FFAMax

> I believe you can already do this with tinygrad by specifying `VISIBLE_DEVICES` see https://docs.tinygrad.org/env_vars/ Got chance to verify it. VISIBLE_DEVICES=0,1,2 and VISIBLE_DEVICES=1,2 will lead to the same result. Also...

This one is correct for CUDA: to use both GPU need run something like CUDA_VISIBLE_DEVICES=0 exo --node-id=1 --node-port=8001 --listen-port=65001 CUDA_VISIBLE_DEVICES=1 exo --node-id=2 --node-port=8002 --listen-port=65002

> I'm curious what the impetus here is. What system are you running this on where you are having trouble writing to /tmp? And do you have any idea why?...

> > I'm curious what the impetus here is. What system are you running this on where you are having trouble writing to /tmp? And do you have any idea...

@AlexCheema this to help in troubleshooting. Instead of saying 'Unable open a socket' will point to more specific issue related to the file permissions. This will not resolve the root...

Hello, @Sean-fn. So, I tried it on Mac Intel-based and regular PC with PCI-E GPU on Debian/Ubuntu, at this point I do not have any success case of using it....

Just FYI. I got more luck on Description: Ubuntu 22.04.3 LTS Release: 22.04 and Description: Debian GNU/Linux 12 (bookworm) Release: 12 Codename: bookworm and Description: Ubuntu 22.10 Release: 22.10 Codename:...

Also tried run Llama-3.1-8B with 1, 2, and finally it stopped crashing only with **3** nodes which are: Linux Box (NVIDIA GEFORCE GTX 1080) 8GB Linux Box (NVIDIA GEFORCE GTX...

Do you wanna see all other peers :D or have a chance to configure peers manually should be enough?

+ one more case: node can stay alive but GPU may stuck, so need also determine no response from backend/call timeout