for my board.
I have not experienced any slowdowns, my CBr23 scores remain at around 23K. What I *HAVE* noticed though is that total package power now never exceeds 155W, whereas before it would routinely hit the 181W limit. Maximum VIDs have reduced by around 30-40mV (averaging just over 1.33V on the P-cores now) as far as I can tell. VRM temps have dropped even further to max out at around 70'C. I generally leave everything on AUTO in the bios except for enabling XMP and rebar and what have you - I'm not interested in overclocking it.
My bump of trouble tells me that most people running B-series chipsets won't experience issues even without an 0x129 update (even with 14900Ks) but have no data to back this suspicion up. B-Series is just pure bondage and discipline for K-series SKUs, although it seems it _had_ been supplying more power than my CPU actually needed at full bore.
you look so good!
I was under the impression that the microcode update applied to non-K processors. I mean when I'm reading the BIOS updates from MSI or ASUS, it seems to be non-K processors. I haven't seen it confirmed for both k and on k
even non-K versions of cpus run above 65watts, so i assume its for EVERYONE.
intel failing to take true accountability for this means i may very likely never buy from their cpu division again
14900k sufferer here. One question that I have is if manually setting that voltage limit to 1.55 makes it safe to turn off the intel defaults?
I'm a big tinkerer, and I'd love to be able to overclock and tinker with my chip. I want to stray away from the Intel defaults in some ways (e.g., I am considering removing wattage limits). Is setting that voltage limit manually going to work equivalent to the Intel defaults (except for where I diverge manually) in terms of chip safety?
Thanks for the video!
But why didn't I try tuned IA AC/DC Loadline (for example 60/110) on Gigabyte default profile?
Also, the "CPU Internal AC/DC load line" enabled on the "Perfomance" on the Gigabyte default profile can help with instability.
So, by saying to not turn off intel default settings, is he saying to select performance or extreme but not disabled?
Yes. Disabled means full manual (or MB juicing) control, with all that implies
Correct
It's strange how some corporations are able to hypnotize customers to the degree, that they buy a product that is advertised for its high performance 6GHz and its reliability and them downclocking it to even reach that reliability and being just fine with lower performance despite them having paid that premium price in the first place. No wonder corporations keep doing what they do.
1:17 Not neccessarily. It would avoid an overwhelming of the replacement process since it is draged out over time.
The CPUs cant maintain or boost to 5.8-6Ghz without requesting 1.4v or higher... and the budget silicon cant handle those voltages
wow thats funny i seen your gamers nexus load line calibration video. will the next video pretty much be an updated version?
I have a question for the AC/DC load line video: Is AC load line measured on a specified frequency or manufacturers tune it for their VRM controller? Simply put PCB traces have different caracteristics on different frequencies, so for example a motherboard have 0,400mR at 100kHz and 0,442mR az 400kHz. Some cheap motherboards have a fixed 200kHz freq, other more expensive ones have a range 200-600kHz.
What is the minimum specified VRM response time for AMD and intel? I mean the time between when the CPU request a Voltage and when it gets the requested Voltage.
I personally turned on intel default settings with the microcode update plus turned on XMP 6000 on the RAM. I have high bandwidth/low latency disabled as well. Before all this information came out I was just running whatever profile was on it, think it was the gigabyte optimization setting then turned off multi core performance. This PC was built only like 9 months ago, i have a 14700k
care to explain how to disable "high bandwidth/low latency"? LOL
@@brunodinis7454 On gigabyte boards you can disable those two features lol. Its gigabyte exclusive so no other board has it
@@brunodinis7454 what ? lol - on gigabyte boards those are two features that can be disabled...
I can say that my old Z690 from Asrock was giving me very reasonable values.
It died and I had to switch to a ROG B760 and the values were completely bonkers.
I tried the X129 and the problem I have is that my 13600k still hits 100C° and keeps thermal throttling. I have a quite decent Cooler master twin Fan waterblock (getting hot, so it works) and I already set the Long duration boost to 125W from the 181W that the Bios sets as standard. (181W for the short duration too.)
It is quite hot these days but a CPU should not hit 100C° under normal gaming load.
The ASUS advanced Bios allows to set a 90C° as a threshold , and still it never hits above 80-85C° under ASUS optimization and I often get better benchmark results, despite manually setting the CPU values to Intel recommended. The Voltage spyke often hits 1.46v through, which is quite high for a 13600K but that happens under Intel default as well.
I will try to set a limit to 1.4V and see if it's stable but I cannot have my CPU thermal throttling constantly just because the Intel default settings on x129 lacks the tools to tame the CPU at reasonable temp values. The CPU also Spyke's anyway to 1.465v for no reason while switching applications on desktop environment: I'm not impressed to say the least!
Even opening HW64 is enough to see spykes.
You are right: the platform is a mess, but it wasn't in 2022. Intel and the MB manufacturers made it like that to be competitive with The Ryzen 7xx0 platform: Intel wants it's offer to look fast and is burning our CPUs to do that.
n1 videos even if sometimes a bit longer. did u test disabled tvb/singlecoreboost to get rid of the high vcore spikes or is this a urban myth in youtube? thx have a n1 one
I think too much change in current is the biggest problem. Set static voltage and frequency.
y-cruncher has a 2nd mode that is mostly CPU intensive, not sure how to use it via bench mate.
If you start y-cruncher directly, select option 1 and then 3 for a quick test or higher numbers for longer tests.
3:40 I have a z790 Aorus Pro X and never touched my AC LL because I never had the experience to change that nor understand it. I knew I was right that the curve my motherboard was doing was not supplying enough voltage on high loads most especially multicore / more than 2 cores active loads. I never figured out how I could set the AVX offset to off instead of the default Auto that keeps changing settings depending on what my motherboard feels like. On default, it kept playing at -5 or -7 and I hated that the negative transients was caushing blue screens on my PC while I was undervolting it. I might as well as copy majority of the settings that you have on your 14900k into my 14700k and then play around with it to get stable. This band-aid fix is REALLY disappointing.
What do you do when you have an HP Workstation with a limited user controled BIOS? One thing I'm going to do is swap in a better screw down thermal clamp on the CPU instead of the default lever clamp. It may void the warranty, but at least the cooling efficiency will be a little better on the chip.
Buildzoid uploads :)
Not a one hour rant :(
Great video. Question is im on z790 aorus prox and i got the intel defult settings after i update my bios what if i go to my old update F3 ( the bios that
My pc came with ) which is the previous update from F4 that shows intel default settings is it safe to do?
Plus after updating the gigabyte control center tells me that under vault protection is enabled in cpu tap in the app I didn’t hat this notification before when i
Funny enough my CPU was spared because when i bought it 2 years ago, top youtubers called to disable multithreading enhancement which is the future that pushes CPU even further. Well, thank you guys for saving my CPUs. I did also applied a small downvolt for temps just enough not to lose performance. Checked my VID now. is below 1.270 using adaptive voltage offset of -0.100
My bios is from january and my Mobo is a Gigabyte z690 UDD4 with f28 bios, I will not update to microcode
Good stuff! Is there any way to set the voltage limit stuff manually if one turns off the Intel default settings?
Yep, check 13:05 , you can set max voltage through the IA VR Voltage Limit, though it may be named differently on other motherboards
@@Summanis the only setting i made was to limit power to 235watts after microcode update... i could not find IA VR or VCore voltage limits, which i wanted to set at 1.4 Volts, i have the same gigabyte motherboard, but slightly different interface, its all so confusing...
@@realracing3specter295 For real. I try searching up what all the settings in my MSI motherboard do and half the time nothing related comes up. I still have no idea what "memory extension mode" is. I haven't even checked to see if my motherboard has a similar voltage limit (12th gen) but I'm sure it would turn into a rabbit hole.
I set mine to 1.45 max. It will hit a 40300 on cinebench with that limit. Now that is higher wattage but voltage limited. Intel defaults are down from 38500 to 37700.
I updated to the beta bios with the new microcode on my 14700K
intel performance (Asus doesnt give me an extreme option). ICCmax to 400A, PL1/2 to 320W, IA volt limit to 1.5V, AC/DC loadline to 0.5/0.5. And the VRM loadline to the recommended 4. As is, that runs CBr15.
Overclocked the cpu to 59x and now looking into undervolting the VF curve...
Seems that this chip runs 59x on 1.44V, but is capable of a -0.05V undervolt. Which is a worse result than before, where it could handle a -0.08V undervolt, so the new microcode is doing something.
Checking the next VF point: 56x runs with a -0.20V undervolt around the 1.22V.
Work in progress....
Still lots more testing needed, but the chip still seems to get way more voltage than it actually needs on the new microcode....
Enable 'Enhanced Thermal Velocity Boost' or in a more techy name 'VMaxStress': is just a voltage reliability on/off switch. Optimized AIB settings always disable it. Also for XTU users: touching anything in XTU to overclock will disable the feature too.
Intel's fault(and the reason they are lax with warranty) was making a mistake in their voltage reliability feature which was allowing the CPU to run at unrealiable voltages: that's what 0x129 microcode fixes.
AIB do know how to properly configure Intel CPU: my ASUS Laptop has AC_LL and DC_LL perfectly configured, as well as other things like TVB, PL1, PL2, ICCMax and even ASUS propietary 'Regulate Frequency Above Threshold'. For Desktop they just use AC_LL for whatever their reasons are: maybe it's some social engineering for home overclockers, ease of configuration,...
Thank you for the video, really helpful. Can we run the XMP mem profile with the intel default settings on ? or is it better to keep the profile disable with the intel default on.
Also after applying the intel default settings the temps seem to be really high 80-100. If this keeps up would you reccomend doing the 1.4V limit? ( if it would help)
I think I'll stick with my March 2023 MSI BIOS LMFAO. I got 1,35V max, load voltage of 1,22V set manually. I dont need intel telling me shit.
People just all core overclock your CPU -> fixed Pcore, Ecore and Vcore. No troubles then.
Or turn indeed the IA VR voltage limit setting on 1.40v (value is 1400 in BIOS). Maybe KS owners do put it at 1.425v (1425), that is stil safe.
Or put fixed max on your Pcores and E cores, then the VID never requests high values, so P core 5500mhz, E core 4200mhz or so.
Good luck it is so easy to solve. Don't do BIOS upgrades, all not needed.
I wonder if there's a specific setting that enforces the 1.55 voltage limit, or if it's really a hidden flag.
So far I was thinking that the Intel Defaults would just be a preset for the available BIOS settings, but they might actually do more, like setting such kind of hidden flags. 🤔
Intel had QA problems way before. This is why Apple broke up with them and pursued designing their own chips with TSMC.
Apple is smart now that u bringed this up now I understand why apple design their own chipsets now, and why consoles use AMD
id really like to learn more about how vid requests are generated
cause it really does seem to me like the microcode fix doesnt get to the root cause here
rather it's the only feasible fix for cpus that have already been manufactured
I just got a computer with the i9 13900k in it and learned about these issues. I did my gigabyte motherboard bios update. Some people told me to do the gigabyte default clock. Do you agree to do that and will loading the defaults hurt anything or delete any data/apps?
I'm about to give up on computers and become Amish instead. It'd be easier.
The GOPsters are making moves to have off-grid people pay for "lost utility company income" in some districts. You'd better churn a lot of butter, Eli!
@@LilMissMurder3409/videos I'm not even American so it doesn't really affect me, but I wouldn't vote for the Democrat scum even if they paid me like they do with everyone else.
Not that I think any of this is relevant to Intel being a terrible company.
Considering I spent 4 grand on my computer and another 4 grand on my sound system for my computer, it would certainly be cheaper to be Amish. No BIOS headaches.
So, if my gigabyte bios has mce on with intel default, just leave it?
no, disable it
I think you are probing the correct way, 50 ohm termination resistor and a coax cable with the difference in shield and signal wires cut as short as possible and ground point soldered as close to the signal wire as possible to decrease the inductive loop. If you find a voltage probe point on the CPU substrate consider using channels 1 and 3 to prevent multiplexing of the ADC's if it has two, otherwise ignore my advise, my own scope is pedestrian at best.
Also YES some explainers on the various BIOS settings and OC flowcharts are very welcome.
No way 1.55 doesn't still cause degradation. Maybe they'll also limit how often you can have your advertised boost clocks and most chips will just about about make it through the warranty?
If I leave intel default enabled my 13600k runs 1.45vcore in Cinebench r23 multicore. and instantly hits 100c+
If I change svid behaviour to typical it lowers vcore to a normal-ish 1.25ish but my r23 multicore score is 2000 less than normal for a 13600.
This is on a 690-A from Asus.
People just all core overclock your CPU -> fixed Pcore, Ecore and Vcore. No troubles then.
Or turn indeed the IA VR voltage limit setting on 1.40v (value is 1400 in BIOS). Maybe KS owners do put it at 1.425v (1425), that is stil safe.
Or put fixed max on your Pcores and E cores, then the VID never requests high values, so P core 5500mhz, E core 4200mhz or so.
Good luck it is so easy to solve. Don't do BIOS upgrades, all not needed.
Same deal with me on 13600k/msi b660 baked. 1.42v and 95+ temps. I enabled ‘normal’ mode and then chose the cpu lite mode setting 6. Temps back down to 75c and at 1.30v. Cinebench score is about 22600, so pretty close to the standard score.
What a stupid fix. I wasn’t having any issues with my 13600k but decided to update the bios for the ‘fix’. The fix would have cooked my CPU.
I had to return my 14700k with the performance degrading by the day for a 14900k. (I know risky.) - But so far, before and after the microcode this thing runs 15c cooler than my 14700k at the same intel default settings that the 14700k was overheating with. With the same exact set up, the 14900k ran cool and no issues so far in comparison to the 14700k was thermal throttling and at times, unstable.
Also.. I've been enjoying your videos and here is my sub!
How long did you have your cpu ? i have a 17 14700f and mine didnt blue screen thankfully i updated my bio’s yesterday i had the cpu for about 6months
@@nightwing8794 In short, I had bought mine at Microcenter on Black Friday it came with the MSI Z790 bundle. About less than a year, approximately around 6 months.
@@nightwing8794 To add to this comment, what's weird is no matter what, it was overheating with even undervolting / underclocking. We thought initially it might be the aio, but like funny enough none of these issues have appeared in the 14900k yet, it's been rock solid thank god.
Hey mate. Thanks for all these videos and tests. Right now, i have 14900k-msi z790 with latest 0x129 non beta update. Im using it with lite load 9 and ia cep disabled because it comes with ll mode 16 and that causing idle voltages being too high. Like 1.3-1.3 vcore and 1.48 vid. And if i dont turn off cep and lower ll mode to 11 or lower, cep kicks in and cpu performance getting capper from %20 to %50.
So i wonder, if i dont change any settings from default intel profile but only change ll mode 9 and disable ia cep, would that also cancel the microcode update? I dont gent voltage spikes in hw info but there still could be transient spikes.
Can you test? Thanks.
What's the default IA VR Voltage Limit on your Gigabyte motherboard?
AC and DD LL explanations would be awesome (and for Asus motherboards too).
Nice one as usual. BTW, with "Renesas", the emphasis is on the first syllable "REnesas", not "reNAYsas" 😉 Source: "We are Renesas: Company Overview"
Does this mean that the VID-Limit will also be disabled when I change the AC/DC LL or set a Vcore offset when using the "Intel default settings"? That would be a bummer. Want a little UV but also the Vcore Limit.
Great video, I'm tempted to set up my Tektronix oscope then capture voltages on my MSI motherboard. I'd love to see what's happening in real time although I'm uncertain where to pick up the signal. CPU power plane is easy to find but would that be the correct voltage. Hmm...
So for my undervolteted 13700k its more save to keep my undervolt settings ? Never hit over 1.3 Volts. I did -95mV all cores expact the 2 Fast cores. They got just -65mV to get the 5.7 GHz done.
haha the date on the BIOS Buildzoid is a time traveller? :P
13900K/MSI Z790-A wifi bought in August 23, since the very beginning I ran adaptive undervolt of -0.035V on frequencies above 5GHz, with the MSI default profile for water cooler (i.e. no other limits). Never experienced any crashes, passed all the usual stress tests like Cinebench, prime95 etc. Just ran Cinebench multicore, 40293 and max VID during the test in HWinfo was 1.42, with averages around 1.33V. Don't think I'll be updating to 0x129.
12th gen could have become a legendary cpu with > 5GHz out of the box, AVX512 and BCLK OC. And then Intel came and did Intel stuff.
I love actually hard core overcloking. it's a lot of fun to watch.
intel, historically i've liked your cpus as well as amds. But honest guys, while i like watching this, i don't want to live it. What i'd like is to forget about voltage and use my computer care free. I'll wait for the update and just use it i guess.
I mean for now i'm running at 90w, and 1.05/-.07v (adaptive offset, because i like cool and quiet anyway 52/22 multipler. I like to set these settings in the board though and don't want to deal with this problem.
BTW i can't tell the difference from the voltage melting settings, and the computer run quiet with no heat. Food for thought. Maybe we've all gotten too focused on benchmarks. Amd is paying that price now.
maybe we are all to blame for this. to be sure it's probably mainly intel, and some measure of board. I mean those defaults are strange. This is what happens when everyone keeps stripping away every last bit of buffer. There was a time 15 years ago where cpus had a 20% buffer. It was actually pretty nice. That was a bigger process node too which are way less susceptible to electromigration. I m not even convinced we are entirely mitigating that. What's 3nm and smaller going to be like. This can't continue forever unless we want our stuff lasting for 2 years.
1.4V IA VR Voltage limit enabled! Thank you Buildzoid! My idiot CPU peaked at 1.5V in that bios ! (14700KF @ Z790 AORUS ELITE X WIFI7) (Bios ver. F7f)
I did 1.47 for me. Max I have was 1.45V. This is still high (especially for temps) but better than nothing, Intel screwed badly
What should be my ia vr current limit if my cpu is 13700k? Stock ghz. Thanks.
In your opinion, are the 12 GEN processors randomly subject to this problem if we are talking about BCLK OC?
12th gen has nothing to do with this, just 13th and 14th gen
I am getting noise from CPU side of B760M-k motherboard after installing last bios uptade. how can i fix that
Nice 👍
Tl;dr turning it off makes voltage go back up, towards or past unsafe levels.
Totally unrelated to the Intel case, but I have a unique recurring crash on AMD 3900X, on a Gigabyte X570 Aorus Pro rev 1.0 with a 12+2 VRM, which I believe to be voltage related. And I've been suspecting the board. But I don't really have the voltage-expertise. So I guess posing the question here can do no harm. This crash first happened when running Starfield. Consistently, every 5 to 20 minutes of running the game. A crash with a brief moment of the sound hanging, and then a perpetual black screen, with everything including the reset button unresponsive. But turning the system off by holding the power button down was still possible. Then rebooting to a black screen, which stays black for several minutes, after which suddenly the system comes back to life again. When entering the BIOS, it says that the BIOS has been restored. In other words, this crash corrupts the BIOS, and the board flashes it with a backup on reboot, to ... reincarnate. Whatever the hell kind of game crash has the ability to corrupt the BIOS? I have been messing with PC's since the eighties, I've seen a thousand crashes or more. But I had never before seen a crash that corrupted the BIOS. Somehow something triggers the clear CMOS, or something draws so much power that it might as well be a short, dropping the voltage. Ever since my dance with Sarfield, and swearing the game off completely, my PC has a random crash every few weeks, to the same black screen with the same symptoms. I have seen it happen while idle, running just a browser.
I'm thinking about waiting until a few more 9000 series CPU's are launched and tested, and the upcoming chipsets, and then swapping out the whole system.
I'm running it stock, only EXPO and SMART enabled. Updating the BIOS firmware, GPU drivers and everything else made no difference.
With a good CPU there’s no need to use a new microcode. Just limiting PL1/2 and ICCMAX, and also, if you want, as TVB doesn’t work to cap turbo boost for the best to cores, one option to avoid the Intel algorithm to boost best two cores is just cap this to cores to 5.7 and those cores would be cap, and the remaining will use the Intel algorithm to boost. So, for light loads I use 57x with no crash and for Ycruncher or heavy workloads the PLs cap or downclock the frequency to 5.4MHz, however, for shader compilation or light loads, i always use 5.7 and I do pass CB15 exteeme, link pack and other test. I have a good CPU 14900K SP106, Pcores 115 and Ecores 88 and VF curve at 1.425 max voltage at 6MHz.
Problem is that I'm on an Asus TUF Gaming B760M-PLUS (non-wifi ddr5), and this sh!tty motherboard doesn't have any option to limit the voltage. It even used to have an option to set a voltage offset in the microcode 0x104 but doesn't have that option in the latest "stable" 0x129 microcode, but now the offset option is gone and it only has a manual static option to set the voltage, which feels stupid to even set the voltage to 1.4V at all times as that may do more harm than good. I never thought I'd ever do overclocking or undervolting in order to squeeze more performance out of the system but Intel degradation situation has got me wanting to do this for a completely different reason that I never even thought I'd encounter in a modern CPU nowadays.
How can anything in the BIOS override what the microcode does? Isn't the microcode hard coded into the CPU itself?
Do if I have already undervolt with bios setting, do I still need the new microcode?
I updated my Asus motherboard to the last bios with Intel microcode 0x129.
If I use Intel Default settings, I reach 100 degrees Celsius in all games.
If I use asus profile, I get 80 degrees as max, but you are telling me it disables the VID/VCORE LIMIT...
Idk what to do lol, any help?
My PC used to crash.
However, I solved it after getting a new case that has better airflow.
My problem was temps of my boot drive. I’m someone that only likes using mini-itx cases, so temps are higher than if I used a mATX or ATX case.
If using an all core, fixed voltage manual setting (5.4p/4.2e/4.8 ring 1.32 vcore set in bios) with a standard LLC (MSI level 6 in my case, 1.248 vcore under load) rather than the adaptive/default cores/voltages and advanced tweaking with ac/dc llc, is it still required to tweak these settings shown for safety purposes?