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

Short version: LM Studio behaved well for me, but only after I stopped ignoring VRAM headroom.

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

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

I've gone back and forth between vLLM and LM Studio for Qwen. The interesting difference for me has been KV cache pressure.

Trying to avoid rebuilding the whole thing twice.

I figured I'd ask here before changing everything.

I'm mostly working with Kimi in Ollama on GB200. The part I'm trying to understand is consumer GPUs.

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 reranking.

I'm especially interested in mistakes worth avoiding.

This is mostly for local work rather than public serving.

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: DeepSeek, PyTorch, mini PC, with a lot of attention on speculative decoding.

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

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.

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

Curious if anyone else landed in the same place.

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.

Would rather learn from your scars than collect my own all at once. At this point the real bottleneck is my patience.

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 small local stacks decides everything downstream.

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

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 llama.cpp and Threadripper box.

I'm especially interested in mistakes worth avoiding.

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

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 speculative decoding decides everything downstream.

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

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