Perfect CPU does not exist. Buy Intel stocks - they are cheap

Intel 13gen and 14gen problem

As you have likely heard some (top?) 13gen and 14gen Intel CPUs may suffer from manufacturing issues (oxidation in early CPU batches?) and/or improper self-configuration that may lead to irreversible CPU degradation (also known as: "$590 top desktop Intel 13gen/14gen CPU may perform suicide").

It would generally be okay if manufacturer (Intel) confirmed what exactly are the problems. There are still lots of speculations, with no clear and public answers. For example people (including me) wonder why mobile gen13 and gen14 processors are "safe from troubles".

Intel? vs AMD? - CPUs became so complex that they must have bugs

You may have impression that Intel is bad and AMD is "good" (just compare stock prcies), but things are not black and white in CPU market.

Here is the story of most known CPU security flaws in recent years:

• 2005 – 2017 – Branch Target Buffer Attacks and other CPU side-channel attacks: Colin Percival’s research demonstrated a practical attack on modern processors using the branch target buffer. This attack could potentially leak information and was demonstrated by extracting an OpenSSL RSA key. Several other papers on the subject were published during this time, however, mainstream attention was not garnered until Spectre and Meltdown were published in 2018.
• January 2018 – Spectre –? Researchers at Google Project Zero and elsewhere disclosed the Spectre vulnerability (CVE-2017-5753 and CVE-2017-5715). Spectre is a class of speculative execution vulnerabilities that affect a wide range of processors.
• January 2018 – Meltdown –? Researchers disclosed the Meltdown vulnerability (CVE-2017-5754), which primarily affected Intel processors. Meltdown allowed unauthorized access to kernel memory. Note: In 2021 Spectre v2 Variants were disclosed (Spectre-BHB, Spectre-PR2, Spectre-SSB, and others)
• 2018 – RIDL (Rogue In-Flight Data Load) –?Affecting Intel chips produced as early as 2008 the RIDL vulnerability was discovered by the same researchers that disclosed the ZombieLoad and Fallout vulnerabilities. RIDL be used to leak data from the vulnerable CPU’s various internal buffers (select regions of allocated memory used to store or load data).?
• May 2018 – Fallout –? Researchers disclosed the Fallout vulnerability (CVE-2018-12126, CVE-2018-12127, CVE-2018-12130, and CVE-2019-11091), which exploited speculative execution to read data from Intel CPUs’ microarchitectural buffers.
• August 2018 – Foreshadow –? Researchers revealed the Foreshadow vulnerability (CVE-2018-3615 and CVE-2018-3620), which targeted Intel SGX (Software Guard Extensions), allowing attackers to extract sensitive information.
• May 2019 – ZombieLoad – Researchers disclosed the ZombieLoad vulnerability (CVE-2018-12126, CVE-2018-12127, CVE-2018-12130, and CVE-2019-11091). It affected Intel processors and allowed attackers to exploit speculative execution to read data from other processes.
• January 2020 – CacheOut – CacheOut (CVE-2020-0549) was disclosed. This speculative execution attack could leak data from other processes or virtual machines running on the same CPU, even when SMT (Simultaneous Multi-Threading) is disabled.
• March 2020 – LVI (Load Value Injection) – Another speculative execution vulnerability that can be used to inject data into a victim’s transient execution. It affects Intel processors and can leak sensitive information.
• July 2023 – AMD Zenbleed – The Zenbleed vulnerability (CVE-2023-20593) is yet another speculative execution bug that allows data to be exfiltrated at high rate of speed (for this type of attack) clocking in at 30kb per core, per second.?
• August 2023 – AMD Inception – AMD’s Zen 3 and Zen 4 CPUs are affected by the ‘Inception’ vulnerability. Like many of the attacks on this list, inception is a side-channel attack that can lead to the exposure of secure data.
• August 2023 – Intel Downfall – Identified as CVE-2022-40982, Downfall is one of the most recent examples of speculative execution bugs, this time taking advantage of memory optimization features in Intel processors to access user data.

Self-destroying 13gen and 14gen Intel CPUs

Intel published short info on their site on July 22nd, 2024: https://community.intel.com/t5/Processors/July-2024-Update-on-Instability-Reports-on-Intel-Core-13th-and/m-p/1617113

They claim that elevated voltages in 13gen/14gen Desktop CPUs may happen due to buggy CPU microcode. People speculate that all 13gen/14gen CPUs are affected as CPUs share likely similar microcode and similar design.

It is worth to remind that per official specifications desktop CPUs:

• 13900K (launched Q4'22) may reach max 5.80 GHz Turbo Frequency
• 14900K (launched Q4'23) may reach max 6.00 GHz Turbo Frequency

with 253W being a Maximum Turbo Power for those processors and 125W being a base power.

So generally speaking those top desktop CPUs may sometimes consume twice as much energy as they normally do. That will result in lots of extra heat, when CPU decides to "boost" clock frequency to handle your heavy workload (eg. while you are gaming).

Let's take a look at top mobile CPUs:

• i9-13900H (launched Q1'23) may reach max 5.40 GHz Turbo Frequency
• i9-14900HX (launched Q1'24) may reach max 5.80 GHz Turbo Frequency

with 115W and 157W being a Maximum Turbo Power for those processors and 45W and 55W being a base power respectively.

Here we have a situation that 14gen mobile CPU can consume 3x more power than normally!

Intel? Turbo Boost 3.0 Technology

As per Intel webpages:

CPUs don’t always need to run at their maximum frequency. Some programs are more dependent on memory to run smoothly, while others are CPU-intensive. Intel? Turbo Boost Technology is an energy-efficient solution to this imbalance: it lets the CPU run at its base clock speed when handling light workloads, then jump to a higher clock speed for heavy workloads.

Running at a lower clock rate (the number of cycles executed by the processor every second) allows the processor to use less power, which can reduce heat and positively impact battery life in laptops. But when more speed is needed,?Intel? Turbo Boost Technology dynamically increases the clock rate to compensate. This is sometimes called “algorithmic overclocking”.

Intel? Turbo Boost Technology?can potentially increase CPU speeds up to the Max Turbo Frequency while staying within safe temperature and power limits. This can increase performance in both single-threaded and multithreaded applications (programs that utilize several processor cores).

If you’re wondering how to enable Turbo Boost, don’t worry — it’s enabled by default. You don’t need to download or configure anything.

Intel? Turbo Boost Max Technology 3.0 identifies the best performing core(s) on a processor and provides increased performance on those cores through increasing frequency as needed by taking advantage of power and thermal headroom. Intel? Turbo Boost Max Technology 3.0 frequency is the clock frequency of the CPU when running in this mode.

To have a full picture:

Intel Turbo Boost tech is to make sure your silicon can be used in most efficient manner.

CPUs are not only increasing clock frequency, but also do increase voltage using faulty (?) "algorithmic overclocking" ;-) Increased CPU voltage usually helps to stabilize higher clock frequencies, so is expected what happens behind the scene.

All modern Intel and AMD CPUs communicate with motherboard and request different voltage depending on load level.

VID

VID is a value that the CPU sends to the motherboard. The CPU is telling the motherboard what voltage it would like. VID may be used in "auto", "offset" or "adaptive" modes, where the motherboard determines what voltage should be sent to the CPU.

It is particularly important for mobile and battery powered devices because it allows much better battery life as the CPU can tell the host what level of voltage it thinks it needs now based on the load it's running at. When it's at low load, the voltage can be reduced significantly (thus reducing the current) and saving power drain.

VID is not used at all if you set your CPU to a fixed voltage (as many overclockers do) because setting a fixed voltage tells the motherboard that you want to send that voltage to the CPU no matter what the situation is.

At the factory each CPU is tested and a VID table is programmed into the CPU that determines for a number of different load levels (and perhaps even core by core), what voltage the CPU should request. That table is then consulted for a given load level to determine what VID value to send to the motherboard at that particular moment. The VID table can vary from one chip to the next based on how "good" a chip is. Really good chips will request less voltage for the same load compared to not as good chips.

As per observations (HWMonitor and CPU-Z tools both report VID max that was configured in factory) some sample values found in internet were:

• 13900K/KF (sample) - VID max is 1.5v
• 14700kf (sample) - VID idle is around 1.44v and VID max is 1.53v

(they may likely vary per particular particular CPU instance)

Overvoltage = CPU slowly making a suicide

Some people say that with disabled Turbo Boost 3.0 and disabled Enhanced Turbo they observed VID = 1.65v on Intel 14900KF desktop CPU with 6.0 GHz clocked P-cores. This is certainly above max VID = 1.53v which was reported as limit by this particular CPU.

Researcher claimed that 2M-3M stressing of such CPU would lead to irreversible changes in silicon and instable (=junk) CPU. He claimed he can reproduce problem easily, just time and heavy workload is required.

Researcher claimed also that his CPU cooling was perfect, but CPU enjoyed too much listening of "Killing Me Softly with His Song".

Fear not - we have a "fix" for you - microcode update released

On August 9th, 2024 Intel published microcode update version 0x129 for desktop 13gen/14gen processors. Nobody seen it published there: https://github.com/intel/Intel-Linux-Processor-Microcode-Data-Files/tree/main/intel-ucode or https://github.com/intel/Intel-Linux-Processor-Microcode-Data-Files/releases/, but some manufactuers such as MSI or Asus already released BIOS/UEFI updates same day to selected motherboards.

I hope that Intel will publish microcode 0x129 update together with some description on github.com as they used to do in the past.

I am counting also that they will refer to mobile processors situation, because 14900HX is truly getting huge power consumption increase in Turbo Boost 3.0 mode.

Officially they claim problem is not appearing in mobile processors.

What can you do for your 13gen/14gen CPU?

No matter whether you have desktop or mobile CPU it would be wise to make sure you have quality cooling (thermal paste on CPU may be worth replacing; clean up of PC interior is also welcome). Please observe your CPU temperature in full load and make sure that it does not exceed 85-90C (100C is max operating temperature for 13gen/14gen CPUs).

Secondly please search for latest BIOS/UEFI updates (dated after 2024-08-08) in your PC vendor page. Update your BIOS/UEFI, because CPU microcode is normally embedded in BIOS/UEFI flash (FIT):

BIOS/UEFI update causes that (updated) microcode for Intel processor will be loaded earliest as possible, no matter which OS you are going to use.

If there is no BIOS/UEFI (firmware) update available for your PC, Operating System that you are using will soon very likely update CPU microcode for you. Intel names it as "Early OS Microcode Update" and microcode update processes are well described here: https://www.intel.com/content/www/us/en/developer/articles/technical/software-security-guidance/best-practices/microcode-update-guidance.html.

Linux situation is quite well described here: https://www.linuxfromscratch.org/blfs/view/svn/postlfs/firmware.html

Microsoft did not release Intel microcode update yet, but very likely Windows update and KB page will appear.

So generally speaking:

• improve cooling if you can
• install BIOS/UEFI update if you can
• update your OS system if you can
• try to undervolt a bit your CPU if you have "overclocker" colleague and your BIOS/UEFI supports some advanced settings (each non-standard configuration needs to be "load tested" for errors)
• pray that mobile processors are free from this bug (Intel is not going to release any microcode update for them?)

13gen and 14gen desktop processors have also received warranty extension. Normally warranty period was 3 years and now it became 5 years.