Spent some time on this today and wanted to compare notes.

I've gone back and forth between TensorRT-LLM and SGLang for Mixtral. The interesting difference for me has been tail latency.

Trying to avoid rebuilding the whole thing twice.

This keeps coming up in conversation so I figured I'd start a thread.

One thing I keep noticing is that people underestimate consumer GPUs until they actually try to run this stuff for more than one afternoon.

Curious whether that lines up with what other people are seeing.

Curious what people here are seeing.

My current setup is MLX with GB200, and I keep running into prefill/decode split.

If you've tested something similar, what did you keep and what did you stop bothering with?

For context, I'm not chasing leaderboard numbers. I mostly want the system to feel consistent day to day.

Posting this in case it saves someone else a little time.

I'm helping a small team set up a local workflow. We don't need a giant stack, but we do care about INT4.

Simple and boring is a valid answer.

For context, I'm not chasing leaderboard numbers. I mostly want the system to feel consistent day to day.

This keeps coming up in conversation so I figured I'd start a thread.

It feels like the conversation has shifted hard toward inference and serving details lately, and honestly that makes sense once you start dealing with workstation thermals.

Wondering if this is just where the whole space is heading now.

Thought I'd post a small recipe because a lot of threads ask similar things.

Maybe useful as a starting point: LM Studio on 4090, keeping workstation thermals in mind from the start.

Maybe useful as a starting point for newer folks.

Posting this in case it saves someone else a little time.

I'm not trying to build a perfect architecture diagram. I just want a sane starting point around PyTorch and 4090.

Would rather learn from your scars than collect my own all at once.

This is mostly for local work rather than public serving. The logs were somehow calmer than I was.

Did a bit of testing and thought I'd post the result here.

Short version: PyTorch behaved well for me, but only after I stopped ignoring tokens/sec.

Nothing scientific here, just practical notes from a real setup.

This keeps coming up in conversation so I figured I'd start a thread.

The more I work with local setups, the more it seems like prefix caching decides everything downstream.

Maybe obvious to some of you, but it clicked for me this week.

I care more about stability than squeezing the last possible tokens/sec out of it.

I expected the biggest difference to come from the model, but the stack around it mattered at least as much.

Spent some time on this today and wanted to compare notes.

I'm comparing two approaches here: keep it simple on dual 3090, or squeeze more out of the stack and deal with the complexity.

Trying to avoid rebuilding the whole thing twice.

My use case is pretty ordinary: local chat, some document work, and occasional batch jobs.