Intel Core 13th and 14th Gen Vmin Shift Instabilty Update - New Microcode Update (0x12F)
Posted by imaginary_num6er@reddit | hardware | View on Reddit | 97 comments
bygphattyplus@reddit
New bios update? Do I need to update my i713700f? I have the x12B and haven't had any issues. This is ANOTHER fix or just an improvement?
PruneCharacter645@reddit
En la práctica, la mayoría de los equipos de escritorio y gaming no se verán afectados por la situación que soluciona 0x12F. Pero para laboratorios, servidores pequeños o cualquier configuración donde se deja la CPU inactiva con tareas ligeras durante días, este parche podría facilitar el proceso.
bygphattyplus@reddit
I had to translate, but thank you. I had a huge panic attack updating my bios last time so hearing this makes me feel much better.
PruneCharacter645@reddit
No panic attacks... the only problem is if you have power outages at home without a UPS... in that case, updating the BIOS and the power going out could mean having to replace the BIOS chip on the motherboard...
bizude@reddit
What in the heck? Are you freaking telling me that Intel is STILL finding more problems?! How the heck did they become this incompetent?!
PruneCharacter645@reddit
En la práctica, la mayoría de los equipos de escritorio y gaming no se verán afectados por la situación que soluciona 0x12F. Pero para laboratorios, servidores pequeños o cualquier configuración donde se deja la CPU inactiva con tareas ligeras durante días, este parche podría facilitar el proceso.
imaginary_num6er@reddit (OP)
Planning to redo testing again after all microcode updates?
bizude@reddit
Well, let's see if I'm still part of this industry in a few weeks. I'll consider it if I am, but no promises.
TalkingCrap69@reddit
What in the world did you do this time?
bizude@reddit
There's a vendor which has been sending me products over the past year, percentage wise has had a 75% defect rate and the last two products of theirs I tested were electrically defective and damaged motherboards as a result.
I barely managed to escape a violent gang which targeted me in Arkansas two months ago, I really don't have the financial resources to be wasting time on vendors whose QA standards are so low it's threatening my ability to be financially stable between damaging equipment I can't afford to replace and lost income from another review I can't send to Tom's Hardware
And I had an angry moment where I told them exactly how sick and tired I am of dealing with their lack of standards. If this is a reviewer's experience with sampled products, I can only imagine the situation is equally bad for consumers.
imaginary_num6er@reddit (OP)
I hope everything goes well
bizude@reddit
Yeah, so do I
But I had a few choice words with a vendor after having certain inexcusable repeat problems occur with their hardware and that pissed people off because my uncensored anger is rather raw.
BrightCandle@reddit
The product is faulty by design, they run at clock speeds needing voltages that kill the chips. The only solution is a substantial reduction of clock speed. If they do that they will be forced to refund and likely recall. So instead they are just doing what they can to string people along and reduce the failures a little and hope to ride it out. The consequences is huge on their reputation but hoped to be less than the cost of recall of two generations of CPUs.
Cheeze_It@reddit
Capitalists pushing for more profit instead of better products.
Reactor-Licker@reddit
At this point I think it’s safe to assume all Raptor Lake desktop CPUs are suspect and potentially a ticking time bomb. I remember it being pretty clear that the last patch was supposedly the “final” one that fixed everything. Arrow Lake, for all its faults, at least isn’t this unreliable (that we know of yet).
PruneCharacter645@reddit
En la práctica, la mayoría de los equipos de escritorio y gaming no se verán afectados por la situación que soluciona 0x12F. Pero para laboratorios, servidores pequeños o cualquier configuración donde se deja la CPU inactiva con tareas ligeras durante días, este parche podría facilitar el proceso.
LynxFinder8@reddit
The weird part is. I run a 14th gen undervolted and OCed. The 0x12C firmware not only improved stability but also dropped temperatures under load by a couple of degrees.
I think Intel still has room to tweak the V-F characteristics in these CPUs
hackenclaw@reddit
I wonder how many will survive after 10yrs. (yes some motherboard can survive this long, I have a Q9650 rig still working)
24core CPU might still relevant in 10yrs. Many of the HEDT Haswell-E, Broadwell-E still pretty good for most usage.
BrightCandle@reddit
This is the fifth patch to fix the problem since September. There will be more.
LuminanceGayming@reddit
so is this an indication that the previous microcode update was not sufficient to prevent vmin shift?
T0talN1njaa@reddit
It’s a mess if it hasn’t been but the message from Intel isn’t really clear it seems to tiptoe around the fix being about improving stability..
If it is in fact another fix then this would mean all the chips already rmaed with a fresh install of 0x12b could now be subject to some sort of damage without 0x12f..
Intel need to clarify further
PruneCharacter645@reddit
En la práctica, la mayoría de los equipos de escritorio y gaming no se verán afectados por la situación que soluciona 0x12F. Pero para laboratorios, servidores pequeños o cualquier configuración donde se deja la CPU inactiva con tareas ligeras durante días, este parche podría facilitar el proceso.
fafatzy@reddit
They will never clarify it… they still haven’t
T0talN1njaa@reddit
Completely agree 💯
Affectionate-Memory4@reddit
It appears to have been sufficient for the vast majority, but additional support for the platform making it even less likely are welcome.
theholylancer@reddit
Yeah but... This feels like the rrod issue where it's when not if when the things die
Sure that was due to bump gate but this still feels like they can't hit the spec without degredation long term and is just patching it until it's no longer an issue.
Perfect_Cost_8847@reddit
I agree. This looks like “managing” the issue, not fixing it. What supports this notion is that Intel couldn’t afford a mass recall, so they had no option but to proclaim they had fixed the issue, even if they hadn’t.
theholylancer@reddit
and also as an avid OCer since the days of the Athlon 64, its how heavily OCed chips die over time.
They degrade in terms of what voltages you need to hit a clock, then the ability to hit said clock at all, then maybe eventually an actual death. Unlike say the asrock X3D situation where its likely some sort of over current event that pops them, this is slow OC degradation and they needed to either redesign the thing, or admit failure and drop official spec way lower back to if not 13th gen levels but 12th gen since there are also dead 13th gen chips.
Which would piss off way more ppl
VenditatioDelendaEst@reddit
Chips degrade that way at stock too. There's just enough voltage margin that the computer is typically replaced before you run out of it. Overclocking works by trading away almost all of the margin for performance, so OC'd chips will start miscomputing well inside their useful life, even if the degradation rate isn't meaningfully higher.
(That's not to say it's not meaningfully higher; it's all current, temperature, and voltage, after all.)
theholylancer@reddit
yep, but most normal chips, IE non K, and non top end chips have plenty of margin to really run for 10 years+
Like stock 6000 series and older intel chips are in ebay listing all the time in office machines, and they are cooled by shitty ass OEM stuff, but they have so much margin it usually dont matter
its when intel was desperate to fight X3D with clockspeed when it is an issue
nero10578@reddit
It never was lol you can’t fix these except for lowering official clockspeed and admitting a huge fuckup and opening themselves to lawsuits
SkillYourself@reddit
They already did with the 1.55V limit. At stock loadlines, 14900K/14900KS will rarely hit max 2-core turbo within that VID. Temperatures and predicted current would need to be low for the peak 2-core turbo clocks to be achiveable under 1.55V VID.
nero10578@reddit
But they still are doing everything they can to not change the official specs to have lower clocks.
BrightCandle@reddit
I think this is the reality of the situation, the CPUs are being run beyond what they can handle and dying early as a result. Intel will never say this they will just keep stringing people on with microcode updates to try and reduce the failures in the usable life of the CPUs and apply the things that only impact performance a little bit.
nero10578@reddit
Exactly. The CPUs target way too high clockspeeds that requests way too high VID than the ring can handle. You can’t fix this except for officially changing the specs to lower clocks and having lower VIDs as a result. All they’re doing now is cutting into the voltage margins for running at the spec voltage and hoping it will reduce failures enough.
imaginary_num6er@reddit (OP)
Advanced-@reddit
I just bought a 14th gen CPU and put mine in 4 days ago.
I leave my PC on all the time. Will be shutting it down until Asus releases this....
PruneCharacter645@reddit
En la práctica, la mayoría de los equipos de escritorio y gaming no se verán afectados por la situación que soluciona 0x12F. Pero para laboratorios, servidores pequeños o cualquier configuración donde se deja la CPU inactiva con tareas ligeras durante días, este parche podría facilitar el proceso.
pdp10@reddit
Microcode patches are loaded at each bootup by the systemboard firmware and/or the OS. If your OS has the microcode patch, then for the time being you're not missing anything by not having an updated motherboard firmware.
BrightCandle@reddit
Anyone getting the feeling that Intel has just been gaslighting its customers for more than 2 years now? Since they started delivering patches since September every time they have said this will fix the problem and then another patch and another came out and they just keep coming.
My 13700k died a month ago, its in the process of being replaced but I already bought an AMD because its taking months.
bizude@reddit
My "spider sense" went off the hook when I heard one Intel employee describe the instability issues as like "peeling an onion, and finding more layers with every peel"
Visible_Confection12@reddit
Intel employee told you this? That means this issue is far from over.
crab_quiche@reddit
Ogre Lake
greggm2000@reddit
Onion Lake :)
SUPERSAM76@reddit
Is this something we can apply ourselves or do we have to wait for our motherboard manufacturers to spit out a BIOS update?
alfiejr23@reddit
Wait for your motherboard vendor.
seohyunfan@reddit
how do I know which microcode package update this will be in on Linux?
alfiejr23@reddit
Just check it at motherboard's bios page. Asrock already has it.
hagar-dunor@reddit
Running my 13900KS slightly underclocked but heavily undervolted since day 1. Never had a crash and still going strong. No microcode updates applied. I had my share of "You're losing so much potential bro, 1.55V is fine if Intel says so, 100°C is totally normal bro (intel subreddit)"
The fact remains that Intel massively fucked up power delivery on these generations, and you had to invest quite some effort in tuning this, which no user should be forced to. What a trainwreck.
goobdaddi@reddit
So do I have to refresh asus's bios page until there is an update or can this be installed some other way?
VenditatioDelendaEst@reddit
It may be possible for your OS to load the new microcode on boot, but IIRC Intel didn't use this mechanism for the previous updates for some reason. It's possible that whatever parameters are involved in this issue are locked before control is handed off to the OS kernel.
pdp10@reddit
Does the Lnux redistributable microcode package have the previous microcode patch, do you know?
VenditatioDelendaEst@reddit
If I'm reading this right, it's up to 0x38 released on Feb 11. That references this spec update. AFAICT, "v_min shift" is erratum RPL063, listed as "no fix" and "Workaround: It may be possible for the BIOS to contain a mitigation for this erratum."
Cradenz@reddit
the last time intel rolled out a cpu microcode update through windows update it caused A LOT of problems. a lot of blue screens. They actually did when 13th gen released. They never did it again.
l1qq@reddit
I'm so done with Intel after this, lol...nVidia looks to get next on my chopping block.
d1stor7ed@reddit
I switched to team red after I RMAd my 4th 14th gen.
l1qq@reddit
yup, my next personal build will be an x3d and I just finished builds for 2 others last week using AM5.
Myrang3r@reddit
I would get one, but AMD also joined the greed train. From a 350€ 7800x3d to now asking 550+ for the 9800, yeah, no thanks…
greggm2000@reddit
Intel and AMD and Nvidia have always been on the “greed train”, companies charge what they think consumers will tolerate, bc that’s how markets work.
If you aren’t going to get Intel, and you aren’t going to get AMD, what are you going to get?
trololololo2137@reddit
intel 265k is now $300 and is way faster than 9800x3d for everything except gaming
greggm2000@reddit
But it's a lot slower for gaming, and if your workload is non-gaming, you'd choose a different AMD CPU than the 9800X3D. As well, sure the 265K is $300 at the moment, but that ignores the cost of the motherboard, and it's also a non-upgradeable platform.
In short, I don't see where the 265K makes sense for most people.
ezkeles@reddit
this is why i wish china succeed make their own cpu and gpu
anything they touch uis cheap for long time
greggm2000@reddit
CPUs and GPUs are mostly already made in China (including Taiwan).
Now, if you mean desktop/laptop CPUs and GPUs that are totally designed and made in Mainland China, then those exist, but they're not nearly as performant, and then on top of that, you have the software stack, as an issue.
SimpleNovelty@reddit
Seriously, people think these companies are non profits or something? They're going to try to maximize their profits. Pray you don't get priced out or buy used.
greggm2000@reddit
Indeed. And then on top of that, there's the tariffs if you're in the US, which is going to make things much worse if you live there.. as if we wanted higher prices.
vg_vassilev@reddit
RemindMe! 4 months
Jayram2000@reddit
4 is crazy, were they early i9s?
Unironickek2@reddit
He's making shit up
LuluButt3rs@reddit
Dont believe what everyone says online
Aggravating_Ring_714@reddit
Be sure to switch to the fantastic 9800x3d on an epic asrock mainboard. Zero issues guaranteed 😂
Eat-my-entire-asshol@reddit
Or just don’t get an asrock board for now until they figure it out?
Sleepyjo2@reddit
Or a MSI board, or an ASUS. Or really just skip the chip if you’re worried about it. 7800x3d is more than fine and quite a bit cheaper anyway.
I do have a 98 on an asrock board, which does instill some worry but I’ve also been running it at jedec speeds (intentionally, I know) with the lowest power draw I can manage since I bought it so I’m hoping it’ll be alright until I get rid of it.
(The same failures have been reported from multiple brands, just less frequently. Reddit is, unfortunately, not a good way to determine failure rates but we have no other info.)
TenshiBR@reddit
yikes, I am staying with my 78 for now I guess
98 price is plummeting and I was getting tempted
Eat-my-entire-asshol@reddit
Huh? Im on an asus board since november and overclocked without issues and a 9800x3d.
Multiple friends on msi and asus boards since november no issue
Meanwhile my brothers 13700k is crashing, my 13900ks now needs 1.394v to run stock speeds and it needed much less when i first got it
Sleepyjo2@reddit
I don't know why you brought up 13th gen, I didn't say anything about Intel.
I also don't know why you brought up you and your friends. I'm on ASRock since launch and also fine, that means literally nothing. All I said was that there are multiple posts about people with non-ASRock boards also having failures within the last month or two, I even pointed out that ASRock has more frequently reported issues so you're basically arguing what I already said.
Reddit posts are not the full picture. Are there more reddit posts with ASRock boards having failures? Yes. Are there nearly zero reddit posts of other boards like you see with essentially any other product? No.
Despite popular belief using reddit as a basis for failure rate analysis is worthless. More people than not will literally just return it instead of making a post.
That tells us basically nothing. Is it probably an ASRock issue? Maybe, but literally nobody has any failure rate information except board partners and AMD and none of them are speaking. If you're worried about the chip dying just don't buy it, regardless of board brand. It'll be there in 6 months or you can get a cheaper option on the same platform.
bushwickhero@reddit
That’s wild. My 9th gen got repurposed for my True NAS when I upgraded to a 9800x3d.
NoStructure5034@reddit
4 faulty chips? Yikes. Glad I stuck with 12th gen.
LightningAndCoffee@reddit
With Intel there are viable alternatives - in fact Intel is arguably the bottom of the market.
Nvidia though - they are best in class in both gaming and AI and neither are close. Good luck with that.
Emergency-Soup706@reddit
Pakistan may have half the GDP but they are getting state of the art equipment from china for literally no cost. We are in the process of trying to create indigenous tech which will take at least 20 years
If it came to total war they wouldn't stand a chance even if their air force may be superior.
The BrahMos missiles hit the targets in Pakistan they were supposed to hit and we can manufacture them in house
Let's see, if anything this should speed up the pace of modernisation of the armed forces
fine_printer@reddit
Time to buy Loongson CPUs, then.
Sopel97@reddit
you're gonna run out of companies soon
vg_vassilev@reddit
Apple Silicon next on the chopping board
Gatortribe@reddit
Being loyal to brands (and doing the opposite) is such a terrible mentality, why would you join any "teams" in the first place? Just buy the best product and ignore all of that. 14th gen Intel definitely was not that, and if they ever come back from the dead you'll just burn yourself again on worthless brand loyalty.
trytoinfect74@reddit
Considering that it's the 2nd or 3rd "final fix that would resolve the issue for good", I personally believe that all of the 13th and 14th are more or less "affected batches" (occurance of the issues tied to voltage and clock as we learnt from the stats) and should be ditched whenever possible but Intel will never admit it.
Advanced-@reddit
I am happy I bought in the platform with DDR4 ram. Just Upgraded my 12600k to a 14600k, with the extended 5 year warranty that should be just about when my PC will need a real upgrade to keep up in games.
Going to get 5 years out of this and move one with a full rebuild. Anyone buying to keep this system for longer than that really needs to make other plans though.
A_NON99@reddit
I think the direct cause of the degradation is the TVB applying high voltages that shorten the life of the device, and the only countermeasure is to prevent the 14900k from using that function, but why does Intel's response keep going around and inflating misconceptions about the problem?
No-Improvement-8316@reddit
Guys, trust us, the issue has been fixed.
Henrarzz@reddit
_asfasd
starcube@reddit
_Hotfix_2.2
vg_vassilev@reddit
_copy
Impossible-Glass-487@reddit
I have a predator with its own updates. Am I supposed to install this or not?
krw755@reddit
Can someone please ELI5 what Vmin means?
BrightCandle@reddit
Electron migration, the volts cause the hole to bigger and bigger requiring an increase in voltage to make the transistors switch until it burns out completely. Intel CPUs are acruing damage due to being overvolted and so they have to keep pumping up the minimum voltage to keep them working which only accelerates the CPU death more.
SkillYourself@reddit
The VF table never changes - it's burned into the factory. The cores just degrade from high voltage until the stock VF table isn't high enough to be stable.
airfryerfuntime@reddit
Vmin = voltage minimum. In 13th and 14th gen intel chips, there's a problem with the clock circuit in part of the core that handles instruction sets. Basically, the minimum voltage required for stable operation literally shifts, so the core needs more and more power. This results in heat and further instability, which is what's causing crashes.
Eat-my-entire-asshol@reddit
Vmin shift instability in Intel CPUs refers to a problem where the minimum voltage required for a processor to operate stably increases over time, often due to elevated voltages and temperatures.
slither378962@reddit
If min volts too high, then your cpu goes bye bye