![]() Next, the graphics drivers have to be installed. I had to disable the following, otherwise my eGPU was not detected: On my system, I had to disable some security features in the UEFI/BIOS settings. The test-system in my case was the ThinkPad X1 Carbon 6th Gen. In this post, I wanted to share how I achieved to run a simple TensorFlow r1.6 example on the eGPU. While the setup of the eGPU on Windows is literally plug & play, there is much more to do on Ubuntu 17.10. ![]() Ultimately, I ended up with the following components: Regarding the latter one, I simply wanted to give it a shot. To compensate the former, I ordered a used one on Ebay. no official support and only very few resources for most eGPU systems on Linux.high GPU prices due to crypto-currency mining.Of course, there were two big bullet points which made me hesitate some more days: And that's why I thought: well, then do it yourself! Most or even all of them are focused on pure gaming. There was basically no benchmark available regarding eGPU setups for Machine Learning. However, this is a dual-GPU and you only get access to one of them, so the performance is actually worse than it looks like on most benchmarks.Īfter checking out some of these cloud platforms, I was still curious about how an eGPU performs with TensorFlow. The free version comes with 10 hours of GPU access on a Nvidia Tesla K80. Last but not least, I checked out FloydHub, which actually worked quite well. Unfortunately, the offer was limited to a 8-core CPU instance, while no GPU instance was available. Next, I checkout out the 300$ free credit on Google compute engine. This means that you have to pay for a full hour, even when you just run a simple example for 1 minute. I did not try out a GPU-enabled instance on AWS, because the use a billing based on a hourly rate. That's why I tried to get access to a high-performance graphics card in order to be able to train non-trivial networks and so some more serious research.Īt first, I had a look at some offers in the cloud. And it's well known that taking advantage of a GPU boosts training time by a huge margin. But everything faded into obscurity because I almost lost full interest into gaming the last years.īut this changed, since I'm spending a lot of time in deep learning since about two years. The symbiosis of having a light weight laptop at university or on the go, but still having a desktop like power horse when having some spare-time at home sounded like a dream. Therefore I am not able to use the external monitors.I remember when I read about eGPUs for the first time. The problem is that the external displays are showing my cursor (as a weird cross), but are not recognized as displays in the Ubuntu settings and do not show any windows. I updated the drivers to nvidia-driver-460Įverything seems to work fine since the eGPU is recognized by nvidia-smi:Īlso the external displays which are attached with hdmi to the eGPU are recognized by NVIDIA X Server Settings.I use lightdm as my display manager, instead of gdm3. ![]() In the nf file I also added ‘Option “AllowExternalGpus” “True”’ to the devices sections of the Geforce RTX 3070.I set the ‘Thunderbolt Security Level’ to ‘PCIe and DisplayPort - No Security’ in the bios. ![]() I also searched the internet for solutions which caused me to perform the following actions To configure the eGPU I used the following tutorial: I use a fresh install of Ubuntu 20.04 LTS as my operating system. ![]() These are used in combination with a HP Zbook studio which supports thunderbolt 3. I just bought a Razer Core X Chroma in combination with a Gigabyte RTX 3070. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |