Haswell i5 4200U Speed being held down to 0.78GHz for periods for no apparent reason

[Martinthefizzio]

I have a Sony Vaio SVF14N2A4E laptop with a Haswell i5 4200U 2 core 1.6GHz (base speed 2.29GHz) processor.

It's quite an old laptop I suppose hailing from circa 2012 but it works fine and is absolutely usable and capable of doing what I need of it which is fairly basic browsing word processing and very occasional digital image work.

When the processor is operating correctly the observed speed (from Windows 10 Task Manager) varies in the range 1.6-2.2GHz and the PC works fine.

The problem is that regularly the PC slows down markedly in its speed of use sometimes making it vertually unusable. The observed speed (in Task Manager) when this happens is a rock steady 0.78GHz. This "lockdown" continues for a seemingly random period before returning to normal functionality.

There are no options in the BIOS (I have the latest) to change the operation of SpeedStep

I have the latest version and updates of Windows 10 and the latest drivers for all devices except apparently for the Intel Management Engine, Intel 8 Series USB Enhanced Host Controller.

The processor temperature seems to be reasonable at around 53C and the issue does not appear to be related to the temp.

The intel diagnostics test run while the problem was evident reports the expected frequency as 1.6 and the measured frequency as 2.29GHz all whilst the processor was running at 0.78GHz???

What's going on?

How can I get this problem sorted?

Is there another way to disable/control the SpeedStep technology if this is the problem? I'd be happy to have the processor run at a fixed rate of 1.6GHz as at least this would be better than the mega slow 0.78GHz...

[AlHill]

You may want to re-do your thermal paste.  While there, make sure the heatsink is clean and the fan is working.

Doc (not an Intel employee or contractor)
[Waiting for Windows 12]

[Martinthefizzio]

Thanks for the advice AlHill, but since the process temp is well below a critical temp when the throttling starts it's not necessarily first thing in my sights.

Since posting the original question i'v run throttlestop to try to help diagnose the problem. This has been very useful in showing that PROCHOT flag is the cause of the slowdow even though the individual core and package temperatures are well within allowed range.

So it would appear that something else is setting the PROCHOT true. Anyone know what the source of this might be (Intel built in GPU? or something else). Is there a way of monitoring these other source temperatures?

At the end of the day I suppose I can always use TS to simply disable PROCHOT to allow the processor to operate normally. Since it's not the cores themselves that are overheating there shouldn't be that much of a risk considering my general usage...?

Martin

[JosueO_Intel]

Hello Martinthefizzio,

Due to this product being discontinued, Intel Customer Service no longer supports inquiries for it, I am glad to see that fellow community members have the knowledge and they jumped in and are helping you. You may also find the Discontinued Products website helpful to address your request. 

https://www.intel.com/content/www/us/en/support/discontinued-products.html

You can verify this product's discontinuance status at the Intel Product Specifications website > Product Status > "Discontinued".

Please keep in mind that this thread will no longer be monitored by Intel. Thank you for your understanding.

Best regards,

Josue O.

Intel Customer Support Technician

[n_scott_pearson]

PROCHOT can be exerted by devices external to the processor. It is typically asserted by the VRM, which often includes a capability to assert PROCHOT if a thermal sensor placed in the vicinity of the FETs should show overheating. The answer Al gave still applies; there is likely an issue with the thermal paste.
Hope this helps,
...S

[n_scott_pearson]

Oops, forgot to mention, external PROCHOT assertion cannot be blocked.
...S

[Martinthefizzio]

Hi Scott

Thanks for the info. Will definitely look at the heat sink paste and replace if I feel confident.

As far as PROCHOT not being able to be blocked...

I've found that using ThrottleStop and unchecking BD PROCHOT the processor speed goes back to normal operation and even with extended use the core and pkg temperatures don't go above 53C.

Would reducing the voltage to the processor reduce the heat generated in the VRM and hence prevent the PROCHOT being asserted?

M

[n_scott_pearson · Accepted solution]

Sorry, I said it incorrectly. While the PROCHOT signal into the processor is non-maskable, you can configure the processor to ignore it. Not a good practice, but perhaps the only solution. The problem, of course, is that the signal is being generated for a reason. While there is the possibility of a bad sensor, the probability is pretty low.

Sadly, if you've updated the firmware on your board, it may have the newer version of the microcode that disables support for undervolting. If you (still) can, you can try undervolting, but I do not think that that is the right answer. I think the right answer is to address the source of the assertion. This may mean having to approach the motherboard vendor and have them replace it.

...S

[Martinthefizzio]

Thanks for the clarification.

Since Sony have sadly pulled out of the laptop market getting support from them is a no-no

I'll certainly experiment with the under-volting using Throttlestop if my BÌOS allows it and ĺook at replacing the heat sink paste.

Thanks everyone for your help here, there is some very good information and possible solutions for others with similar problems in this thread.

M

[n_scott_pearson]

The good (?) news is that your laptop is so old that there probably haven't been any BIOS updates that would introduce the microcode updates disabling undervolting (they're susceptible to the Plundervolt! vulnerability instead ;^) )

...S