Hi everyone!
I have Lenovo X1 Carbon 6th Gen (i7 8550u + 16GB RAM). Recently I bought Aorus Gaming Box with RTX 2070 on board, then setup it easily on Windows 10 and Linux kubuntu 19.10 without any problem.
I use it for gaming on Windows 10 and for neural networks on Ubuntu (in development purposes for my job).
While there are no problems in Windows 10, I am experiencing some issues with it on my Ubuntu system. All works fine when the system is booted and I connect the eGPU. The problems begin when I just disconnect it from my laptop:
- After disconnecting, hibernate doesn't work for some reason (the laptop goes to black screen and then it can be awaken pressing any key, it is even not like suspend)
- After disconnecting, the whole system freezes some time later, only mouse works. Even other terminals by pressing Ctrl+Alt+F2 and so on don't work. However, reboot with pressing Ctrl+SysRq (PrtSc) and REISUB works (instead of hard reboot)
Note: I am not using Nvidia card to output/render on any display on Ubuntu. Just computing/CUDA/neural networks.
There is a message in syslog which might be related to disconnecting, I am not sure:
kernel: [ 2177.344903] snd_hda_codec_hdmi hdaudioC1D0: HDMI: invalid ELD buf size -1
Output of uname -a:
Linux thinkpad 5.3.0-46-generic #38-Ubuntu SMP Fri Mar 27 17:37:05 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux
Output of nvidia-smi:
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 440.64.00 Driver Version: 440.64.00 CUDA Version: 10.2 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 GeForce RTX 2070 On | 00000000:0A:00.0 Off | N/A |
| 58% 32C P0 36W / 175W | 0MiB / 7982MiB | 0% Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| No running processes found |
+-----------------------------------------------------------------------------+
I can provide more details and any useful info to solve the problem, please help.
To do: Create my signature with system and expected eGPU configuration information to give context to my posts. I have no builds.
.
I always get kernel panics if I just unplug the egpu. There has to be some switching process that I think unloads the driver and sends an ACPI signal to the gpu. I haven't tried it yet, but I think Pop!_OS has a robust system for switching nvidia GPUs so I would try that.
Mid-2012 13" Macbook Pro (MacBookPro9,2) TB1 -> RX 460/560 (AKiTiO Node/Thunder2)
+ macOS 10.15+Win10 + Linux Mint 19.1
2012 13" MacBook Pro [3rd,2C,M] + RX 460 @ 10Gbps-TB1 (AKiTiO Thunder2) + macOS 10.14.4 [build link]
Hey, I also have an RTX2070Super with a Razer Core X for training NN connected to my XPS15 on ubuntu 19.04.
It worked straight of the box, but I need to boot with the eGPU plugged in.
I am also seeing some bottleneck on the Thunderbolt3 interface when training large models, I am never attaining 100% usage, around 85%.
I am very happy with the setup btw.
Have not tried to connect a display btw.
To do: Create my signature with system and expected eGPU configuration information to give context to my posts. I have no builds.
.UPD: Solved the problem with the fresh installation of Ubuntu 20.04. I guess the problem was related to tlp/power settings or kernel version. Now I have kernel 5.4 btw
To do: Create my signature with system and expected eGPU configuration information to give context to my posts. I have no builds.
.