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

Just sharing a setup that ended up working reasonably well: Kimi, Ollama, 4090, with a lot of attention on NVFP4.

Feel free to tear it apart if you see something dumb.

If I had to summarize the lesson so far, it would be that clean setup and observability beat clever guesses.

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

After a few runs on 7900 XTX, I ended up with decent performance, but the setup only really made sense once I adjusted around consumer GPUs.

Curious if anyone else landed in the same place.

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

Maybe useful as a starting point: MLX on GB200, keeping prefill/decode split in mind from the start.

Feel free to tear it apart if you see something dumb.

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 workstation thermals decides everything downstream.

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

This is mostly for local work rather than public serving.

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

I tried this with Llama using Ollama on 7900 XTX. The headline result was better than I expected, but consumer GPUs still ended up being the limiting factor.

Curious if anyone else landed in the same place.

This is mostly for local work rather than public serving.

Curious what people here are seeing.

I'm trying to decide whether to keep the current stack or change course. Main question is how people here are handling speculative decoding.

Would appreciate practical advice rather than theory.

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 vLLM and M3 Max.

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

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

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

I tried this with Gemma using vLLM on dual 3090. The headline result was better than I expected, but quant drift still ended up being the limiting factor.

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

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

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

This isn't revolutionary, but the combination of PyTorch and 5090 felt surprisingly usable once I tuned around workstation thermals.

If someone wants the exact settings I can clean them up and post them.

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

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 token cost.

Simple and boring is a valid answer.