i9 13900HX dying? (laptop)

[Ritzs]

System: Acer predator helios 16 (PH16-71)
Processor: i9 13900HX
GPU: RTX 4070

Hello Everybody, I come to you with a problem that has me and some others stumped. Also I'm very new to the entire diagnosing PC problems thing so please excuse my naïveté.

I purchased this laptop in April 2023. So I'm out of luck in the warranty department as Acer's carry-in is only for one year. (Also for some reason their "extended warranty" webpage never seems to work, story for another day though)

Now the problems started a couple months ago when Apex legends started crashing constantly with DXGI_DEVICE_ERROR_HUNG. Literally crashing almost immediately after joining an online match. The crash doesn't seem to happen in the lobby or the firing range.

Following this happening, I thought it would be a graphics driver issue and posted this thread on the Geforce forums.

After a brief conversation there, I realized that rolling back drivers seems to work for a while before it starts crashing again. Following this I was asked to run benchmarks and HWinfo logs on my laptop, which I did.

We ruled out the monitor being a problem already.

Drive link with logs and HWinfo screenshots.

We noticed random CPU thermal spikes which were leading to thermal throttling and weird behavior of the memory.
To rule out thermal paste, I got it re-pasted in a local acer outlet. The thermal paste seems to be working because afterwards the idle temperatures went down quite a bit.

However, this didn't solve the issue and we're back to square one now. I tried playing apex legends, the finals and cyberpunk2077 after that with all of them crashing after a few mins of gameplay. The laptop seems to work fine otherwise. You'll find the crash logs here

Feel free to go through these as well:

EA forums

Acer forums

I'm at a dead end now. I contacted Acer support and they did their usual rigamarole. The Acer service center guy told me it'll take a while for him to find a spare stick of ram for me to test if the memory is the problem and I'm currently waiting for that.

My main questions would be:

1. Are there any more diagnostics I can do to pin-point the issue?
2. What are the chances that I have a failing intel CPU?
3. In case of a CPU failure, do I have any "fixes" other than a complete motherboard replacement?

 

[Zach_01]

Is there a way to try to lower the boost of the CPU?
Say below 5Ghz?
Like 4.8?

Just to see if that fixes the problem.

 

[ShrimpBrime]

I would try.

Disable iGpu in device manager.
Disable iGpu graphics in bios also.
DDU ALL display drivers, restart.
Install fresh up to date GPU driver for the 4070 only.

GL.

 

[Ritzs]

Is there a way to try to lower the boost of the CPU?
Say below 5Ghz?
Like 4.8?

Just to see if that fixes the problem.

Acer has unabled any kind of CPU voltage customization. Being an HX CPU this is a weird move by acer and I see a lot of people complaining about it on the acer forums. On the Neo models, you can access an advanced BIOS utility, but on the model I have (Non-neo) this route also doesn't seem possible. Again, very weird considering the neo model is the more 'budget-friendly' option of the two.

 

[Ritzs]

I would try.

Disable iGpu in device manager.
Disable iGpu graphics in bios also.
DDU ALL display drivers, restart.
Install fresh up to date GPU driver for the 4070 only.

GL.

iGPU is already disabled in both Bios and predatorsense (Acer's software). DDU was already done when I first noticed the issue. Still no luck

 

[Zach_01]

Acer has unabled any kind of CPU voltage customization. Being an HX CPU this is a weird move by acer and I see a lot of people complaining about it on the acer forums. On the Neo models, you can access an advanced BIOS utility, but on the model I have (Non-neo) this route also doesn't seem possible. Again, very weird considering the neo model is the more 'budget-friendly' option of the two.

I didn't mean changing voltage directly.
Just if you can get the max multi from 52x to 48x
Voltage will drop accordingly I believe

 

[Laurijan]

Is this CPU affected by the recent Intel 13th and 14th gen CPU problems?

 

[Zach_01]

Is this CPU affected by the recent Intel 13th and 14th gen CPU problems?

If by downclocking it doesnt crach anymore then its highly possible.

 

[Laurijan]

If by downclocking it doesnt crach anymore then its highly possible.

Would not hurt to update bios and microcode too although I heard if Intel CPU is already unstable it most likely doesnt help. Latest microcode that addresses the problems is 129
Also if the CPU had degraded even though using stock clock settings there is a chance that warranty is longer than 1 year. In EU its usually 2 years for most stuff but if a product is defect from the start like with some 13th and 14th gen CPUs you still can get warranty support (under fault responsibility) here even if over 2 years old.

 

[Ritzs]

Would not hurt to update bios and microcode too although I heard if Intel CPU is already unstable it most likely doesnt help. Latest microcode that addresses the problems is 129

Acer hasn't released a microcode BIOS update yet

I didn't mean changing voltage directly.
Just if you can get the max multi from 52x to 48x
Voltage will drop accordingly I believe

Multiplier option on throttlestop doesn't seem to be selectable unfortunately. No external programme works on this laptop. I've tried XTU, afterburner and throttlestop.

Is this CPU affected by the recent Intel 13th and 14th gen CPU problems?

This might be the case.

 

[Zach_01]

Would not hurt to update bios and microcode too although I heard if Intel CPU is already unstable it most likely doesnt help. Latest microcode that addresses the problems is 129
Also if the CPU had degraded even though using stock clock settings there is a chance that warranty is longer than 1 year. In EU its usually 2 years for most stuff but if a product is defect from the start like with some 13th and 14th gen CPUs you still can get warranty support (under fault responsibility) here even if over 2 years old.

I know this, but it would be good to test it like as is just by down clocking, before attempting the new microcode

Acer hasn't released a microcode BIOS update yet


Multiplier option on throttlestop doesn't seem to be selectable unfortunately. No external programme works on this laptop. I've tried XTU, afterburner and throttlestop.


This might be the case.

BIOS is locked I presume...

Have you checked this?

[INTEL]-How To Update Your Microcode for Intel HX 13/14th Gen. CPUs Laptops/Mobile Easily.

Let me start with a small background info... Since we know all Intel's fiasco about what is happening with Raptor Lake/Raptor Lake Refresh which is Intel's codename for the 13th and 14th generations of Intel Core processors, Most OEM/Vendors doesn't want to provide BIOS update that includes...

www.techpowerup.com

 

[Laurijan]

maybe this thread helps:

[INTEL]-How To Update Your Microcode for Intel HX 13/14th CPUs Laptops/Mobile Easily.​

 

[Ritzs]

I know this, but it would be good to test it like as is just by down clocking, before attempting the new microcode


BIOS is locked I presume...

Have you checked this?

[INTEL]-How To Update Your Microcode for Intel HX 13/14th Gen. CPUs Laptops/Mobile Easily.

Let me start with a small background info... Since we know all Intel's fiasco about what is happening with Raptor Lake/Raptor Lake Refresh which is Intel's codename for the 13th and 14th generations of Intel Core processors, Most OEM/Vendors doesn't want to provide BIOS update that includes...

www.techpowerup.com

Yep BIOS is locked. I'll check the microcode link out now! Thank you so much!

 

[Zach_01]

Yep BIOS is locked. I'll check the microcode link out now! Thank you so much!

Its convenient because you can uninstall it anytime

Yep BIOS is locked

Some laptops used to have some key combo function that would make settings appear in BIOS...
I dont have clue if this is still a thing with latest.

I hate this kind of lock up. They should at least make it unlocked for downward settings

 

[Ritzs]

Meanwhile I found this. Seems like Acer in their infinite wisdom have locked undervolting at the CPU microcode level. So I have no clue how safe updating the microcode on my laptop will be.

Its convenient because you can uninstall it anytime


Some laptops used to have some key combo function that would make settings appear in BIOS...
I dont have clue if this is still a thing with latest.

I hate this kind of lock up. They should at least make it unlocked for downward settings

The Neo models allow getting into the advanced bios with the key combo fn+tab. However this doesn't seem possible on the non-neo model which I have. I feel like I'm out of luck. I'll backup my work and try the microcode update and hope for the best.

 

[Ritzs]

Applied the microcode. Seems to have been successful. I'll test tonight and update this thread.

 

[Laurijan]

Applied the microcode. Seems to have been successful. I'll test tonight and update this thread.

Is is said in longer explanations of this topic that once CPU is unstable microcode probably wont help anymore. Its to save those non damaged CPUs mainly. But doesnt hurt to try and run tests.

 

[SL2]

Is there a way to try to lower the boost of the CPU?
Say below 5Ghz?
Like 4.8?

Just to see if that fixes the problem.

Would windows settings do as a plan b?

 

[Ritzs]

Would windows settings do as a plan b?
View attachment 359921

This is already at 95

 

[Ritzs]

Ran the Intel diagnostic tool. It says everything is working A-ok. So now I'm beyond confused. Could be a fluke. Could be my memory that's the actual problem. Could also be my 4070.

 

[iameatingjam]

Wtf you can't even underclock your cpu in the bios? Man what the hell. Throttlestop didn't work as well eh. Windows has built in frequency limiter thats normally hidden (not the maximum state one, but actual frequency). You can unhide it with a registry change.

How to Add or Remove 'Maximum processor frequency' in Windows 10 & 11 - MajorGeeks

Maximum processor frequency specifies the operating frequency of your CPU's core. Maximum processor frequency is hidden and set to 0 by default, which means your processor runs as fast as allowed. Advanced users can slow down the processor if your computer is running hot or consuming too much power.

www.majorgeeks.com

Is this CPU affected by the recent Intel 13th and 14th gen CPU problems?

Intel says no but its the same silicon so I don't see how it could not be affected. Just probably less-so due to the... power saving nature of laptops.

 

[Zach_01]

Ran the Intel diagnostic tool. It says everything is working A-ok. So now I'm beyond confused. Could be a fluke. Could be my memory that's the actual problem. Could also be my 4070.

How about those crashes though?

Did you run the Intel diagnostic tool before apply the microcode?

 

[Ritzs]

How about those crashes though?

Did you run the Intel diagnostic tool before apply the microcode?

Unfortunately I didn't run the diagnostic before applying the microcode... I played cyberpunk for a bit today (after microcode application) and it crashed again. P-core 2 and 3 seem to thermal throttle immediately followed by the entire package. The CPU sustained temps around 94-96 degrees. Here's the hwinfo log.

Wtf you can't even underclock your cpu in the bios? Man what the hell. Throttlestop didn't work as well eh. Windows has built in frequency limiter thats normally hidden (not the maximum state one, but actual frequency). You can unhide it with a registry change.

How to Add or Remove 'Maximum processor frequency' in Windows 10 & 11 - MajorGeeks

Maximum processor frequency specifies the operating frequency of your CPU's core. Maximum processor frequency is hidden and set to 0 by default, which means your processor runs as fast as allowed. Advanced users can slow down the processor if your computer is running hot or consuming too much power.

www.majorgeeks.com

Intel says no but its the same silicon so I don't see how it could not be affected. Just probably less-so due to the... power saving nature of laptops.

Usually on Acer laptops you can hold fn+tab before turning it on to get into the advanced BIOS where these options are available. Acer has a line of 'helios neo' laptops where this is possible. Now the 'neo' line is the more budget-friendly counterpart of the normal helios laptops. But me and other users of the normal helios line of laptops can't seem to replicate this method. So my best guess is Acer somehow skipped or ignored this in the helios models.

Thanks for the registry trick. I'll try this, test and update this thread.

Wtf you can't even underclock your cpu in the bios? Man what the hell. Throttlestop didn't work as well eh. Windows has built in frequency limiter thats normally hidden (not the maximum state one, but actual frequency). You can unhide it with a registry change.

How to Add or Remove 'Maximum processor frequency' in Windows 10 & 11 - MajorGeeks

Maximum processor frequency specifies the operating frequency of your CPU's core. Maximum processor frequency is hidden and set to 0 by default, which means your processor runs as fast as allowed. Advanced users can slow down the processor if your computer is running hot or consuming too much power.

www.majorgeeks.com

Intel says no but its the same silicon so I don't see how it could not be affected. Just probably less-so due to the... power saving nature of laptops.

Alright I tried setting the max frequency to 4800Mhz. However this does not seem to work as HwInfo is still reporting a clock of above 5000Mhz.

Immediately after starting apex the processor goes up to 96° and thermal throttles. Weird thing is the game never crashes in the lobby / firing range or bot royale (only bot enemies) mode. But when I open an online match, it immediately crashes with either a "DXGI_DEVICE_ERROR_HUNG" or "DXGI_DEVICE_ERROR_REMOVED"

If I force dx12 in game options on steam, it takes a while to crash (10ish mins). Same behavior both ways.

I have no clue why this happens as I'm connected to ethernet (It happens on wifi too) and everything else seems to work.

Other games crash 15-30 mins after opening (Finals, cyberpunk) without any error message.

 

[Laurijan]

I would test ram with memtest86+ and a bootable usb stick next as well as check SSD/HDD smart data for errors with crystaldiskinfo

 

[Ritzs]

I would test ram with memtest86+ and a bootable usb stick next as well as check SSD/HDD smart data for errors with crystaldiskinfo

Doing it next.