I recently purchased a NUC10i7FNH to use at home. I had previously purchased a Razer Core X eGPU and Nvidia Geforce RTX 2060 for my laptop.
I have the latest versions of all drivers, software and I've tried multiple different Windows builds, including fast track, slow track and retail releases. I'm currently on Windows fast track build 2004.
The NUC works just fine with no issues if I just use the built in video card connected via HDMI.
The eGPU and video card have worked without issue for the past year connected to my Razer Blade Stealth.
When paired with the NUC10i7FNH the eGPU exhibits various issues.
- Sometimes it will not get detected on boot and I will have to unplug/replug it in.
- Once connected it will work for a random amount of time before it acts like its unplugged/plugged back in. It will do this constantly, sometimes it will come back and sometimes it wont.
- I've tried different TB3 cables in case it was damaged but same issue
I've scoured through Event Viewer for errors. I don't see errors each time it disconnects but sometimes I will see the below errors:
- The driver \Driver\WudfRd failed to load for the device PCI\VEN_8086&DEV_15E8&SUBSYS_20818086&REV_06
- Which after some googling and verifying in devicemanager, thats the thunderbolt 3 controller
- Another one is relating to this: Driver nvlddmkm Ddi Add Device_Add Device Failure
EDIT 1: None of the errors I've encountered are related to TDR.
EDIT 2: I've uploaded my system report from Intel Driver and Support Assistant
Thank you for submitting your question on this Intel® Community.
To better assist you, we would like to gather more information about your system configuration.
- Please generate a system report with the Intel® System Support Utility (Intel® SSU) and attach it to this thread.
- Intel® SSU Download link
- Open the application and click on "Scan" to see the system and device information. By default, Intel® SSU will take you to the "Summary View".
- Click on the menu where it says "Summary" to change to "Detailed View".
- To save your scan, click on "Next", then "Save".
- What is the Thunderbolt* 3 version currently running on the computer?
- Firmware and drivers versions the Razer Core X is running.
Intel Customer Support Technician
I attached the out put from the driver update tool, it just needs to be renamed from .txt to .HTML.
As far as drivers, it's whatever the latest from the Intel website is for thunderbolt and for razwr core x, I'm not sure as razer doesn't provide driver downloads, so it's whatever windows update installed.
Thank you for your response.
Based on the report, you are running a preview build of Windows*. When running a preview build, systems are prone to instability, which may cause Windows* to stop responding while running applications or performing specific tasks.
- Is there any chance you can let us know if this issue occurs on the latest version of Windows* 10 1909 (18363) November 2019 Update?
- Also, try to install the latest BIOS and Thunderbolt* 3 firmware versions available at https://downloadcenter.intel.com/
- BIOS Update Version: 0039
- Thunderbolt™ 3 DCH Driver for Windows® 10 for Intel® NUC Version: 64. Please follow the steps available on the detailed description of the driver.
In addition, try these steps:
- Set the Power Mode in NUC BIOS to: High Performance
- Switch off PCI Link Power Management in Windows Power Setting
- Set the power mode in NVIDIA control panel to Maximun Performance. Go to 3D Setting>Global Setting>Power Management Setting>Maximum Performance
Intel Customer Support Technician
I've tried on the retail build of windows and I encountered the same error. My laptop (razer blade stealth) is on the preview as well and have no issues using the eGPU.
I updated the BIOS to the one linked and installed the Thunderbolt 3 driver you listed, however it downgraded the version from 1.41.789 to 1.41.769, not sure if that is intentional or not. Version 789 must have been installed by Windows from Windows Update.
I've completed the other settings listed as well previously trying to get this to function.
I'll report back if/when I encounter the error again.
Attached the output from SSU
It doesn't seem to disconnect however when trying to play a game it completely crashes everything requiring a hard reboot.
EDIT: Seems because I had both the internal GPU and eGPU connected to different monitors it caused the hard crash. Once I disabled the internal GPU and disconnected it the crash seems to have disappeared.
EDIT 2: Nevermind, when it disconnects it seems to crash everything.
Is there a way for me to troubleshoot if the hardware in the NUC is bad? I was looking over bios settings and it was doing it, so could the thunderbolt hardware be bad?
EDIT: I was also just using the eGPU with my Razer Blade Stealth for a few hours playing a game and I had zero issues.
I was able to capture a memory dump when it crashed earlier, I have attached it.
Disconnects still happen, here is what occurs:
After 15-30 minutes it will disconnect and reconnect, then it will continue to do it for 1-3 minutes then it will BSOD or it will just lock up requiring me to hold down power for 10 seconds to force it off.
I have been internally working on this issue along with Wanner and Alex and I have to let you know that we are investigating the root cause for the behavior that you are reporting but this may take time and on top of that we are also dealing with limited resources due to the global pandemic situation that we all are facing.
Having said that, we have two options: a) wait for the product team to investigate and come up with their findings which may not be positive (we can't promise anything at this time) and this will take time or b) reach out to our contact center via phone or chat to explore other alternatives to resolve this issue, here is our contact us information: https://www.intel.com/content/www/us/en/support/contact-support.html#@7. Please make reference to case number 4558730.
I apologize for the inconvenience this issue has generated, I totally understand your position about this issue.
We are currently investigating this issue.
We have sent you a private message to request personal information. Please reply to this message at your earliest convenience.
Intel Customer Support Technician
ADeFr2 - saw your reply in NIlls thread. Replying here to keep that thread clean. Note: I am no expert and have not solved my random freezes.
You might read through my thread (start after the wireless issue). https://forums.intel.com/s/question/0D70P000006vrOe/nuc10i7fnh-get-error-messages-about-wifi-6-ax201...
I say that because you, like I, appear to have newer Intel generic drivers for several items. I had to turn off Windows 10 automatic updating of drivers along with disconnecting from the internet to install and keep the NUC specific drivers.
Nills said to change a power setting and add two items to the registry. I have yet to do that have you? This post explains how to add to the registry. https://docs.substance3d.com/spdoc/gpu-drivers-crash-with-long-computations-128745489.html
Researching based off his post, I found this. https://www.reddit.com/r/eGPU/comments/ds4m9v/core_x_chroma_rtx_2070_super_nuc8i5_finally/ Though not the same NUC, it goes beyond the power setting/registry additions and says to use the legacy thunderbolt drivers.
If this helps, let me know. I have changed what I think is the Intel equivalent power setting (Intel Dynamic Power Technology to Custom, which has a Max Performance setting), but still had a freeze. I will add to the registry, and possibly try the legacy thunderbolt driver. In the meantime, I disconnected my thunderbolt (which drives a second monitor) because the freezes have become too frequent for the NUC to be useable.
I have exactly the same issue with my Intel® NUC Kit NUC6i7KYK.
i have an external GPU (Gigabyte GeForce GTX 1080 G1 Gaming (8GB) in a Akitio Node and it has worked perfectly well for around two years via thunderbolt 3 connection.
All of a sudden a few weeks ago it suddenly kept disconnecting. THe only thing that happened during that time was some updates to Windows 10 (around Mid March). Since then i have updated all the drivers, firmware, bios,. I rolled back windows to an earlier version and removed updates but none of this has worked.
As far as i can tell the eGPU, case, cables etc. are all working properly. And my conclusion so far is that the likely cause has been a windows update rolled out Mid March.
Has there been any analysis of whether this is the cause (e.g. did windows update modify settings which have caused this issue), or did the nuc system just suddenly mess up?
Thank you very much for the feedback, this type of data is very important for the product team that is looking into these issues.
I really don't have anything to add other than let you know that we are currently investigating these issues, that we have limited resources to do that, and that with the Covid-19 situation everything is a bit more difficult.
I am compiling a list of all of the reports and whenever we have anything new, I will come back to the community and will post a notification.
Unfortunately, this is all I have for now.
thanks for the information
an update - I reset windows and reinstalled and the issue persists - so a clean reinstall did not resolve
the eGPU is identified and connects to the nuc, and as long as i dont run any programs or do anything it is somewhat stable - however as soon as I open a program it disconnects
any help on this would be appreciated
Again, my problem is worse with a monitor connected via thunderbolt, but I have random freezes without, so the problem may well not be limited to eGPU/thunderbolt. I have updated my long thread https://forums.intel.com/s/question/0D70P000006vcqr/nuc10i7fnh-get-error-messages-about-wifi-6-ax201...