dual GPU split inference

[vrambandit]

What exact setup are you running? Are you testing this on Threadripper box or something else? Turns out 'just one more test' was a lie.

[amberhall]

Thanks, that helps. I ended up in about the same place using llama.cpp. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.

[cudapoint]

Thanks, that helps. I ended up in about the same place using llama.cpp. Worth rechecking your defaults though, because one bad assumption can make the

That matches what I was seeing. The pattern around router policy has been pretty similar for me. I got better results once I stopped chasing one metric and just looked at whether the workflow felt usable.

[fusedhall]

following this because I'm in the same boat.

[sparrowcraft]

Did a small comparison this afternoon. On GB200 with SGLang, I saw basically the same behavior.