After upgrading legacy Thunderbolt drivers to DCH ones, there is question mark next to AMD eGPU (Radeon VII or RX Vega 64, both have the issue):
This device cannot find enough free resources that it can use. (Code 12)
If you want to use this device, you will need to disable one of the other devices on this system.
I saved a screenshot of working resource allocation and compared it against failing one. There is definitely a difference with PCI Express Switch Port (Downstream/Upstream). One of ranges is missing in failing case. I can provide screenshots.
I cannot roll back to legacy Thunderbolt driver. It seems no longer works with Windows 19H2 - complains about control application
Thank you for posting on the Intel® communities.
In order to check this behavior futher, could you please confirm the following informaiton?
1- External Graphics Processing Unit (eGPU) Enclosure model that you are using with the AMD* video cards:
2- Just to make sure regarding the screenshots attached, it the first one (top) the one related to when this configuration is working with the Thunderbolt™ 3 Legacy Driver? And is the second one (bottom) the one with issues with Thunderbolt™ 3 DCH Driver?
3- Please run the Intel® System Support Utility (Intel® SSU) and attach the report to this thread.
3.1- Download the Intel® SSU and save the application on your computer.
3.2- Open the application and click "Scan" to see the system and device information. The Intel® SSU defaults to the "Summary View" on the output screen following the scan. Click the menu where it says "Summary" to change to "Detailed View".
3.3- To save your scan, click Next and click Save.
3.4- To attach a file, you must click the “Attach” option on the bottom left-hand corner of the response box.
Intel Customer Support Technician
1 - Model is Razor Core X
2 - Both screenshots are taken with the same Thunderbolt DCH driver. First screenshot taken on 1st boot, 2nd screenshot was taken after reboot (power cycle). All subsequent boots are exposing same problem.
Thank you for your response and the information provided.
Please allow us to look into this behavior and as soon as we have more details available we will be posting back in this thread.
Intel Customer Support Technician
I appreciate all comments and reports regarding this issue as well as your patience. We have been looking at this but unfortunately, I don't have any recommendations at this point.
We are trying to allocate internal resources to do some more digging and debugging but this may take extra time and I cannot promise any resolution at this moment.
I am tracking all eGPU compatibility related cases in order to provide details to the product team for further debugging and I will get back to this community post as soon as I hear from them.
I would like to be of much help to you but for now, this is all I got.
According to the forums on egpu.io, the problem was introduced with the Windows 10 Build 1903/1909. After reverting pci.sys to an older version from Build 1809 the problem seems goes away. I did not verify it myself, but maybe Intel has a leverage on Microsoft to investigate and fix?
Thanks for the details and also for reporting.
We are looking into this issue but it may take time especially if this is somehow Windows* related.
I would recommend that you also open a ticket with Microsoft.
Let me do some more digging and I will come back to this post with an update however I can't promise any fix at this point.