Hello World,

I’m working on getting a ClearML Server spun up on the same computer as GPUs. Question about implementation details..

Fresh build using this guide, 128mb RAM, 2 GPUs

https://gist.github.com/amir-saniyan/b3d8e06145a8569c0d0e030af6d60bea

I was told it would work better to have a second hard drive for the server. I’ve studied a lot but never implemented.

So question, it appears, in a practical sense, this means to mount the second drive, and then bind them in ClearML’s configuration? Also, thinking about Docker and slower access speed for Docker mounts and such, I’m thinking it may be more performant, in general, to never bind to the m2 ssd for anything but training neural nets? True? False? Any advice from people with practical experience here?

edit: Tiny raft, big ocean, thank you :)