You need a final BIOS update to fix the Intel CPU stability issue for good
Intel has released another microcode update for 13th and 14th generation Intel Core processors to address a serious issue that can cause crashes and even damage the processors in the long run.
Intel says that this update, which is the third of its kind, addresses the root cause of the instability issue.
The details:
- The patch addresses issues that occurred during idle or light activity periods.
- It is caused by a circuit that does not play well with elevated voltage and temperature levels.
- Intel says that the patch does not affect the performance in games or when running applications noticeable.
Issue has been plaguing users for months
It took months for Intel to identify the root cause of the issue. Intel users started to report crashes months ago and speculation about the cause of the experienced issues ran rampart for some time on the Internet.
Intel released two microcode updates for affected processors, which addressed three of the four identified issues. These two updates were released in June and August of 2024.
The list of affected processors includes 13th and 14th generation Intel Core i9, i7, and i5 processors.
i9-13900KS i9-14900KS
i9-13900K i9-14900K
i9-13900KF i9-13900KF
i9-13900F i9-14900F
i9-13900 i9-14900
i7-13700K i7-14700K
i7-13700KF i7-14700KF
i7-13790F i7-14790F
i7-13700F i7-14700F
i7-13700 i7-14700
i5-13600K i5-14600K
i5-13600KF i5-14600KF
BIOS Update required
The latest patch needs to be released as a BIOS update. Motherboard manufacturers have access to the patch already.
It is not clear when first BIOS updates will become available. It may take weeks for beta update releases and then more weeks before the final version of the BIOS update gets released.
It is still likely that first BIOS updates will be released as early as October 2024.
Intel recommends that customers check for updates at the website of the manufacturer of the motherboard. The Compatible Products website that Intel operates may also provide information on the latest available updates for the processor.
Windows users may load System Information to display information about the motherboard. Just open Start, type system information and press the Enter-key to open the tool.
Check the system manufacturer and model there, or the baseboard listings. Use the information to open the website of the manufacturer of the motherboard and search for updates on it.
Closing Words
Tests and benchmarks will show if the latest microcode update addresses the issue once and for all, and if performance is not impacted noticeably.
Do you use a computer with one of the affected processors? Feel free to leave a comment down below.
I’ve built my system back in November 2022 after I bought the parts mid October but couldn’t finish it yet due to a bend Corsair H170i.
The PC consists of:
– CoolerMaster HAF 700 EVO
– i9 13900K
– Asus ROG Maximus Z790 Extreme
– 128GB Corsair Dominator Platinum RGB DDR5-5600 CL40 @ 5400 CL40 XMP (4x32GB)
– EVGA GTX1080Ti FTW3 11GB
– Corsair H170i 420mm AIO with 6x 140mm Noctua NF-A14 IndustrialPPC 3000RPM fans
– Corsair AX1200i PSU
– 2x Samsung 990 Pro 4TB NVMe @ PCI-E 4.0 4x
– 1x Samsung 840 Pro 256GB
– 1x Samsung 860 Pro 2TB
– 3x HGST 7K4000 Deskstar 4TB (one disconnected for now)
– 3x HGST He10 Ultrastar 8TB (one external)
The CPU is set to -0.06v undervolt through Global SVID but all power limits are removed.
The CPU is running the earliest 0x10B microcode.
The Motherboard is running the earliest non-publicly available 0219 bios from 1 September 2022 which I acquired a copy of through Asus support.
So far I’ve never seen dangerous voltages, not the VIDs nor the Vcore.
The system has been running stable for close to 15K hours in total now.
The bios reports a Vcore of about 1.376v.
In idle state the VIDs never exceed 1.41v and are generally between 1.35-1.38v
The system agent reports 1.382v in the OS
CPU-Z multi-threaded stress-test will drop it down further to ~1.26v average throughout all VIDs
In idle state the Vcore averages between 1.288-1.343v in idle as seen in applications such as AI Suite 3, Aida64 and CPU-Z.
CPU-Z multi-threaded stress-test will drop it down further to 1.243-1.252v
The CPU during Cinebench R23 will stay at maximum around 290W instead of 320-330W without an undervolt.
Temperatures are kept around peak ~90-95°C even in a 30-minutes stress test at average 22°C ambient.
These tests are fully stable, even if Cinebench is set to Realtime which has aided me to find the real stable undervolt setting before the application starts to either crash or BSOD the system.
The maximum load this CPU has endured is Prime95 (FPU) stress without undervolt hitting ~376W power draw and 100°C throttling but was quickly ended after a minute, this was not within cooling capacity but it has neither broken the CPU.
In short, I am not updating the bios to a newer version.
I dodged the bullet by not updating at all as the newer microcodes such as 0x125/0x129 introduced problems that actually damaged the CPUs.
Even though in some cases it may improve performance and not lose much, bios updates may actually cause changes in behaviour in terms of memory stability that I am not giving up.
Later bios versions also introduced undervolt protection where people complained it did not actually apply the undervolt the CPU despite turning everything off which I am not going to risk happening as bios downgrades are not possible beyond a certain point.
CPU-Z Validation with 17063 score: https://valid.x86.fr/c6fhp7
Aida64 Benchmarks: https://imgur.com/a/i9-13900k-128gb-ddr5-5400-benchmarks-aida64-mEPV7FP
Cinebench R23 scored a maximum of 41K points: https://i.imgur.com/L5hHDsN.png
Cinebench 2024 scored a maximum of 2299 points: https://i.imgur.com/uWm9Yzq.png
Just updated my ASUS ROG Strix Z790-H (running i7-14700K) to the latest BIOS even thought it was a beta version. BIOS description read:
1.Updated with Intel microcode 0x12B to address elevated voltage requests during idle or light activity, further mitigating Vmin Shift instability issues.
2.The option to disable C1E in the BIOS has been removed to ensure it remains enabled, reducing processor stress and maintaining stability per Intel’s recommendation.
All went well although a recurring problem with updating the BIOS in this computer has been that upon reboot it gives the message “Inaccessible Boot Device”. The fix on my system has been to search for VDM in the BIOS setup and change “Map PCIe Storage under VDM” & Map SATA Controller under VDM” both to Enabled.
FYI | Just for fun, about a month ago I cleaned up and revived an old 2013, 9020 Dell-mini-desktop, that only runs @ 22 Watts max, on a locked down Win7 OS, and it runs fast and cool as all git-out and does everything ‘securely’ online I want it to do (even in the Texas Heat with low voltage and low rural data)…
Am also using an old bought and paid for Netlimiter-Pro-4.0.40.0 and an old amp meter to prove everything I’m sayin.
And I don’t have to worry about stupid processor issues… I’m usin it right now as I write this with minimal
data usage.
(((Back to the Future Baby)))
These look like they are all desktop or workstation CPU’s. If so, you should mention that in the article Martin. I’m using a 13th gen Intel CPU at work that’s performing great and without the problems you’ve listed, but it’s a laptop CPU, not one of these you listed. Also the laptop came with Dasharo Coreboot, and Dasharo fixes a lot of these kinds of problems themselves.
Intel offering a “fix” widely to millions of non tech users who have no clue this is even an issue. Meanwhile PC manufacturers still selling gear with chips having this flaw, that will never get fixed and end up bricking for many. Leaving perplexed owners to spend double to replace bricked PC unless they press a warranty. I avoided by going to Core Ultra 7 155H, so far unaffected, but it cost me extra to avoid Intel’s mess.
I use Intel Driver and Support Assistant.
https://www.intel.com/content/www/us/en/support/intel-driver-support-assistant.html
It says there are no updates. I went on the site provided in the article and manually searched.
There was nothing found for my board or processor.
It is amazing this issue has gone on so long before a remedy was developed. I do not use any of the affected CPU’s but I certainly understand the owners of these CPU’s frustrations. No doubt Intel will have an uphill struggle regaining confidence in their customers after the mess.
I installed it a few weeks ago for my i9-14900K on MSI MAG Z790.
After install you are forced to choose a performance profile for the CPU by a popup.
There’s 3 choices, though I don’t remember the exact values they are for the power and current limits I manually set before after learning about the issue.
I’m using the perfomance (2/3) setting which is has a slightly higer current limit then what I had it set at. The third setting is the same as the old defualt with the insane values of 4,096 watts and 512 amps.
I have an i7-14700F processor and did the test with the “Intel Processor Diagnostic Tool” program.
I passed it, all green.
Yet it is one of the incriminated processors.
What does that mean?
Things like this is why I switched to AMD long ago, Intel messes up all the time and if you follow these updates your CPU becomes slower even if they say otherwise. If you pay for 700 dollars of performance and you end up getting downgraded to 400 dollars afterwards it’s like Intel steals 300 dollars from you under the guise of security when it’s really about money.
They need to beat AMD at the first few days of reveal & public release of these chips even if it kills it eventually.
It gets worse when manufacturing itself created hardware flaws which degraded the chips even further.
That’s why I am gonna avoid intel maybe 2 generations later just to be sure there won’t be any manufacturing flaws & unless they increase platform support & not just 2 years.
AMD also has problems. People are reporting that this year’s gaming CPUs are well underpowered comparing to specs. But nothing as egregious as Intel.
@boris AMD’s gaming CPUs are X3D chips and they are not released yet. Regular Ryzen 9000 series are for general computing and while they do not offer substantial perf. gains over 7000 series, they are much better efficiency wise.
Completely agree.
Thankfully, I have Lenovo with Intel Core i5-13400. It’s not on the list, but my power supply is wonky. I remember when I bought the computer half an 8 month ago, it had full 65W power at startup and the computer would boot in 40 seconds. Now, power is constrained to 25W at boot and computer starts in 3–4 minutes. I though Intel capped maximum CPU power because of the issues. Now it is a different problem. But it is still great that my CPU is not affected. Thank you for getting the final list of affected processors and everything else.