Leaked internal reports allegedly reveal Intel's instability problems are not over — elevated voltages could be only one of the causes of CPU crashing

Intel Raptor Lake processors for desktops
(Image credit: Intel)

Intel announced a few days ago that elevated voltage levels are the root cause of Raptor Lake's instability issues. The company will address this problem with a microcode update in mid-August. However, German publication Igor's Lab claims that Intel didn't share the whole story and is still investigating other root causes of Raptor Lake instability.

According to the publication's sources, elevated voltages aren't the true root cause—or perhaps the only root cause. The claims only somewhat contradict Intel's official statement, which mentions elevated operating voltage as the root cause of the issues but does say that validation work is ongoing to ensure the microcode patch addresses the entirety of the issue. 

Igor's report also highlighted details of the elevated voltage behavior causing Raptor Lake instability and what precisely Intel is adding to its mid-August microcode update. Intel has purportedly observed a significant increase in minimum operating voltage (Vmin) across multiple cores on unstable chips returned from customers. The elevated minimum operating voltage is purportedly caused by the "elevated voltage, high frequency, and elevated temperature" conditions that Intel has observed in chips that have been returned. 

This unsafe minimum operating voltage affects the chip even at idle conditions. Intel has observed sporadic elevated voltages when Raptor Lake chips resume low-power states to execute background operations before entering a low-power state again.

To fix the issue, Intel's microcode update will reportedly limit VID requests above 1.55V to rectify the voltage-related problems. This change will not impact performance much. However, Intel says that further analysis is required to see whether or not the mid-August microcode update will fully mitigate all instability issues due to the reasons above.

Here's a snippet of the purportedly leaked communications given to Igor's Lab:

– Intel observes a significant increase to the minimum operating voltage (Vmin) across multiple cores on returned affected processors from customers.
– This increase is similar in outcome to parts subjected to elevated voltage and temperature conditions for reliability testing.
– Factors contributing to this Vmin increase include elevated voltage, high frequency, and elevated temperature.
– Even under idle conditions at relatively cool temperatures, sporadic elevated voltages are observed when the processor is resumed from low power states in order to service background operations before entering a low power state again.
– At a sufficiently high voltage, these short-duration events can accumulate over time, contributing to the increase in Vmin.
– Intel analysis indicates a need to reduce the maximum voltage requested by the processor in order to reduce or eliminate accumulated exposure to voltages which may result in an increase to Vmin.
– 
While Intel has confirmed elevated voltages impact the increase in Vmin, investigation continues in order to fully understand root cause and address other potential aspects of this issue.
––Intel—

Aaron Klotz
Contributing Writer

Aaron Klotz is a contributing writer for Tom’s Hardware, covering news related to computer hardware such as CPUs, and graphics cards.

  • rluker5
    I think 1.55v is erring on the side of performance. 1.5 max requested would probably be better and 1.4v would be safe so long as the motherboard weren't cranking the volts with some auto tuning.

    I also don't think it is the only issue as the unlimited power and poor motherboard "stock" behaviors are still out there. That and motherboards can give more volts than requested by the CPU. Really need to get motherboard vendors in line.
    Reply
  • HyperMatrix
    Using the latest bios with the intel default specs shows the predicted voltage for 6GHz, which is the stock boost clock for the 14900K, is 1.609V. lol. That is nuts. In comparison, 5.3GHz is just 1.22V and 5.7GHz is 1.44V.

    I decided to set a per-core max multiplier of x57 and disabled TVB. So it works like normal but doesn’t try to pull over 1.6V for an extra 300MHz in lightly threaded applications.

    The “big numbers” marketing is what’s leading to all these problems. All core clocks are set to 5.7GHz anyway. 6GHz+ is only when a couple cores are being used. And it’s really not worth such a big voltage spike. Should have just advertised it as 5.7 or 5.8GHz and left it at that.
    Reply
  • diminishedfifth
    I dunno about it just being a voltage issue. Didn't Intel and MB manufacturers decrease the volts in their updates and still the issue persisted?
    Reply
  • greymaterial
    diminishedfifth said:
    I dunno about it just being a voltage issue. Didn't Intel and MB manufacturers decrease the volts in their updates and still the issue persisted?
    the newer chips like rpl has power gates located inside the core for each core, therefore a low latency when the core needs high freq requesting high voltage will happen within the core itself. MB bios is basically setting the range of voltage and power allowed to supply to the chip externally, where the power gates in core is controlled by microcode only.
    Reply
  • ThatMouse
    Shameful article. No mention of the oxidation problem.
    Reply
  • JRStern
    Is this only or mostly on unlocked chips, or other identifiable set of product numbers or features?
    Reply
  • -Fran-
    Intel not being truthful?

    kermitsurprisedface.gif

    Regards.
    Reply
  • d0x360
    diminishedfifth said:
    I dunno about it just being a voltage issue. Didn't Intel and MB manufacturers decrease the volts in their updates and still the issue persisted?
    It's not, it also an issue with the interconnects (probably between ecores - cache) that is being damaged or was always defective and has been since 13th gen.

    Probably

    The voltage issues might have made the problem much worse but it's probably not the root cause it's a bandaid to try and slow the failure rate but almost all if not all of these chips are going to self destruct earlier than they should.
    Reply
  • bit_user
    Who was the purported recipient of this memo? The way it's written like "Intel this", "Intel that", and signed "--Intel--" makes it sound like it was something sent out to some of their partners or investors.
    Reply
  • bit_user
    HyperMatrix said:
    I decided to set a per-core max multiplier of x57 and disabled TVB. So it works like normal but doesn’t try to pull over 1.6V for an extra 300MHz in lightly threaded applications.

    The “big numbers” marketing is what’s leading to all these problems. All core clocks are set to 5.7GHz anyway. 6GHz+ is only when a couple cores are being used. And it’s really not worth such a big voltage spike. Should have just advertised it as 5.7 or 5.8GHz and left it at that.
    That sounds wise, but perhaps not sufficient. The article cites:
    "This unsafe minimum operating voltage affects the chip even at idle conditions. Intel has observed sporadic elevated voltages when Raptor Lake chips resume low-power states to execute background operations before entering a low-power state again."
    This could mean there are spikes not affected by that multiplier and that are too quick for you to see, but perhaps can still cause damage? Plus, the continued hunt for a root cause suggests it's perhaps not all just about voltage.
    Reply