best local runtime for mixed models for Gemma

[fusedhall]

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 small local stacks.

I'm especially interested in mistakes worth avoiding.

This is mostly for local work rather than public serving.

[amberhall]

What worked for me was pretty simple. If you're on llama.cpp, I'd confirm the version first and then reduce variables until the behavior is obvious. Once you know the plain path works, then add the faster path back in piece by piece.

[vectorhouse71]

I hit almost the same issue last night.

[latentscope35]

Did a small comparison this afternoon. I couldn't reproduce the slowdown until I pushed context much higher. The surprise was that memory pressure showed up before throughput collapsed.

[fusedhall]

Can you fill in a bit more context? Is this local-only or are you routing anything externally?

[mellowwire98]

Can you fill in a bit more context? Is this local-only or are you routing anything externally?

Good point. That matches what I saw on M3 Max.

[amberhall]

I don't think that's the whole story. I've seen people blame tool calling when the real issue was much more mundane. Worth testing the simpler explanation before adopting the more interesting one.

[vectorember]

I hit almost the same issue last night.

[quantshed]

Interesting, I hadn't tried that. That matches what I saw on GB200.

[vectorember]

Interesting, I hadn't tried that. That matches what I saw on GB200.

A couple things I'd check first. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at continuous batching.